Compare commits

...

43 Commits

Author SHA1 Message Date
Jan Höppner
04407d24fc New release s390-tools-2.19.0
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 17:05:37 +01:00
Jan Höppner
bcb8ea09d1 lshwc: Fix compile error for gcc <8.1
With gcc < 8.1 the following compile error occurs:

lshwc.c:52:37: error: initializer element is not constant
 static unsigned int read_interval = ioctlsleep, cfvn, csvn, authorization;
                                     ^~~~~~~~~~

Even though this gcc bug was fixed in gcc 8.1+, change the constant
definition to a #define to make the tool compile with older compilers as
well.

Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69960
Link: https://gcc.gnu.org/bugzilla/show_bug.cgi?id=66618
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Benjamin Block
88619b6dba chreipl-fcp-mpath: documentation and man page
Add a README.md for chreipl-fcp-mpath and infrastructure to build and
install a man page generated from the README.md. Building and installing
of the man page is only done when supplying the make option `ENABLE_DOC=1`
(per default disabled). The `pandoc` utility is used to build the man page.

The README.md is reused since there is no other tool or command as part of
chreipl-fcp-mpath that needs to be documented in a separate man page.

Also add information about chreipl-fcp-mpath to the main README.md and
CHANGELOG.md.

Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Benjamin Block
e127b89624 chreipl-fcp-mpath: optional dracut config to bundle toolset into initrd
Add a dracut configuration file that might optionally be installed along
with the toolset (per default to: $(USRLIBDIR)/dracut/dracut.conf.d). It
causes the udev rules and helper scripts to be included into an initial
ramdisk, when (re-)built with dracut; along with any dependency.

This is only done when passing `HAVE_DRACUT` set to 1 during the
installation (default is 0), e.g.: make install HAVE_DRACUT=1.

Enabling the toolset during the initial ramdisk phase has the advantage
that paths can be audited earlier in the life cycle of this Linux
runtime, e.g. to record the WWID in the ID-file.

Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Benjamin Block
3591b94c42 chreipl-fcp-mpath: try to change the re-IPL target if the old one went away
In the final step of the udev rules toolset, we either know that the
current event subject is the re-IPL target, a dm-multipath device with
the re-IPL target, or represents the same volume based on its WWID. As
such it is a candidate to replace the current re-IPL target.

The new helper `chreipl-fcp-mpath-try-change-ipl-path` will use the
subject itself - in case it is a single scsi disk -, or the dm-multipath
device, to test whether there is a path in a good state.

"Good state" is based on the zfcp device driver state
(<SDEV>/zfcp_failed, <SDEV>/zfcp_in_recovery), the scsi_transport_fc
port state (<SDEV>/../../fc_remote_ports/rport-*/port_state), and the
scsi device state (<SDEV>/state). A path is only considered if they all
indicate a device in good conditions.

If such a path is found, the helper will try to set it as new re-IPL
target regardless of whether the current re-IPL is still in good shape
or not. This is by design, and done to reduce complexity in further
state checking, and prevention of races with overlapping events in udev
(when executing in parallel workers).

Whenever a new re-IPL target is selected and set in
/sys/firmware/reipl/fcp/{device,wwpn,lun}, the helper will also update
the records in the ID-file - otherwise they might appear as stale, when
they aren't.

This step in the udev rule processing might also result in log messages
written to the syslog (using the utility `logger` from util-linux;
writing to /dev/log).

In case the re-IPL target is changed, a message with level notice is
logged, informing about the new target.

In case no good path is found as part of a dm-multipath device, a
message with level critical is logged, as it might result in a failed
re-IPL if no path is available.

Lastly, when changing the information in /sys/firmware/reipl/fcp/ fails
for whatever reason, a message with level alert is logged, as the
written information might be inconsistent and must be audited/corrected
manually by an operator.

Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Benjamin Block
2a29a28f78 chreipl-fcp-mpath: if event subject is not re-IPL target, test if WWID matches
When the first test, whether the current event subject represents the
re-IPL target failed, we attempt a second test, using the recorded WWID
of a previously successful attempt.

The new helper `chreipl-fcp-mpath-is-ipl-vol` utilizes the information
previously stored in the ID-file (read under lock via `flock`).

The recorded WWID is compared with the one of the event subject; if it
matches we know that it addresses the same volume, and might be used as
alternative re-IPL target.

Additionally we also compare the current re-IPL triplet with the one
recorded in the ID-file. If it doesn't match anymore, we assume the
recorded WWID is stale - e.g., because the operator changed the re-IPL
target manually - and the previous check invalid.

If both the WWID match, and the re-IPL triplet is still the same as
previously recorded, the toolset may continue with the subject, and use
it as replacement for the current re-IPL target.

Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Benjamin Block
cb831aea44 dbginfo.sh: collect config- and debug-data for chreipl-fcp-mpath
When built with `D=1` chreipl-fcp-mpath automatically creates debug data
for the helpers that run during udev event processing. This is the only
way to effectively debug what is happening during event processing in
chreipl-fcp-mpath.

Collect this data under its default path in /run/udev/.

Also collect the chreipl-fcp-mpath statefile which contains WWID, Device
Bus-ID and LUN of the last path that was detected to represent a FCP
re-IPL device.

Suggested-by: Steffen Maier <maier@linux.ibm.com>
Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Benjamin Block
04be704083 chreipl-fcp-mpath: record the event subject WWID if it repr. the tgt
When we have identified the current event subject to represent the
re-IPL target, we record its WWID for future identification in a
stateful ID-file (per default: /run/udev/chreiplzfcpmp-ipl-volume-id). In
addition to the WWID, we also record the current re-IPL triplet
(<Dev-Bus-ID>:<WWPN>:<LUN>), so that when that changes - e.g. due to an
operator manually changing the re-IPL target -, we know that the
recorded WWID is stale.

This record may be used in cases when the current re-IPL target is
completely gone from the system, so we can't used it as comparison
object for when events arrive for paths that go to the same volume, but
don't have the same I_T_L nexus. They however have the same WWID. We
may use these (new) paths as replacement for the one that is completely
gone.

The new helper `chreipl-fcp-mpath-record-volume-identifier` uses the
kernel scsi-device attribute `wwid` as source for the WWID (verbatim).

As with reading the re-IPL firmware information, when writing to the
ID-file, a lock is taken via `flock`, to prevent overlapping
writes/reads to the file.

Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Benjamin Block
1bcfcd3253 chreipl-fcp-mpath: test if event subject represents re-IPL target
Now that we know that the system uses s390x IPL and the current re-IPL
target is from FCP, whenever we get a udev event that indicates a path
state changed, we have to figure out whether it affects the path that is
currently set as re-IPL target, or goes to the same volume (so we might
use it as alternative re-IPL path).

Add a new helper `chreipl-fcp-mpath-is-ipl-tgt` for this task.

Based on the information provided in
/sys/firmware/reipl/fcp/{device,wwpn,lun} it figures out whether the
current event subject has the same triplet <Dev-Bus-ID>:<WWPN>:<LUN> (in
T10 SCSI: I_T_L nexus), or whether one of its parts (in case of
dm-multipath) has.

If true, we know for sure, that we deal with an event affecting the
current re-IPL setting, and continue.

When accessing the re-IPL firmware information a lock is taken via the
`flock` utility (from util-linux). This is done so concurrent changes
from the toolset, that are added in a later commit, don't result in
inconsistent reads.

Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Benjamin Block
2dbaf9f991 chreipl-fcp-mpath: test if the system uses s390x IPL and re-IPL is from FCP
Add second filter stage to the udev rules: only continue with the
toolset, if the system uses s390x IPL and if the current re-IPL target
is from FCP.

For the test, whether the next re-IPL target is planned to be done from
FCP, a helper `chreipl-fcp-mpath-is-reipl-zfcp` is added, and the
information in /sys/firmware/reipl/reipl_type is used.

If either one or both of the tests are not true, the udev rules will
exit.

For debugging, the toolset can be built with the Make variable `D` set
to 1 (e.g.: make D=1; make install D=1). This causes the toolset to
write shell traces of the helper to $(DEBUGOUTDIR) (per default:
/run/udev); otherwise no output is generated. `dbginfo.sh` will be
changed to collect this data automatically.

Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Benjamin Block
2ea20094d0 chreipl-fcp-mpath: toolset skeleton with initial set of udev rules
Add a new tool to s390-tools: chreipl-fcp-mpath. A toolset based on udev
to change the active re-IPL target if the current target goes into an
error-state, due to for example a pulled cable.

Add the base infrastructure consisting of a Makefile and a base set of
udev rules. Integrate the Makefile into s390-tools Make machinery.

The rules filter events base on the event subject type (dm-multipath,
scsi disk), udev action (change, or add), and device mapper action
(path failure, or path reinstated). No further processing is done right
now in case of a match; if neither of these criteria matches, the
rules exit, and the toolset won't be invoked.

Reviewed-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Joern Siglen
a80c14d72c dbginfo.sh: addtional timeout and omitting a report
running mpio data disks with offline paths might cause dbginfo
to hang up on the command "blockdev --report" -> we remove
it to ensure a non blocking data collection.
In addition we add a second kill step with the same timeout
value to retry stopping a command

Signed-off-by: Joern Siglen <siglen@de.ibm.com>
Reviewed-by: Mario Held <mario.held@de.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Thomas Huth
10518da618 zipl/boot: Always clear COMMAND_LINE_EXTRA region
Always clear the COMMAND_LINE_EXTRA region, also in case sclp_setup()
fails. If the region is not cleared properly, there might be junk
in there after a reboot, causing confusion for the kernel later.

While we're at it, also switch to COMMAND_LINE_EXTRA_SIZE instead
of COMMAND_LINE_SIZE, since that constant is more appropriate here.

Closes: https://github.com/ibm-s390-linux/s390-tools/pull/122
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Thomas Huth <thuth@redhat.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Marc Hartmayer
e9d1e168b2 zdump: rename ZG_ACTION_STDOUT to ZG_ACTION_COPY
With an upcoming patch the output target can be different to `stdout`.
Therefore, the term `ZG_ACTION_STDOUT` is no longer generic enough. In addition,
rename `do_write` to `do_copy` to match the enum name.

Reviewed-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Marc Hartmayer
7567506cca zdump: stdout_write_dump: replace magic value
Replace magic value with something easier to read.

Reviewed-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Marc Hartmayer
524017ac06 zdump: stdout_write_dump: if dfo_size() == 0 then don't enter the loop
While at it, decrease the scope of @cnt and @rc.

Reviewed-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Marc Hartmayer
15774dbe32 zdump: stdout_write_dump: replace dfo_size() calls
Read the `dfo_size()` value only once and then reuse it.

Reviewed-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Marc Hartmayer
f247850ff0 zdump: only mount command uses the optional arguments
Currently, for some commands (e.g. `zgetdump unmount`) unused positional
arguments are silently ignored, for some commands (`zgetdump mount`) they're
not. Therefore let's fix this by always returning an error message and exit code
if too many positional arguments are given.

Reviewed-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexandra Winter
800ef7cb13 hsci: Compatibility with old hsci tool
In case the hsci tool was upgraded on a running system, it can happen
that an hsci interface exists, that was created by an older version of
the tool (and thus has no veth pair). Add support to show and delete
such an hsci interface.

Signed-off-by: Alexandra Winter <wintera@linux.ibm.com>
Reviewed-by: Wenjia Zhang <wenjia@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexandra Winter
48cf3b3809 hsci: Compatibility with old kernel
Without kernel commits
f7936b7b2663 s390/qeth: Update MACs of LEARNING_SYNC device
4e20e73e631a s390/qeth: Switchdev event handler
60bb1089467d s390/qeth: Register switchdev event handler
HSCI cannot be used with multiple MACs, but the single MAC usecase needs
to be supported even with old kernel. So manually setting the same single
MAC on hsci, HiperSockets and the external interface is still required.
The hsci itself is not a bridgeport, so a static forwarding rule in the
bridge is also required. It seems cleaner to use the initial MAC of the
veth hsci as single static MAC, than the initial MAC of the HiperSockets
interface.

Even with an updated kernel a single static MAC interface, that is not
subject to ageing and re-learning, is beneficial when hsci is used as
a single MAC interface.

Signed-off-by: Alexandra Winter <wintera@linux.ibm.com>
Reviewed-by: Wenjia Zhang <wenjia@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexandra Winter
99c8c27302 hsci: Support multiple MAC addresses
The first implementation of HSCI (HiperSockets Converged Interface)
supports only a single static MAC address per HSCI and thus constructs
like MACVTAP or OpenVSwitch cannot be used on top of it.

Kernel commits
f7936b7b2663 s390/qeth: Update MACs of LEARNING_SYNC device
4e20e73e631a s390/qeth: Switchdev event handler
60bb1089467d s390/qeth: Register switchdev event handler
changed the behaviour of a LEARNING_SYNC HiperSockets (HS) device
such that any MAC address that is added or learned on a non-isolated
bridgeport of the HSCI bridge is set on the LEARNING_SYNC HS device and
its isolated siblings.

So no new trigger is required for a HS device with the bridge attribute
LEARNING_SYNC to support multiple MACs on a northbound HSCI interface.
However the first implementation of the hsci tool uses the bridge itself
as northbound HSCI interface. This is not treated as a real bridgeport.
So we add a veth bridgeport as northern link for HSCI.

Signed-off-by: Alexandra Winter <wintera@linux.ibm.com>
Reviewed-by: Wenjia Zhang <wenjia@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexandra Winter
a5c3af1361 hsci: Idempotency and cleanup
Instead of checking for ideal starting conditions, try to achieve
the desired outcome and accept conditions that do not prevent it.

This allows the user to get to a defined state, even if the
hsci interface was manipulated by other means than the hsci tool,
e.g. by ip commands or bridge commands

Minor cleanup:
hsci show:
	- display PNET_ID, even if only one hsci port has one
	  (e.g. if External is a bond)
	- only 1 header for mulitple hscis
hsci add:
	- sanitize error messages
	- allow nonsense HSCI of 2 HiperSockets

Signed-off-by: Alexandra Winter <wintera@linux.ibm.com>
Reviewed-by: Wenjia Zhang <wenjia@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexander Egorenkov
353c23a157 zdump: Include dfi_mem_chunk.h only if required
Don't include dfi_mem_chunk.h in dfi.h because not every module including
dfi.h needs it.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexander Egorenkov
d371ba0db8 zdump: Cleanup zgetdump.h C header
Separate the definition of a struct from the declaration of a global
variable of the type. To improve readability.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexander Egorenkov
8308143008 zdump/stdout: Extract stdout.h header from zgetdump.h
Create a separate C header for STDOUT to separate independent things and
make it more readable.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexander Egorenkov
d8871197ea zdump/opts: Make command-line argument parsing independent of DFO
Set DFO format after the parsing of command-line arguments is done.
The command-line argument parsing should be independent of DFO.
This improves testability of the command-line argument parsing and
separates independent things from each other.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexander Egorenkov
df338a3bac zdump/opts: Make parsing of command line arguments testable
Make the C module responsible for the parsing of command-line arguments
independent of other global variables. This improves its testability.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexander Egorenkov
1005e7be7e zdump: Don't use god-like C headers
Get rid of C headers which include everything. This is bad for testing and
maintainability. C files should include only what they need and no more.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexander Egorenkov
afb14cba7a zdump/zfuse: Extract zfuse.h header from zgetdump.h
Create a separate C header for FUSE to separate independent things and
make it more readable.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 15:12:10 +01:00
Alexander Egorenkov
1e450e6f7b zdump/dt: Move declaration of DT structs to dt.h
Move DT structs to the header where they belong to.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 14:59:57 +01:00
Alexander Egorenkov
696a007b62 zdump/dfi: Move declaration of DFI structs to dfi.h
Move DFI structs to the header where they belong to.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 14:59:57 +01:00
Alexander Egorenkov
ea5f3f4a1c zdump/dfo: Move declaration of DFO structs to dfo.h
Move DFO structs to the header where they belong to.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 14:59:57 +01:00
Alexander Egorenkov
6afd5e97d1 zdump/dfi: Extract dfi_vmcoreinfo.h header from dfi.h
Create a separate C header for DFI VMCOREINFO to separate
independent things and make it more readable.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 14:59:57 +01:00
Alexander Egorenkov
b1da8bbbe9 zdump/dfi: Introduce symbolic constants for OLDMEM base and size
Replace magic values with descriptive names to improve readability.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 14:59:57 +01:00
Alexander Egorenkov
6ece595532 zdump/dfi_mem_chunk: Validate parameters passed to dfi_mem_phys_read()
To prevent segmentation faults when the function dfi_mem_phys_read() is
given invalid parameters, check that the given memory range falls within
a known physical memory chunk, just like it is done for dfi_mem_virt_read().

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 14:59:57 +01:00
Alexander Egorenkov
40d4aa38e9 zdump/dfi_mem_chunk: Rename dfi_mem_read_rc() to dfi_mem_virt_read()
The new name describes the purpose of the function better and
emphasizes its difference with the function dfi_mem_phys_read().

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 14:59:57 +01:00
Alexander Egorenkov
c58cfd1ec1 zdump/dfi_mem_chunk: Replace dfi_mem_read() with dfi_mem_read_rc()
Always use the function dfi_mem_read_rc() which verifies that the given
address and size fall within a valid memory chunk. We cannot trust
user's input.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 14:59:57 +01:00
Alexander Egorenkov
02086f26d2 zdump/dfi_mem_chunk: Don't expose private function dfi_mem_chunk_virt_add()
The function dfi_mem_chunk_virt_add() is used only internally and should
not be exported to prevent misuse.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 14:59:57 +01:00
Alexander Egorenkov
7a2666d727 zdump/dfi: Remove redundant extern from function declarations in headers
The keyword extern is redundant for function declarations in C headers.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 14:11:29 +01:00
Alexander Egorenkov
2df532c1fa zdump: Extract dfi_mem_chunk module
To make dfi_mem_chunk API unit testable.

Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 11:50:58 +01:00
Eduard Shishkin
969a439aaa zipl: check add_envblk predicate when setting stage3_parms
Set stage3_parms.envblk_len to 0, if add_envblk is false.
This fixes segmentation fault when performing job with
not defined zIPL environment (e.g. installing SCSI dump)

Signed-off-by: Eduard Shishkin <edward6@linux.ibm.com>
Tested-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 11:50:58 +01:00
Fedor Loshakov
f8331a4b8e ziomon: correct throughput calculation in ziorep_printers.cpp
During evaluation of the collected ziomon data from the system with following
multipath configuration:

$ multipath -ll
3600507630bffc3200000000000005294 dm-0 IBM,2107900
size=20G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='service-time 0' prio=50 status=active
  |- 1:0:0:1083457618 sdd 8:48  active ready running
  `- 0:0:0:1083457618 sda 8:0   active ready running
3600507630bffc3200000000000005394 dm-1 IBM,2107900
size=20G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='service-time 0' prio=50 status=active
  |- 1:0:0:1083457619 sdf 8:80  active ready running
  `- 0:0:0:1083457619 sdb 8:16  active ready running
3600507630bffc3200000000000005494 dm-2 IBM,2107900
size=20G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='service-time 0' prio=50 status=active
  |- 1:0:0:1083457620 sdg 8:96  active ready running
  `- 0:0:0:1083457620 sdc 8:32  active ready running
3600507630bffc3200000000000005594 dm-3 IBM,2107900
size=20G features='1 queue_if_no_path' hwhandler='1 alua' wp=rw
`-+- policy='service-time 0' prio=50 status=active
  |- 1:0:0:1083457621 sdh 8:112 active ready running
  `- 0:0:0:1083457621 sde 8:64  active ready running

there were low throughput values for adapter noticed in ziorep_utilization
virtual adapter report:

$ ziorep_utilization adp_line_speed.log
...
CHP Bus-ID  |qdio util.%|queu|fail|-thp in MB/s-|I/O reqs-|
 ID            max   avg full  erc     rd    wrt   rd  wrt
2021-08-18 12:56:44
 61/0.0.1946  32.8   3.1    0    0   0.0   84.5    21 204K
12:57:04
...

there were also low values for throughput of devices noticed in
ziorep_traffic report. And were no aggregation of throughput for
different aggregation parameters provided:

$ ziorep_traffic adp_line_speed.log -i 0
       WWPN                LUN       |I/O rt MB/s|thrp in MB/s-|...
                                        min   max    avg  stdev ...
2021-08-18 12:58:24
0x500507630b09c320:0x4052409400000000   0.0 219.2  85.7  3.352K ...
0x500507630b09c320:0x4053409400000000   0.0 348.6  84.6  3.330K ...
0x500507630b09c320:0x4054409400000000   0.0 268.6  84.0  3.317K ...
0x500507630b09c320:0x4055409400000000   0.0 354.2  83.7  3.312K ...

$ ziorep_traffic adp_line_speed.log -i 0 -Cu
 Bus-ID |I/O rt MB/s|thrp in MB/s-|...
           min   max    avg  stdev ...
2021-08-18 12:58:24
0.0.1946   0.0 354.2  84.5  3.328K ...

although, iostat tool showed decent throughput values for each scsi disk, which
belongs to investigated zfcp adapter. In this example each scsi disk has
throughput more than 300MB/s, which is much larger, than 84.5MB/s
throughput for the adapter:

$ iostat -x 1
...
Device            r/s     w/s     rkB/s     wkB/s ...
dasda            0.00    0.00      0.00      0.00 ...
dasdb            0.00    0.00      0.00      0.00 ...
sda              0.00 2556.00      0.00 327168.00 ...
sdc              0.00 2554.00      0.00 326912.00 ...
sdb              0.00 2559.00      0.00 327552.00 ...
sdf              0.00 2865.00      0.00 366720.00 ...
sdd              0.00 2843.00      0.00 363904.00 ...
sdg              0.00 2845.00      0.00 364160.00 ...
sde              0.00 2471.00      0.00 315416.00 ...
sdh              0.00 2768.00      0.00 353768.00 ...
dm-0             0.00 5399.00      0.00 691072.00 ...
dm-1             0.00 5424.00      0.00 694272.00 ...
dm-2             0.00 5400.00      0.00 691200.00 ...
dm-3             0.00 5240.00      0.00 669312.00 ...

Use Frameset interval size for calculation of throughput of zfcp adapter
instead of d2c (dispatch to complete) time of request, which is used for I/O
rate calculation. Also use Frameset interval instead of total_latency.sum
(which in fact is a sum of d2c times for each repuest in specifed interval).
Use calc_avg() function for throughput calculation.

With fix applied, virtual adapter report now contains correct value for adpater
throughput for specified interval:

$ ziorep_utilization adp_line_speed.log
...
CHP Bus-ID  |qdio util.%|queu|fail|-thp in MB/s-|I/O reqs-|
 ID            max   avg full  erc     rd    wrt   rd  wrt
2021-08-18 12:56:44
 61/0.0.1946  32.8   3.1    0    0   0.0  1.336K   21 204K
...

With fix applied, traffic report now contains correct values of
throughput for each device and aggregation now works correctly:

$ ziorep_traffic adp_line_speed.log -i 0
       WWPN                LUN       |I/O rt MB/s|thrp in MB/s-|...
                                        min   max    avg  stdev ...
2021-08-18 12:58:24
0x500507630b09c320:0x4052409400000000   0.0 219.2 339.6  3.352K ...
0x500507630b09c320:0x4053409400000000   0.0 348.6 335.4  3.330K ...
0x500507630b09c320:0x4054409400000000   0.0 268.6 333.1  3.317K ...
0x500507630b09c320:0x4055409400000000   0.0 354.2 331.9  3.312K ...

$ ziorep_traffic adp_line_speed.log -i 0 -Cu
 Bus-ID |I/O rt MB/s|thrp in MB/s-|...
           min   max    avg  stdev ...
2021-08-18 12:58:24
0.0.1946   0.0 354.2 1.340K 3.328K ...

Signed-off-by: Fedor Loshakov <loshakov@linux.ibm.com>
Reviewed-by: Benjamin Block <bblock@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-11-10 11:50:58 +01:00
Jan Höppner
895dc80580 Prepare for next release
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
2021-10-01 17:45:29 +02:00
61 changed files with 3059 additions and 1012 deletions

View File

@@ -2,6 +2,7 @@ List of all individuals having contributed content to s390-tools
----------------------------------------------------------------
- Alexander Egorenkov
- Alexandra Winter
- Alexey Ishchuk
- Andreas Herrmann
- Andre Wild
@@ -110,6 +111,7 @@ List of all individuals having contributed content to s390-tools
- Swen Schillig
- Taraka R. Bodireddy
- Thomas Heidrich
- Thomas Huth
- Thomas Richter
- Thomas Spatzier
- Thomas Weber

View File

@@ -1,5 +1,24 @@
Release history for s390-tools (MIT version)
--------------------------------------------
* __v2.19.0 (2021-11-10)__
For Linux kernel version: 5.15
Add new tools / libraries:
- chreipl-fcp-mpath: New toolset that uses multipath information to change
the configured FCP re-IPL path on detecting issues with the current path
Changes of existing tools:
- dbginfo.sh: Add retry timeout and remove possible blocking "blockdev --report"
- dbginfo.sh: Collect config- and debug-data for chreipl-fcp-mpath
- hsci: Add support for multiple MAC addresses
Bug Fixes:
- lshwc: Fix compile error for gcc <8.1
- zdump: Various clean-ups and fixes
- ziomon: Correct throughput calculation in ziorep_printers
- zipl: Fix segmentation fault when setting stage3_parms
* __v2.18.0 (2021-10-01)__
For Linux kernel version: 5.14

View File

@@ -16,7 +16,7 @@ TOOL_DIRS = zipl zdump fdasd dasdfmt dasdview tunedasd \
vmconvert vmcp man mon_tools dasdinfo vmur cpuplugd ipl_tools \
ziomon iucvterm hyptop cmsfs-fuse qethqoat zfcpdump zdsfs cpumf \
systemd hmcdrvfs cpacfstats zdev dump2tar zkey netboot etc zpcictl \
genprotimg lsstp hsci hsavmcore
genprotimg lsstp hsci hsavmcore chreipl-fcp-mpath
SUB_DIRS = $(BASELIB_DIRS) $(LIB_DIRS) $(TOOL_DIRS)

View File

@@ -270,6 +270,10 @@ Package contents
can be time consuming and prevent the HSA memory from being reused
by other LPARs.
* chreipl-fcp-mpath:
Use multipath information to change the configured FCP re-IPL path on
detecting issues with the current path.
For more information refer to the following publications:
* "Device Drivers, Features, and Commands" chapter "Useful Linux commands"
@@ -307,11 +311,11 @@ build options:
This table lists additional build or install options:
| __COMPONENT__ | __OPTION__ | __TOOLS__ |
|------------------|:----------------------------:|:--------------:|
| dracut | `HAVE_DRACUT` | zdev |
| initramfs-tools | `HAVE_INITRAMFS` | zdev |
| | `ZDEV_ALWAYS_UPDATE_INITRD` | zdev |
| __COMPONENT__ | __OPTION__ | __TOOLS__ |
|------------------|:----------------------------:|:-----------------------:|
| dracut | `HAVE_DRACUT` | zdev, chreipl-fcp-mpath |
| initramfs-tools | `HAVE_INITRAMFS` | zdev |
| | `ZDEV_ALWAYS_UPDATE_INITRD` | zdev |
The s390-tools build process uses "pkg-config" if available and hard-coded
compiler and linker options otherwise.
@@ -465,3 +469,14 @@ the different tools are provided:
(libxml2-devel.rpm), and libcurl version 7.59 or newer (libcurl-devel.rpm).
Tip: you may skip the libkmipclient build by adding `HAVE_OPENSSL=0`,
`HAVE_JSONC=0`, `HAVE_LIBXML2=0`, or `HAVE_LIBCURL=0` to the make invocation.
* chreipl-fcp-mpath:
For a complete list and documentation of the requirements, installation and
uninstallation, please see
[chreipl-fcp-mpath/README.md](chreipl-fcp-mpath/README.md).
Summarized: chreipl-fcp-mpath requires GNU Bash, GNU Core Utilities,
util-linux, udev, and multipath-tools. When using `HAVE_DRACUT=1` with the
make invocation, it also requires dracut. When using `ENABLE_DOC=1` with the
make invocation to build a man page and render the README.md as HTML, make
further requires pandoc, GNU awk, and GNU Gzip for the build process.

13
chreipl-fcp-mpath/.gitignore vendored Normal file
View File

@@ -0,0 +1,13 @@
# build artifacts
/chreipl-fcp-mpath-common.sh
/chreipl-fcp-mpath-is-ipl-tgt
/chreipl-fcp-mpath-is-ipl-vol
/chreipl-fcp-mpath-is-reipl-zfcp
/chreipl-fcp-mpath-record-volume-identifier
/chreipl-fcp-mpath-try-change-ipl-path
/dracut/dracut.conf.d/70-chreipl-fcp-mpath.conf
/README.html
/README.pdf
/chreipl-fcp-mpath.md
/chreipl-fcp-mpath.7
/chreipl-fcp-mpath.7.gz

259
chreipl-fcp-mpath/Makefile Normal file
View File

@@ -0,0 +1,259 @@
# SPDX-License-Identifier: MIT
#
# chreipl-fcp-mpath: use multipath information to change FCP IPL target
# (C) Copyright IBM Corp. 2021
#
# Uses the following system-utilities (and shell-builtins):
# Utilities list in GNU Make Conventions:
# https://www.gnu.org/software/make/manual/make.html#Utilities-in-Makefiles
# Those necessary for sourced Makefiles:
# - ../common.mak
# - chreipl-fcp-mpath.mak
# bash:
# - bash
# If $(ENABLE_DOC) is `1`:
# GNU Gzip:
# - gzip
# GNU awk:
# - gawk
override SHELL := /bin/bash
override .SHELLFLAGS := -O globstar -O nullglob -O extglob -c
# Include common s390-tools definitions
include ../common.mak
# Include common chreipl-fcp-mpath definitions
include chreipl-fcp-mpath.mak
# Local setting: .make.config
# You may create a file named like this in the same directory as this
# Makefile, and customize the build this way (e.g. re-define variables
# set in `chreipl-fcp-mpath.mak`, or define a `CHREIPLZFCPMP_POST_INSTALL`
# that is automatically called after each installation)
ifneq ($(wildcard .make.config),)
include $(wildcard .make.config)
endif
#
## Build
#
.PHONY: chreipl-fcp-mpath chreipl-fcp-mpath-clean
chreipl-fcp-mpath:
chreipl-fcp-mpath-clean:
all: chreipl-fcp-mpath
clean: chreipl-fcp-mpath-clean
# common function used in the helper scripts
CHREIPL_FCP_MPATH_COMMON := \
chreipl-fcp-mpath-common.sh
CHREIPL_FCP_MPATH_UDEV_HELPER := \
chreipl-fcp-mpath-is-ipl-tgt \
chreipl-fcp-mpath-is-ipl-vol \
chreipl-fcp-mpath-is-reipl-zfcp \
chreipl-fcp-mpath-record-volume-identifier \
chreipl-fcp-mpath-try-change-ipl-path
$(CHREIPL_FCP_MPATH_UDEV_HELPER) $(CHREIPL_FCP_MPATH_COMMON): $(MAKEFILE_LIST)
$(CHREIPL_FCP_MPATH_UDEV_HELPER) $(CHREIPL_FCP_MPATH_COMMON): % : %.in
$(call chreiplzfcpmp-sed-buildvar-replace,$(<),$(@))
chmod a+x $(@)
$(CHREIPL_FCP_MPATH_UDEV_HELPER): $(CHREIPL_FCP_MPATH_COMMON)
.PHONY: chreipl-fcp-mpath-udev-helper-clean
chreipl-fcp-mpath-udev-helper-clean:
rm -f $(CHREIPL_FCP_MPATH_UDEV_HELPER) $(CHREIPL_FCP_MPATH_COMMON)
udev/rules.d/70-chreipl-fcp-mpath.rules: $(CHREIPL_FCP_MPATH_UDEV_HELPER)
chreipl-fcp-mpath: udev/rules.d/70-chreipl-fcp-mpath.rules
chreipl-fcp-mpath-clean: chreipl-fcp-mpath-udev-helper-clean
dracut/dracut.conf.d/70-chreipl-fcp-mpath.conf: $(MAKEFILE_LIST)
dracut/dracut.conf.d/70-chreipl-fcp-mpath.conf: % : %.in
$(call chreiplzfcpmp-sed-buildvar-replace,$(<),$(@))
.PHONY: chreipl-fcp-mpath-dracut-clean
chreipl-fcp-mpath-dracut-clean:
rm -f dracut/dracut.conf.d/70-chreipl-fcp-mpath.conf
chreipl-fcp-mpath: dracut/dracut.conf.d/70-chreipl-fcp-mpath.conf
chreipl-fcp-mpath-clean: chreipl-fcp-mpath-dracut-clean
ifeq ($(ENABLE_DOC),1)
.PHONY: chreipl-fcp-mpath-doc
chreipl-fcp-mpath-doc: README.html chreipl-fcp-mpath.7
ifeq ($(ENABLE_DOC_PDF),1)
chreipl-fcp-mpath-doc: README.pdf
README.pdf: PANDOCFLAGS += -M title="chreipl-fcp-mpath" -M subtitle="README"
endif
README.html: PANDOCFLAGS += -M title="chreipl-fcp-mpath" -M subtitle="README"
chreiplzfcpmp-doc-man-meta = \
-M title="CHREIPL-FCP-MPATH" \
-V header="Administrator Manual" \
-V section="7" \
-V footer="s390-tools $(S390_TOOLS_RELEASE)"
chreipl-fcp-mpath.7: PANDOCFLAGS += $(chreiplzfcpmp-doc-man-meta)
.INTERMEDIATE: chreipl-fcp-mpath.md
chreipl-fcp-mpath.md: README.md
gawk -- '/NOT-IN-MAN \{/,/NOT-IN-MAN \}/ { next } { print }' $(<) > $(@)
.PHONY: chreipl-fcp-mpath-doc-clean
chreipl-fcp-mpath-doc-clean:
rm -f README.html README.pdf chreipl-fcp-mpath.md chreipl-fcp-mpath.7
chreipl-fcp-mpath: chreipl-fcp-mpath-doc
chreipl-fcp-mpath-clean: chreipl-fcp-mpath-doc-clean
endif
#
## Install
#
.PHONY: chreipl-fcp-mpath-install
# The content of `CHREIPLZFCPMP_POST_INSTALL` (bash script) is automatically
# called *after* installing chreipl-fcp-mpath during `make install`. If not
# defined (the default), nothing happens. You may define this on the make
# command line, or by creating a `.make.config` and defining the variable in
# there.
chreipl-fcp-mpath-install:
$(CHREIPLZFCPMP_POST_INSTALL)
install: chreipl-fcp-mpath-install
# install udev rules
INSTDIRS += $(UDEVRULESDIR)
.PHONY: chreipl-fcp-mpath-install-udev-rules
chreipl-fcp-mpath-install-udev-rules: | $(DESTDIR)$(UDEVRULESDIR)
chreipl-fcp-mpath-install-udev-rules: udev/rules.d/70-chreipl-fcp-mpath.rules
$(INSTALL_DATA) -t $(DESTDIR)$(UDEVRULESDIR) \
udev/rules.d/70-chreipl-fcp-mpath.rules
chreipl-fcp-mpath-install: chreipl-fcp-mpath-install-udev-rules
# install udev helper programs
INSTDIRS += $(UDEVDIR)
.PHONY: chreipl-fcp-mpath-install-udev-helper
chreipl-fcp-mpath-install-udev-helper: | $(DESTDIR)$(UDEVDIR)
chreipl-fcp-mpath-install-udev-helper: $(CHREIPL_FCP_MPATH_UDEV_HELPER)
$(INSTALL_EXEC) -t $(DESTDIR)$(UDEVDIR) $(CHREIPL_FCP_MPATH_UDEV_HELPER)
chreipl-fcp-mpath-install: chreipl-fcp-mpath-install-udev-helper
# install common library files
INSTDIRS += $(CHREIPLZFCPMPDIR)
.PHONY: chreipl-fcp-mpath-install-libfiles
chreipl-fcp-mpath-install-libfiles: | $(DESTDIR)$(CHREIPLZFCPMPDIR)
chreipl-fcp-mpath-install-libfiles: $(CHREIPL_FCP_MPATH_COMMON)
$(INSTALL_DATA) -t $(DESTDIR)$(CHREIPLZFCPMPDIR) \
$(CHREIPL_FCP_MPATH_COMMON)
chreipl-fcp-mpath-install: chreipl-fcp-mpath-install-libfiles
ifeq ($(HAVE_DRACUT),1)
# install dracut config files
INSTDIRS += $(DRACUTCONFDIR)
.PHONY: chreipl-fcp-mpath-install-dracut-config
chreipl-fcp-mpath-install-dracut-config: | $(DESTDIR)$(DRACUTCONFDIR)
chreipl-fcp-mpath-install-dracut-config: dracut/dracut.conf.d/70-chreipl-fcp-mpath.conf
$(INSTALL_DATA) -t $(DESTDIR)$(DRACUTCONFDIR) \
dracut/dracut.conf.d/70-chreipl-fcp-mpath.conf
chreipl-fcp-mpath-install: chreipl-fcp-mpath-install-dracut-config
endif
ifeq ($(ENABLE_DOC),1)
# chreipl-fcp-mpath: install man page
INSTDIRS += $(MANDIR)
.PHONY: chreipl-fcp-mpath-install-man-page
chreipl-fcp-mpath-install-man-page: | $(DESTDIR)$(MANDIR)/man7
chreipl-fcp-mpath-install-man-page: chreipl-fcp-mpath.7
$(GZIP) -fk --best chreipl-fcp-mpath.7
$(INSTALL_DATA) -t $(DESTDIR)$(MANDIR)/man7 \
chreipl-fcp-mpath.7.gz
chreipl-fcp-mpath-install: chreipl-fcp-mpath-install-man-page
.PHONY: chreipl-fcp-mpath-install-man-page-clean
chreipl-fcp-mpath-install-man-page-clean:
rm -f chreipl-fcp-mpath.7.gz
chreipl-fcp-mpath-clean: chreipl-fcp-mpath-install-man-page-clean
endif
#
## Utility
#
# Utilities for the debug feature of chreipl-fcp-mpath-common.sh.
#
# When `chreipl-fcp-mpath` is built with D=1 (default is D=0), each run
# of one of the helper scripts will create a debug log in $(DEBUGOUTDIR)
# (default: /run/udev) which among other things contains the complete shell
# trace of that script run, with some added information that would not be
# inspectable otherwise with just the trace.
#
# There is currently no way of enabling/disabling this at runtime.
#
# chreipl-fcp-mpath-common.sh defines debug log file name as
# "chreiplzfcpmp-${debug_trace_tag}-${SEQNUM:-0}.XXXXXXXXXX" where SEQNUM is a
# udev rule environment variable and each X is replaced with some [[:alnum:]]
# by mktemp. For a definition of `debug_trace_tag`, please see the comments
# in the source.
#
# The following targets can be used for some simple access and filtering of the
# logs during development.
DEBUG_LOG_GLOB = $(DEBUGOUTDIR)/chreiplzfcpmp-[[:digit:]][[:digit:]][[:alpha:]][[:alpha:]][[:alpha:]][[:alpha:]]-+([[:digit:]]).[[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]]
# display all debug log files on the system
.PHONY: chreipl-fcp-mpath-debug-logs
chreipl-fcp-mpath-debug-logs:
@ls -1d $(DEBUG_LOG_GLOB)
# display only debug log files of script runs that exited with status 0 (= good)
#
# XXX: `sed -n -e '$p'` is used instead of `tail -n1` to prevent an other
# dependency just for that (both invocations do the same thing).
.PHONY: chreipl-fcp-mpath-debug-logs-filter-good
chreipl-fcp-mpath-debug-logs-filter-good:
@for lg in $(DEBUG_LOG_GLOB); do \
sed -e '/^+ trap_exit$$/,/^+ trap - EXIT$$/d' "$${lg}" \
| sed -n -e '$$p' \
| grep -q -e '^+ exit 0$$' || continue; \
ls -d "$${lg}"; \
done
# display only debug log files of script runs that didn't exit with status 0
# (= bad)
.PHONY: chreipl-fcp-mpath-debug-logs-filter-bad
chreipl-fcp-mpath-debug-logs-filter-bad:
@for lg in $(DEBUG_LOG_GLOB); do \
sed -e '/^+ trap_exit$$/,/^+ trap - EXIT$$/d' "$${lg}" \
| sed -n -e '$$p' \
| grep -q -e '^+ exit 0$$' && continue; \
ls -d "$${lg}"; \
done
.PHONY: chreipl-fcp-mpath-debug-logs-clean
chreipl-fcp-mpath-debug-logs-clean:
rm -f $(DEBUG_LOG_GLOB)

269
chreipl-fcp-mpath/README.md Normal file
View File

@@ -0,0 +1,269 @@
<!-- markdown documentation: https://github.github.com/gfm/ -->
NAME
====
chreipl-fcp-mpath - use multipath information for re-IPL path failover on a
running Linux instance
DESCRIPTION
===========
The IPL process of Linux on Z or LinuxONE from an FCP-attached SCSI volume uses
exactly one path to the volume. If this path is unavailable, the IPL fails.
The **chreipl-fcp-mpath** toolset monitors **udev** events about paths to the
re-IPL volume. If the currently configured re-IPL path becomes unavailable, the
toolset checks for operational paths to the same volume. If available, it
reconfigures the re-IPL settings to use an operational path.
Thus, re-IPL from an FCP-attached SCSI volume can be successful despite path
failures on a running Linux instance if at least one path to the re-IPL volume
remains operational.
**Chreipl-fcp-mpath** requires **udev**, **multipathd** and **dm-multipath**.
Once installed, the toolset runs automatically and autonomously. No user
intervention is possible or required.
Other than installing the toolset, there is no user interface for
**chreipl-fcp-mpath**.
Requirements
------------
The **chreipl-fcp-mpath** tool has the following requirements on the
Linux instance that is being monitored:
- The Linux instance must have started successfully, during IPL.
- The running Linux instance must use **dm-multipath** and **multipathd** for
the configured re-IPL volume - a volume that contains a zipl boot record
and has one of its paths used in the re-IPL configuration.
- **udev** must run.
- The toolset must observe at least one event about the configured re-IPL
path. Examples for such events are: the SCSI disk comes online, or a path
of the corresponding multipath device goes down or comes back online.
- The WWID of the re-IPL volume must not change while the Linux instance is
running.
- When the configured re-IPL path becomes unavailable while the Linux
instance is running, at least one operational path to the re-IPL volume
must be available, or must become available. If no such path is available
when the Linux instance is rebooted, the re-IPL path is not changed.
- The tool assumes that any manually reconfigured re-IPL device is valid and
operational.
The tool treats a newly configured re-IPL device like the initially
configured re-IPL device. In particular, if the newly configured re-IPL
device fulfills the requirements of the tool, re-IPL path failover takes
place if the configured re-IPL path becomes unavailable.
Caution with Manual Changes to the Configured re-IPL Target
-----------------------------------------------------------
**chreipl-fcp-mpath** is designed to accept operator-inititated changes of the
re-IPL device. However, concurrent changes by the operator and tool driven
changes can result in the operator change being overwritten.
To avoid this problem, change the re-IPL device only during steady-state
operations, when no path events happen. Alternatively, make sure that no events
are processed while you change the device. See [EXAMPLES](#examples) for one
way to suspend event processing.
MESSAGES
========
During monitoring and event processing, **chreipl-fcp-mpath** writes messages
to the syslog.
When the configured re-IPL path is changed to a different path to the same
volume (priority *daemon.notice*):
> Changed re-IPL path to: \<device-bus-id\>:\<wwpn\>:\<lun\>.
When a path event indicates that the last available path has become
non-operational (priority *daemon.alert*):
> The re-IPL device cannot be changed because no operational path to the
> re-IPL volume remains. The next re-IPL might fail unless you re-attach or
> enable at least one valid path to the re-IPL volume.
When changing the configured re-IPL device failed because of an error with the
used Linux kernel interface (priority *daemon.crit*):
> Changing the re-IPL device failed. The current re-IPL settings might be
> inconsistent. Check and correct the settings (see the README.md of
> chreipl-fcp-mpath) to make sure that the current re-IPL device is valid.
A failure to change the re-IPL device can indicate an inconsistent setting that
cannot be corrected automatically by **chreipl-fcp-mpath**. As a result, the
next re-IPL might fail or might not use the intended re-IPL device.
You can use the following tools to check and correct the current settings:
- **lsreipl** to confirm that the intended re-IPL device is configured;
- **chreipl** to change the re-IPL device;
- **lszfcp** to inspect the state of available paths to the re-IPL device.
<!-- NOT-IN-MAN { -->
SOFTWARE REQUIREMENTS
=====================
**chreipl-fcp-mpath** integrates into s390-tools's build and install
infrastructure. Use **make** to build it. No explicit dependency management is
in place, but the toolset has some software dependencies besides the
requirements in section [Requirements](#requirements):
- GNU Bash;
- GNU Core Utilities (mktemp, readlink, sync, truncate);
- util-linux (flock, hexdump, logger);
- udev / systemd-udev;
- multipath-tools.
To make use of the optional dracut configuration you need: dracut.
To build and install the documentation (man page) you need:
- pandoc;
- GNU Core Utilities (date);
- GNU awk;
- GNU Gzip.
INSTALLATION
============
If your distribution includes a packaged version of **chreipl-fcp-mpath**,
either as a separate package or as part of a **s390-tools** package, install
that package. Otherwise, you can either install it from source as part of
**s390-tools** or separately.
To install **chreipl-fcp-mpath** as part of **s390-tools**, use **make** on the
top-level directory of your **s390-tools** distribution. Installing the entire
distribution might overwrite other already installed tools.
To install the tool separately, change into the **chreipl-fcp-mpath**
directory, and use **make** there.
You need *root* privileges to install the tool into the root file system.
Calling **make** runs the build steps. Calling **make install** runs the build
steps and copies the resulting components to their final destination.
**s390-tools** offers more options and targets to customize the build (see
**make help**).
**chreipl-fcp-mpath** has the following optional build options:
| Option | Values | Default | Effect
| :----- | :----: | :-----: | :-----
| HAVE_DRACUT | 0, 1 | 0 | Install a dracut configuration file that includes **chreipl-fcp-mpath** in the initial ramdisks built with **dracut**.
| ENABLE_DOC | 0, 1 | 0 | Build and install a man page for **chreipl-fcp-mpath**.
Specify any options as arguments for both the **make** and **make install**
command as shown in the following example:
~ # cd chreipl-fcp-mpath/
~ # make HAVE_DRACUT=1 ENABLE_DOC=1
~ # make HAVE_DRACUT=1 ENABLE_DOC=1 install
After the installation, reload the udev rules database:
~ # udevadm control --reload
*The toolset is now active on your running Linux instance.*
If you use the *HAVE_DRACUT=1* option, also rebuild your
initial ramdisk, to immediately include the toolset instead of
waiting for the next kernel update.
How to rebuild the initial ramdisk and the naming scheme for the
resulting file or files depends on your distribution.
The following example applies to Fedora and to Red Hat Enterprise Linux:
~ # dracut --force /boot/initramfs-"$(uname -r)".img "$(uname -r)"
For SUSE Linux Enterprise Server run for example:
~ # dracut --hostonly --force /boot/initrd-"$(uname -r)" "$(uname -r)"
These commands replace the initial ramdisk for the currently running kernel.
If your distribution uses **zipl** as its boot loader, run **zipl** to refresh
the boot record to find the new initial ramdisk.
~ # zipl
With dracut and documentation enabled, **make install** deploys the following
files to these default locations:
/usr/lib/chreipl-fcp-mpath/chreipl-fcp-mpath-common.sh
/usr/lib/dracut/dracut.conf.d/70-chreipl-fcp-mpath.conf
/usr/lib/udev/chreipl-fcp-mpath-is-ipl-tgt
/usr/lib/udev/chreipl-fcp-mpath-is-ipl-vol
/usr/lib/udev/chreipl-fcp-mpath-is-reipl-zfcp
/usr/lib/udev/chreipl-fcp-mpath-record-volume-identifier
/usr/lib/udev/chreipl-fcp-mpath-try-change-ipl-path
/usr/lib/udev/rules.d/70-chreipl-fcp-mpath.rules
/usr/share/man/man7/chreipl-fcp-mpath.7.gz
UNINSTALL
=========
If your distribution includes a separately from **s390-tools** packaged version
of **chreipl-fcp-mpath**, uninstall that package.
For installations without distribution packaging, you cannot uninstall
**chreipl-fcp-mpath** with **make**.
Instead, remove the toolset by deleting the installed files as listed in
[INSTALLATION](#installation)), reload the udev rules database, and rebuild all
modified initial ramdisks as described in [INSTALLATION](#installation)).
<!-- NOT-IN-MAN } -->
EXAMPLES
========
Manual Changes to the Configured re-IPL Device
----------------------------------------------
As outlined in [DESCRIPTION](#description), be cautious when manually changing
the configured re-IPL device. Assure that your reconfiguration actions do not
collide with concurrent automatic event processing by **chreipl-fcp-mpath**.
You can avoid such collisions, by stopping event processing, making your
changes, and then re-enabling event processing. You need *root* privileges for
running the commands in the following example:
~ # udevadm settle
~ # udevadm control --stop-exec-queue
~ # chreipl ...
~ # udevadm control --start-exec-queue
Listing messages with journalctl
--------------------------------
If your Linux instance includes **journalctl**, use the following command to
list all messages that are issued by **chreipl-fcp-mpath**:
~ # journalctl -t chreipl-fcp-mpath
To list only messages that were issued since the last IPL, use this command:
~ # journalctl -t chreipl-fcp-mpath -b
REPORTING BUGS
==============
Use the **Issues** functionality on GitHub to report any bugs in
**chreipl-fcp-mpath**:
[s390-tools Issues](<https://github.com/ibm-s390-linux/s390-tools/issues> "Link to the s390-tools Issues page").
SEE ALSO
========
**chreipl**(8), **dracut**(8), **journalctl**(1), **lsreipl(8)**,
**lszfcp**(8), **multipath**(8), **multipathd**(8), **udev**(7),
**udevadm**(8), **zipl**(8)

View File

@@ -0,0 +1,447 @@
#!/bin/bash
# SPDX-License-Identifier: MIT
#
# chreipl-fcp-mpath: use multipath information to change FCP IPL target
# (C) Copyright IBM Corp. 2021
#
# Uses the following system-utilities (and shell-builtins):
# GNU coreutils:
# - mktemp
# - readlink
# - sync
# util-linux:
# - flock
# - logger
# Makes use of udev event environment variables:
# SEQNUM
# (1) expand failed globs to an empty string
# (2) extended pattern matching to strip leading/trailing whitespaces
shopt -s nullglob extglob
# (1) don't overwrite existing files using redirects (e.g.: `>`)
set -o noclobber
# make sure any state files created are only writeable by the owning user
umask 027
# create log if DEBUG is enabled (with Make: D=1)
#
# Each script importing this library and expecting a debug log to be created
# must declare a *trace tag* in a variable `debug_trace_tag`. This is used as
# identifier in the log file name. The format is:
#
# [[:digit:]][[:digit:]][[:alpha:]][[:alpha:]][[:alpha:]][[:alpha:]]
# \ /\ /
# --------\ /-------- -------------------\ /-------------------
# \/ \/
# relative position of some unique abbreviation for the script
# execution in the name, excluding any common prefix
# udev rules
if '@DEBUG@' && [ -v debug_trace_tag ] && tlg="$(
mktemp -p '@debugoutdir@' \
"chreiplzfcpmp-${debug_trace_tag}-${SEQNUM:-0}.XXXXXXXXXX" \
2>/dev/null)"
then
readonly tlg
exec >|"${tlg}" 2>&1
set -x
set
else
unset tlg
fi
declare -gr ID_FILE='@chreiplzfcpmp-id-file@'
declare -gr FW_LOCK_FILE='@chreiplzfcpmp-fwlock-file@'
declare -gA TRAP_EXIT_FN=()
declare -gf trap_exit 1>/dev/null
function trap_exit() {
local fn
for fn in "${TRAP_EXIT_FN[@]}"; do
"${fn}"
done
trap - EXIT
}
trap trap_exit EXIT
# Output variables:
# id_file_unlock_exclusive_create() - call to unlock when finished with
# critical section
#
# XXX: `id_file_lock_*` can't be taken recursively
function id_file_lock_exclusive_create() {
declare -g ID_FILE_LOCK=""
# prevent concurrent file creation
#
# First, open the file defined in ${ID_FILE} for writing; this will
# succeed and create the file only if it doesn't exist already. If the
# file already exist, the first open attempt will fail and we fall
# back to opening it only for reading; this will always succeed if the
# file already exists (the reason why the first attempty failed). In
# both cases store the corresponding file descriptor in
# ${ID_FILE_LOCK}.
#
# XXX: This should be race free.
# open() with O_EXCL... is atomic (we set `noclobber` as shell
# option); at least as long as we talk about a local FS.
if ! { exec {ID_FILE_LOCK}>"${ID_FILE}"; } 2>/dev/null; then
{ exec {ID_FILE_LOCK}<"${ID_FILE}"; } 2>/dev/null \
|| return 1
fi
declare -gf id_file_unlock_exclusive_create 1>/dev/null
function id_file_unlock_exclusive_create() {
if [ -v ID_FILE_LOCK ]; then
sync "${ID_FILE}" 2>/dev/null
# release file and implicitly the lock, if taken
exec {ID_FILE_LOCK}>&-
unset ID_FILE_LOCK
fi
unset "TRAP_EXIT_FN[id_file_unlock_exclusive_create]"
}
TRAP_EXIT_FN+=(
[id_file_unlock_exclusive_create]=id_file_unlock_exclusive_create
)
flock --exclusive --timeout 5 "${ID_FILE_LOCK}" || return 2
return 0
}
# Output variables:
# id_file_unlock_exclusive_no_create() - call to unlock when finished with
# critical section
#
# XXX: `id_file_lock_*` can't be taken recursively
function id_file_lock_exclusive_no_create() {
declare -g ID_FILE_LOCK=""
# Open the file defined in ${ID_FILE} for reading, and store the
# corresponding file descriptor in ${ID_FILE_LOCK}.
#
# XXX: return code is used in `chreipl-fcp-mpath-try-change-ipl-path`
{ exec {ID_FILE_LOCK}<"${ID_FILE}"; } 2>/dev/null || return 1
declare -gf id_file_unlock_exclusive_no_create 1>/dev/null
function id_file_unlock_exclusive_no_create() {
if [ -v ID_FILE_LOCK ]; then
sync "${ID_FILE}" 2>/dev/null
# release file and implicitly the lock, if taken
exec {ID_FILE_LOCK}<&-
unset ID_FILE_LOCK
fi
unset "TRAP_EXIT_FN[id_file_unlock_exclusive_no_create]"
}
TRAP_EXIT_FN+=(
[id_file_unlock_exclusive_no_create]=id_file_unlock_exclusive_no_create
)
flock --exclusive --timeout 5 "${ID_FILE_LOCK}" || return 2
return 0
}
# Output variables:
# id_file_unlock_shared_no_create() - call to unlock when finished with
# critical section
#
# XXX: `id_file_lock_*` can't be taken recursively
function id_file_lock_shared_no_create() {
declare -g ID_FILE_LOCK=""
# Open the file defined in ${ID_FILE} for reading, and store the
# corresponding file descriptor in ${ID_FILE_LOCK}.
{ exec {ID_FILE_LOCK}<"${ID_FILE}"; } 2>/dev/null || return 1
declare -gf id_file_unlock_shared_no_create 1>/dev/null
function id_file_unlock_shared_no_create() {
if [ -v ID_FILE_LOCK ]; then
# release file and implicitly the lock, if taken
exec {ID_FILE_LOCK}<&-
unset ID_FILE_LOCK
fi
unset "TRAP_EXIT_FN[id_file_unlock_shared_no_create]"
}
TRAP_EXIT_FN+=(
[id_file_unlock_shared_no_create]=id_file_unlock_shared_no_create
)
flock --shared --timeout 5 "${ID_FILE_LOCK}" || return 2
return 0
}
# Output variables:
# firmware_unlock_exclusive() - call to unlock when finished with critical section
#
# XXX: `firmware_lock_*` can't be taken recursively
function firmware_lock_exclusive() {
declare -g FIRMWARE_LOCK=""
# Open the file defined in ${FW_LOCK_FILE} for reading, and store the
# corresponding file descriptor in ${FIRMWARE_LOCK} (it doesn't matter
# whether this is a normal file or directory). This file descriptor
# will only be used for locking - not for actual I/O.
{ exec {FIRMWARE_LOCK}<"${FW_LOCK_FILE}"; } 2>/dev/null || return 1
declare -gf firmware_unlock_exclusive 1>/dev/null
function firmware_unlock_exclusive() {
if [ -v FIRMWARE_LOCK ]; then
# release file and implicitly the lock, if taken
exec {FIRMWARE_LOCK}<&-
unset FIRMWARE_LOCK
fi
unset "TRAP_EXIT_FN[firmware_unlock_exclusive]"
}
TRAP_EXIT_FN+=([firmware_unlock_exclusive]=firmware_unlock_exclusive)
flock --exclusive --timeout 5 "${FIRMWARE_LOCK}" || return 2
return 0
}
# Output variables:
# firmware_unlock_shared() - call to unlock when finished with critical section
#
# XXX: `firmware_lock_*` can't be taken recursively
function firmware_lock_shared() {
declare -g FIRMWARE_LOCK=""
# Open the file defined in ${FW_LOCK_FILE} for reading, and store the
# corresponding file descriptor in ${FIRMWARE_LOCK} (it doesn't matter
# whether this is a normal file or directory). This file descriptor
# will only be used for locking - not for actual I/O.
{ exec {FIRMWARE_LOCK}<"${FW_LOCK_FILE}"; } 2>/dev/null || return 1
declare -gf firmware_unlock_shared 1>/dev/null
function firmware_unlock_shared() {
if [ -v FIRMWARE_LOCK ]; then
# release file and implicitly the lock, if taken
exec {FIRMWARE_LOCK}<&-
unset FIRMWARE_LOCK
fi
unset "TRAP_EXIT_FN[firmware_unlock_shared]"
}
TRAP_EXIT_FN+=([firmware_unlock_shared]=firmware_unlock_shared)
flock --shared --timeout 5 "${FIRMWARE_LOCK}" || return 2
return 0
}
# Output variables:
# IPL_TYPE
# IPL_BUSID
# IPL_WWPN
# IPL_LUN
function firmware_get_ipl_information() {
declare -g IPL_TYPE="" IPL_BUSID="" IPL_WWPN="" IPL_LUN=""
# Take lock so we don't see any intermediate state from other helpers
# running in parallel
firmware_lock_shared || return 5
{ read -r IPL_TYPE _ < /sys/firmware/reipl/reipl_type; } 2>/dev/null \
|| return 1
{ read -r IPL_BUSID _ < /sys/firmware/reipl/fcp/device; } 2>/dev/null \
|| return 2
{ read -r IPL_WWPN _ < /sys/firmware/reipl/fcp/wwpn; } 2>/dev/null \
|| return 3
{ read -r IPL_LUN _ < /sys/firmware/reipl/fcp/lun; } 2>/dev/null \
|| return 4
firmware_unlock_shared
# show read values in debug log if enabled
if '@DEBUG@'; then
declare -p IPL_TYPE IPL_BUSID IPL_WWPN IPL_LUN 1>&2
fi
return 0
}
# Input:
# 1: absolute canonical path to the scsi device in sysfs, e.g.:
# /sys/devices/css0/0.0.0014/0.0.1700/host1/rport-1:0-0/target1:0:0/1:0:0:1075789848
# Output variables:
# SDEV_LUN
function sdev_get_lun() {
local sdev="${1}" sdev_lun_str
# bash uses `intmax_t` as width for integer variables, and glibc
# defines this either as `long int` on 64 bit systems, or
# `long long int` on other.
local -i sdev_lun=0 fcp_lun=0
sdev_lun_str="${sdev##*:}"
# e.g.: 1075789848
[[ "${sdev_lun_str}" == +([[:digit:]]) ]] || return 1
# "cast" to integer
sdev_lun="${sdev_lun_str}"
# convert the Linux integer LUN format to the hexadecimal 64 bit T10
# LUN representation format used by many s390x interfaces
(( fcp_lun = (((sdev_lun >> 0) & 0xffff) << 48)
| (((sdev_lun >> 16) & 0xffff) << 32)
| (((sdev_lun >> 32) & 0xffff) << 16)
| (((sdev_lun >> 48) & 0xffff) << 0) ))
# the '0x' prefix is part of the length
printf -v SDEV_LUN "%#018llx" "${fcp_lun}"
# show read values in debug log if enabled
if '@DEBUG@'; then
declare -p SDEV_LUN 1>&2
fi
return 0
}
# Input:
# 1: path to the scsi device in sysfs, e.g.:
# /sys/devices/css0/0.0.0014/0.0.1700/host1/rport-1:0-0/target1:0:0/1:0:0:1075789848
# , or a symlink pointing to the scsi device, e.g.:
# /sys/class/block/sds/device
# Output variables:
# SDEV_BUSID
# SDEV_WWPN
# SDEV_LUN
function sdev_get_fcp_addressing() {
local sdev="${1}" fcp_lun rport rport_wwpn zfcp_dev
declare -g SDEV_BUSID="" SDEV_WWPN="" SDEV_LUN=""
sdev="$(readlink -se "${sdev}")" || return 1
# get the LUN for this SDEV
#
# sets ${SDEV_LUN}
sdev_get_lun "${sdev}" || return 2
# get the WWPN of the remote port this SDEV is attached to
printf -v rport "%s" "${sdev}"/../../fc_remote_ports/rport-*:*-*
# e.g.: /sys/devices/css0/0.0.0016/0.0.1740/host0/rport-0:0-1/fc_remote_ports/rport-0:0-1
[ "${rport}" != "" ] || return 3
# XXX: This works even if the rport is currently in a bad
# state, so e.g. when it has just gone down because of a
# cable pull.
{ read -r rport_wwpn _ < "${rport}"/port_name; } 2>/dev/null \
|| return 4
# The Linux kernel doesn't guarantee the same format as in
# /sys/firmware/..., so make sure it is the one we expect.
[[ "${rport_wwpn}" =~ ^0x[[:xdigit:]]{1,16}$ ]] || return 5
# the '0x' prefix is part of the length
printf -v rport_wwpn "%#018llx" "${rport_wwpn}"
# get the Device Bus-ID of the device via which this SDEV is attached
zfcp_dev="$(readlink -se "${sdev}"/../../../..)" || return 6
# e.g.: /sys/devices/css0/0.0.0016/0.0.1740
zfcp_dev="${zfcp_dev##*/}"
# shellcheck disable=2034
SDEV_BUSID="${zfcp_dev}"
# shellcheck disable=2034
SDEV_WWPN="${rport_wwpn}"
return 0
}
# Input:
# 1: path to the scsi device in sysfs
# Output variables:
# SDEV_WWID
function sdev_get_wwid() {
local sdev="${1}"
local -a wwid
declare -g SDEV_WWID=""
# read the volume identifier without stripping any content
#
# XXX: we can read the WWID file, even if the SDEV is currently not
# operational (e.g.: due to the path has gone away), as long as
# the VPD PG 83 is still cached in the kernel; and the page gets
# only released on SDEV device release.
{ readarray -d "" -t wwid < "${sdev}"/wwid; } 2>/dev/null \
|| return 1
if '@DEBUG@'; then declare -p wwid 1>&2; fi
# test whether we read something
#
# This strips all leading spaces from the beginning of the read WWID
# (until the first non-space or NUL character), and checks whether the
# result is empty.
# Hence, we return early if the WWID consists of only whitespace.
#
# XXX: there could be unexpected characters in the returned ID.
# `scsi_id` from the udev helpers sanitizes the strings it reads
# from the devices, so they can be used in environment variables
# without much danger.
# But we don't export anything here, so it should be fine.
[ "${wwid[0]/#*([[:space:]])}" != "" ] || return 2
# shellcheck disable=2034
SDEV_WWID="${wwid[0]}"
return 0
}
# Input:
# 1: path to the scsi device in sysfs
# Return Value:
# == 0: SDEV referenced by `1` in good state
# != 0: otherwise
function sdev_test_path_state() {
local sdev="${1}" state zfcp_failed zfcp_in_recovery rport port_state
sdev="$(readlink -se "${sdev}")" || return 1
{ read -r state _ < "${sdev}"/state; } 2>/dev/null || return 2
{ read -r zfcp_failed _ < "${sdev}"/zfcp_failed; } 2>/dev/null \
|| return 3
{ read -r zfcp_in_recovery _ < "${sdev}"/zfcp_in_recovery; } 2>/dev/null \
|| return 4
printf -v rport "%s" "${sdev}"/../../fc_remote_ports/rport-*:*-*
# e.g.: /sys/devices/css0/0.0.0016/0.0.1740/host0/rport-0:0-1/fc_remote_ports/rport-0:0-1
[ "${rport}" != "" ] || return 5
{ read -r port_state _ < "${rport}"/port_state; } 2>/dev/null \
|| return 6
if '@DEBUG@'; then
declare -p state zfcp_failed zfcp_in_recovery port_state 1>&2
fi
[ "${state}" = "running" ] || return 7
[ "${zfcp_failed}" = "0" ] || return 8
[ "${zfcp_in_recovery}" = "0" ] || return 9
{ [ "${port_state}" = "Online" ] \
|| [ "${port_state}" = "Marginal" ]; } || return 10
return 0
}
# Input:
# *: all input parameters are used as quoted message
function log_note() {
logger -p 'daemon.notice' -t 'chreipl-fcp-mpath' "${*}" &>/dev/null
}
# Input:
# *: all input parameters are used as quoted message
function log_crit() {
logger -p 'daemon.crit' -t 'chreipl-fcp-mpath' "${*}" &>/dev/null
}
# Input:
# *: all input parameters are used as quoted message
function log_alert() {
logger -p 'daemon.alert' -t 'chreipl-fcp-mpath' "${*}" &>/dev/null
}

View File

@@ -0,0 +1,56 @@
#!/bin/bash
# SPDX-License-Identifier: MIT
#
# chreipl-fcp-mpath: use multipath information to change FCP IPL target
# (C) Copyright IBM Corp. 2021
#
# Uses the following system-utilities (and shell-builtins):
# Those necessary for sourced library:
# - chreipl-fcp-mpath-common.sh
# Find out whether the device in udev event environment variable ${DEVPATH}
# represents the device we want to re-IPL from. We do this by comparing
# Device-Bus-ID/Target-WWPN/LUN of the individual SDEVs to the parameters set
# in `/sys/firmware/reipl/fcp/`.
#
# Makes use of udev event environment variables:
# DM_UUID
# SUBSYSTEM
# DEVPATH
# shellcheck disable=SC2034
declare -gr debug_trace_tag=05iilt
# shellcheck disable=SC1091
source '@chreiplzfcpmp-lib@' || exit 127
firmware_get_ipl_information || exit 1
if [[ "${DM_UUID}" == mpath-* ]]; then
# Assume Multipath Device Mapper Device;
# e.g.: DEVPATH = /devices/virtual/block/dm-0
declare sdev
# depends on `nullglob` from `chreipl-fcp-mpath-common.sh`
for sdev in /sys/"${DEVPATH}"/slaves/sd*/device; do
sdev_get_fcp_addressing "${sdev}" || continue
[ "${SDEV_LUN}" = "${IPL_LUN}" ] || continue
[ "${SDEV_WWPN}" = "${IPL_WWPN}" ] || continue
[ "${SDEV_BUSID}" = "${IPL_BUSID}" ] || continue
exit 0
done
unset sdev
elif [ "${SUBSYSTEM}" = block ]; then
# Assume SCSI Disk;
# e.g.: DEVPATH = /devices/css0/0.0.0014/0.0.1700/host0/rport-0:0-0/target0:0:0/0:0:0:1074806808/block/sds
sdev_get_fcp_addressing /sys/"${DEVPATH}"/device || exit 2
[ "${SDEV_LUN}" = "${IPL_LUN}" ] || exit 3
[ "${SDEV_WWPN}" = "${IPL_WWPN}" ] || exit 4
[ "${SDEV_BUSID}" = "${IPL_BUSID}" ] || exit 5
exit 0
fi
exit 6

View File

@@ -0,0 +1,86 @@
#!/bin/bash
# SPDX-License-Identifier: MIT
#
# chreipl-fcp-mpath: use multipath information to change FCP IPL target
# (C) Copyright IBM Corp. 2021
#
# Uses the following system-utilities (and shell-builtins):
# Those necessary for sourced library:
# - chreipl-fcp-mpath-common.sh
# Find out whether the device in environment variable ${DEVPATH} represents the
# _volume_ that we IPL'ed from. We do this by comparing its WWID to the one
# recorded in `@chreiplzfcpmp-id-file@`.
#
# Makes use of udev event environment variables:
# DM_UUID
# SUBSYSTEM
# DEVPATH
# shellcheck disable=SC2034
declare -gr debug_trace_tag=11iilv
# shellcheck disable=SC1091
source '@chreiplzfcpmp-lib@' || exit 127
function id_file_read_ipl_information() {
local -a records
declare -g REC_WWID="" REC_BUSID="" REC_WWPN="" REC_LUN=""
# lock file before reading ID, so we don't see any intermediate state
id_file_lock_shared_no_create || return 1
{ readarray -d "" -t -u "${ID_FILE_LOCK}" records; } 2>/dev/null \
|| return 2
if '@DEBUG@'; then declare -p records 1>&2; fi
id_file_unlock_shared_no_create
[ "${#records[@]}" = "4" ] || return 3
# check that none of the array fields contains whitespace only
[ "${records[0]/#*([[:space:]])}" != "" ] || return 4
[ "${records[1]/#*([[:space:]])}" != "" ] || return 5
[ "${records[2]/#*([[:space:]])}" != "" ] || return 6
[ "${records[3]/#*([[:space:]])}" != "" ] || return 7
REC_WWID="${records[0]}"
REC_BUSID="${records[1]}"
REC_WWPN="${records[2]}"
REC_LUN="${records[3]}"
return 0
}
id_file_read_ipl_information || exit 1
if [[ "${DM_UUID}" == mpath-* ]]; then
# Assume Multipath Device Mapper Device;
# e.g.: DEVPATH = /devices/virtual/block/dm-0
declare sdev found=false
for sdev in /sys/"${DEVPATH}"/slaves/sd*/device; do
if sdev_get_wwid "${sdev}"; then
found=true
break
fi
done
unset sdev
"${found}" || exit 2
elif [ "${SUBSYSTEM}" = block ]; then
# Assume SCSI Disk;
# e.g.: DEVPATH = /devices/css0/0.0.0014/0.0.1700/host0/rport-0:0-0/target0:0:0/0:0:0:1074806808/block/sds
sdev_get_wwid /sys/"${DEVPATH}"/device || exit 3
fi
# set by `sdev_get_wwid` and `id_file_read_ipl_information`
[ "${SDEV_WWID}" = "${REC_WWID}" ] || exit 4
firmware_get_ipl_information || exit 5
# set by `firmware_get_ipl_information` and `id_file_read_ipl_information`
[ "${IPL_BUSID}" = "${REC_BUSID}" ] || exit 6
[ "${IPL_WWPN}" = "${REC_WWPN}" ] || exit 7
[ "${IPL_LUN}" = "${REC_LUN}" ] || exit 8
exit 0

View File

@@ -0,0 +1,25 @@
#!/bin/bash
# SPDX-License-Identifier: MIT
#
# chreipl-fcp-mpath: use multipath information to change FCP IPL target
# (C) Copyright IBM Corp. 2021
#
# Uses the following system-utilities (and shell-builtins):
# Those necessary for sourced library:
# - chreipl-fcp-mpath-common.sh
# Find out whether ReIPL is gonna happen from a SCSI volume attached via zFCP
# shellcheck disable=SC2034
declare -gr debug_trace_tag=00iriz
# shellcheck disable=SC1091
source '@chreiplzfcpmp-lib@' || exit 127
declare reipl_type
{ read -r reipl_type _ < /sys/firmware/reipl/reipl_type; } 2>/dev/null || exit 1
if '@DEBUG@'; then declare -p reipl_type 1>&2; fi
[ "${reipl_type}" = "fcp" ] || exit 2
exit 0

View File

@@ -0,0 +1,73 @@
#!/bin/bash
# SPDX-License-Identifier: MIT
#
# chreipl-fcp-mpath: use multipath information to change FCP IPL target
# (C) Copyright IBM Corp. 2021
#
# Uses the following system-utilities (and shell-builtins):
# Those necessary for sourced library:
# - chreipl-fcp-mpath-common.sh
# GNU coreutils:
# - truncate
# util-linux:
# - hexdump
# Record the identification of the volume we want to re-IPL from
#
# Makes use of udev event environment variables:
# DM_UUID
# SUBSYSTEM
# DEVPATH
# shellcheck disable=SC2034
declare -gr debug_trace_tag=10rvid
# shellcheck disable=SC1091
source '@chreiplzfcpmp-lib@' || exit 127
function id_file_record_ipl_information() {
local sdev_wwid="${1}" ipl_busid="${2}" ipl_wwpn="${3}" ipl_lun="${4}"
# lock file before writing ID, so noone sees any intermediate state
id_file_lock_exclusive_create || return 1
# reset ID without removing the file (necessary for the locking to work
# properly, since the FD we use for locking is on this file/inode)
truncate --no-create --size=0 "${ID_FILE}" || return 3
echo -ne "${sdev_wwid}\x00${ipl_busid}\x00${ipl_wwpn}\x00${ipl_lun}\x00" \
>>"${ID_FILE}" || return 4
if '@DEBUG@'; then hexdump -vC "${ID_FILE}" 1>&2; fi
id_file_unlock_exclusive_create
return 0
}
if [[ "${DM_UUID}" == mpath-* ]]; then
# Assume Multipath Device Mapper Device;
# e.g.: DEVPATH = /devices/virtual/block/dm-0
declare sdev
for sdev in /sys/"${DEVPATH}"/slaves/sd*/device; do
if sdev_get_wwid "${sdev}"; then
break
fi
done
unset sdev
elif [ "${SUBSYSTEM}" = block ]; then
# Assume SCSI Disk;
# e.g.: DEVPATH = /devices/css0/0.0.0014/0.0.1700/host0/rport-0:0-0/target0:0:0/0:0:0:1074806808/block/sds
sdev_get_wwid /sys/"${DEVPATH}"/device
fi
# shellcheck disable=SC2153
[ "${SDEV_WWID}" != "" ] || exit 1
firmware_get_ipl_information || exit 2
# shellcheck disable=SC2153
id_file_record_ipl_information \
"${SDEV_WWID}" "${IPL_BUSID}" "${IPL_WWPN}" "${IPL_LUN}" \
|| exit 3
exit 0

View File

@@ -0,0 +1,158 @@
#!/bin/bash
# SPDX-License-Identifier: MIT
#
# chreipl-fcp-mpath: use multipath information to change FCP IPL target
# (C) Copyright IBM Corp. 2021
#
# Uses the following system-utilities (and shell-builtins):
# Those necessary for sourced library:
# - chreipl-fcp-mpath-common.sh
# GNU coreutils:
# - truncate
# util-linux:
# - hexdump
# Try to change the current re-IPL target to a dfferent, operational path to
# the same volume.
#
# Makes use of udev event environment variables:
# DM_UUID
# SUBSYSTEM
# DEVPATH
# CHREIPL_FCP_MPATH_IS_TGT
# shellcheck disable=SC2034
declare -gr debug_trace_tag=15tcip
# shellcheck disable=SC1091
source '@chreiplzfcpmp-lib@' || exit 127
function apply_ipl_information() {
local sdev_wwid="${1}"
local sdev_busid="${2}" sdev_wwpn="${3}" sdev_lun="${4}"
local ipl_type="${5}" ipl_busid="${6}" ipl_wwpn="${7}" ipl_lun="${8}"
local -a records
local try_update_id_file=true
[ "${ipl_type}" = "fcp" ] || return 1
[[ "${sdev_busid}" =~ ^[[:xdigit:]]{1,3}\.[[:xdigit:]]\.[[:xdigit:]]{1,4}$ ]] \
|| return 2
[[ "${sdev_wwpn}" =~ ^0x[[:xdigit:]]{16}$ ]] || return 3
[[ "${sdev_lun}" =~ ^0x[[:xdigit:]]{16}$ ]] || return 4
# After updating the firmware re-IPL information below we also try to
# update the information stored in the ID file (necessary, so it
# contains the correct Device-Bus-ID/WWPN/LUN after the update). For
# the update of the ID file we try to grab an exclusive lock, so there
# are no overlapping reads/writes.
#
# In case we can't get the lock because the ID file is missing, but we
# have a direct TGT match, we may still try to change the re-IPL
# information, but skip the ID file update.
#
# "direct match" means, the event subject is either the SDEV that is
# currently set as re-IPL target, or it is the dm-multipath device that
# currently contains the re-IPL target.
if ! id_file_lock_exclusive_no_create; then
# rc == 1 --> could not read ${ID_FILE}
[ "${PIPESTATUS[0]}" -eq 1 ] || return 5
# if true, we are dealing with a direct TGT match
[ "${CHREIPL_FCP_MPATH_IS_TGT}" = "true" ] || return 6
try_update_id_file=false
fi
# If we have a direct match (see in the comment above), we know
# that we have a path to the current re-IPL volume - no matter of the
# WWID. Otherwise, we got here by comparing the WWID of the event
# subject with the one recorded in the ID file; in this case we try to
# make sure the information is still up-to-date.
if [ "${CHREIPL_FCP_MPATH_IS_TGT}" != "true" ]; then
# last bail to make sure we don't overwrite user choices..
#
# XXX: this will *NOT* prevent the race completely, but at least
# make it less likely
{ readarray -d "" -t -u "${ID_FILE_LOCK}" records; } 2>/dev/null \
|| return 7
if '@DEBUG@'; then declare -p records 1>&2; fi
[ "${#records[@]}" = "4" ] || return 8
[ "${records[0]}" = "${sdev_wwid}" ] || return 9
[ "${records[1]}" = "${ipl_busid}" ] || return 10
[ "${records[2]}" = "${ipl_wwpn}" ] || return 11
[ "${records[3]}" = "${ipl_lun}" ] || return 12
fi
# Take lock so we don't see any intermediate state from other helpers
# running in parallel
firmware_lock_exclusive || return 13
if ! { echo "${sdev_busid}" >| /sys/firmware/reipl/fcp/device \
&& echo "${sdev_wwpn}" >| /sys/firmware/reipl/fcp/wwpn \
&& echo "${sdev_lun}" >| /sys/firmware/reipl/fcp/lun; };
then
log_alert "Changing the re-IPL device failed. The current re-IPL settings might be inconsistent. Check and correct the settings (see the README.md of chreipl-fcp-mpath) to make sure that the current re-IPL device is valid."
return 14
fi
firmware_unlock_exclusive
if [ "${sdev_busid}" != "${ipl_busid}" ] \
|| [ "${sdev_wwpn}" != "${ipl_wwpn}" ] \
|| [ "${sdev_lun}" != "${ipl_lun}" ]; then
log_note "Changed re-IPL path to: ${sdev_busid}:${sdev_wwpn}:${sdev_lun}."
fi
# Try to update the information in the ID file if we have gotten the
# lock for it.
if ${try_update_id_file}; then
# reset ID without removing the file
truncate --no-create --size=0 "${ID_FILE}" || return 15
echo -ne "${sdev_wwid}\x00${sdev_busid}\x00${sdev_wwpn}\x00${sdev_lun}\x00" \
>>"${ID_FILE}" || return 16
id_file_unlock_exclusive_no_create
if '@DEBUG@'; then hexdump -vC "${ID_FILE}" 1>&2; fi
fi
return 0
}
declare -g SDEV=""
if [[ "${DM_UUID}" == mpath-* ]]; then
# Assume Multipath Device Mapper Device;
# e.g.: DEVPATH = /devices/virtual/block/dm-0
for sdev in /sys/"${DEVPATH}"/slaves/sd*/device; do
if sdev_test_path_state "${sdev}"; then
SDEV="${sdev}"
break
fi
done
# No path of the multipath-device that represents the IPL volume is
# online.
if [ "${SDEV}" = "" ]; then
log_crit "The re-IPL device cannot be changed because no operational path to the re-IPL volume remains. The next re-IPL might fail unless you re-attach or enable at least one valid path to the re-IPL volume."
fi
elif [ "${SUBSYSTEM}" = block ]; then
# Assume SCSI Disk;
# e.g.: DEVPATH = /devices/css0/0.0.0014/0.0.1700/host0/rport-0:0-0/target0:0:0/0:0:0:1074806808/block/sds
if sdev_test_path_state /sys/"${DEVPATH}"/device; then
SDEV=/sys/"${DEVPATH}"/device
fi
fi
[ "${SDEV}" != "" ] || exit 0
sdev_get_wwid "${SDEV}" || exit 0
sdev_get_fcp_addressing "${SDEV}" || exit 0
firmware_get_ipl_information || exit 0
# shellcheck disable=SC2153
apply_ipl_information \
"${SDEV_WWID}" "${SDEV_BUSID}" "${SDEV_WWPN}" "${SDEV_LUN}" \
"${IPL_TYPE}" "${IPL_BUSID}" "${IPL_WWPN}" "${IPL_LUN}" || exit 0
exit 0

View File

@@ -0,0 +1,93 @@
# SPDX-License-Identifier: MIT
#
# chreipl-fcp-mpath: use multipath information to change FCP IPL target
# (C) Copyright IBM Corp. 2021
#
# Uses the following system-utilities (and shell-builtins):
# Utilities list in GNU Make Conventions:
# https://www.gnu.org/software/make/manual/make.html#Utilities-in-Makefiles
# GNU coreutils:
# - mktemp
# If $(ENABLE_DOC) is `1`:
# Pandoc:
# - pandoc
# GNU coreutils:
# - date
#
## Paths and Build Variables
#
# Install the configuration file for dracut, to automatically pull in the
# toolset into the initial ramdisk, when built with it.
HAVE_DRACUT = 0
# Build documentation; requires: Pandoc
ENABLE_DOC = 0
# https://www.gnu.org/software/make/manual/make.html#Directory-Variables
CHREIPLZFCPMPDIR = $(USRLIBDIR)/chreipl-fcp-mpath
UDEVDIR = $(USRLIBDIR)/udev
UDEVRULESDIR = $(UDEVDIR)/rules.d
UDEVRUNDIR = /run/udev
DRACUTDIR = $(USRLIBDIR)/dracut
DRACUTCONFDIR = $(DRACUTDIR)/dracut.conf.d
DEBUGOUTDIR = $(UDEVRUNDIR)
INSTALL_EXEC = $(INSTALL) -g $(GROUP) -o $(OWNER) --preserve-timestamps
INSTALL_DATA = $(INSTALL_EXEC) --mode=0644
# used for data exchange and synchronization across the different helpers
chreiplzfcpmp-id-file = $(UDEVRUNDIR)/chreiplzfcpmp-ipl-volume-id
# file used to implement mutual exclusion when accessing firmware IPL info:
# - this should be something that is (practically) always available, so we
# dont have to worry about fallbacks or error-handling;
# - at the same time, it should not be used by anything else with flock(2) to
# hold a lock for long periods.
chreiplzfcpmp-fwlock-file = /sys/firmware/reipl
.DELETE_ON_ERROR:
# export build-time definitions to the scripts/built-components
define chreiplzfcpmp-sed-buildvar-replace =
tmpout=$$(mktemp -p ./ .make.tmp.XXXXXXXXXXXXXXXX) && { \
$(SED) -E \
-e 's|@DEBUG@|$(if $(filter 1,$(D)),true,false)|g' \
-e 's|@chreiplzfcpmp-id-file@|$(chreiplzfcpmp-id-file)|g' \
-e 's|@chreiplzfcpmp-fwlock-file@|$(chreiplzfcpmp-fwlock-file)|g' \
-e 's|@chreiplzfcpmp-lib@|$(CHREIPLZFCPMPDIR)/chreipl-fcp-mpath-common.sh|g' \
-e 's|@debugoutdir@|$(DEBUGOUTDIR)|g' \
-e 's|@udevdir@|$(UDEVDIR)|g' \
-e 's|@udevrulesdir@|$(UDEVRULESDIR)|g' \
$(1) > $${tmpout} \
&& mv $${tmpout} $(2) \
|| { rm $${tmpout}; false; } \
; }
endef
.PHONY: clean-mk-temp
clean: clean-mk-temp
clean-mk-temp:
rm -f .make.tmp.[[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]][[:alnum:]]
# Definitions for generating documentation when $(ENABLE_DOC) is set to `1`
PANDOCFLAGS = --fail-if-warnings
ALL_PANDOCFLAGS = --preserve-tabs --tab-stop=8 --strip-comments \
--standalone --self-contained \
-M date="$(shell date +'%Y-%m-%d')" \
$(PANDOCFLAGS)
$(eval $(call cmd_define, PANDOC," PANDOC ",pandoc))
%.html : ALL_PANDOCFLAGS += -t html
%.html : %.md
$(PANDOC) $(ALL_PANDOCFLAGS) -f gfm -o $(@) $(<)
%.pdf : ALL_PANDOCFLAGS += -t latex --toc
%.pdf : %.md
$(PANDOC) $(ALL_PANDOCFLAGS) -f gfm -o $(@) $(<)
%.7 : ALL_PANDOCFLAGS += -t man
%.7 : %.md
$(PANDOC) $(ALL_PANDOCFLAGS) -f gfm -o $(@) $(<)

View File

@@ -0,0 +1,16 @@
# SPDX-License-Identifier: MIT
#
# chreipl-fcp-mpath: use multipath information to change FCP IPL target
# (C) Copyright IBM Corp. 2021
add_dracutmodules+=" bash multipath udev-rules "
install_items+=" @udevrulesdir@/70-chreipl-fcp-mpath.rules "
install_items+=" @chreiplzfcpmp-lib@ "
install_items+=" @udevdir@/chreipl-fcp-mpath-is-ipl-tgt "
install_items+=" @udevdir@/chreipl-fcp-mpath-is-ipl-vol "
install_items+=" @udevdir@/chreipl-fcp-mpath-is-reipl-zfcp "
install_items+=" @udevdir@/chreipl-fcp-mpath-record-volume-identifier "
install_items+=" @udevdir@/chreipl-fcp-mpath-try-change-ipl-path "
install_items+=" flock hexdump logger mktemp readlink sync truncate "

View File

@@ -0,0 +1,101 @@
# SPDX-License-Identifier: MIT
#
# chreipl-fcp-mpath: use multipath information to change FCP IPL target
# (C) Copyright IBM Corp. 2021
# Did the event affect a multipath or scsi disk device?
ACTION=="change", KERNEL=="dm-[0-9]*", SUBSYSTEM=="block", \
ENV{DM_UUID}=="mpath-*", ENV{DM_ACTION}=="PATH_FAILED", \
GOTO="chreipl_fcp_mpath_path_change"
ACTION=="change", KERNEL=="dm-[0-9]*", SUBSYSTEM=="block", \
ENV{DM_UUID}=="mpath-*", ENV{DM_ACTION}=="PATH_REINSTATED", \
GOTO="chreipl_fcp_mpath_path_change"
ACTION=="add", KERNEL=="sd[a-z]*", SUBSYSTEM=="block", \
GOTO="chreipl_fcp_mpath_path_change"
GOTO="chreipl_fcp_mpath_end"
LABEL="chreipl_fcp_mpath_path_change"
# Is this system IPL'ed (IOW, are we on s390x)? And do we ReIPL via zFCP?
#
# udev(7): If no absolute path is given, the program is expected to live
# in /usr/lib/udev; otherwise, the absolute path must be
# specified.
TEST!="/sys/firmware/ipl", GOTO="chreipl_fcp_mpath_end"
PROGRAM!="chreipl-fcp-mpath-is-reipl-zfcp", GOTO="chreipl_fcp_mpath_end"
# Consider the following scenarios.
# Either:
#
# (A) We recognized a new SCSI Disk. This might represent:
# (a) the path we want to ReIPL from;
# (b) an alternative path to the volume we want to ReIPL from;
# (c) a path to some unrelated volume.
#
# Or:
#
# (B) We recognized a PATH_ event for a multipath device. This might represent:
# the path we want to ReIPL from:
# (a) went away;
# (b) came back online;
# an alternative path to the volume we want to ReIPL from:
# (c) went away;
# (d) came back online;
# (e) some unrelated multipath device saw an event.
# Test whether the affected device is, or contains, the current IPL target.
#
# This covers scenarios:
# (A) (a),
# (B) (a)/(b)/(c)/(d)
PROGRAM!="chreipl-fcp-mpath-is-ipl-tgt", \
ENV{CHREIPL_FCP_MPATH_IS_TGT}="false", \
GOTO="chreipl_fcp_mpath_not_direct_match"
ENV{CHREIPL_FCP_MPATH_IS_TGT}="true"
# Record the WWID, Device-Bus-ID, Remote WWPN, and LUN of the ReIPL target
# (see `chreipl-fcp-mpath-is-ipl-vol` for usecases). This information
# might change, depending on whether the machine operator changes the ReIPL
# target to a different volume.
#
# XXX: Because the kernel doesn't generate any events upon changing of
# the ReIPL target, the chreipl-fcp-mpath toolset can't take any
# actions until the next path event for the new target is
# generated. Following that, we assume that when the machine
# operator changes the ReIPL target, the new target is reachable
# and in a good state at this point in time.
PROGRAM!="chreipl-fcp-mpath-record-volume-identifier", \
GOTO="chreipl_fcp_mpath_try_change_ipl_path"
GOTO="chreipl_fcp_mpath_try_change_ipl_path"
# If the even subject is not a direct match (not the sdev that is the current
# ReIPL target, and not a mpath device that contains the current ReIPL target)
LABEL="chreipl_fcp_mpath_not_direct_match"
# While this sdev/mpath device doesn't directly correspond to the path
# currently set as ReIPL target, it might still point to the same volume.
#
# For mpath devices this can happen if the original ReIPL target is completely
# gone from the machine, and so there is no way we can successfully, directly
# compare the ReIPL parameters to the sdevs of the mpath device.
#
# For cases like these we recorded the volume identifier, which we now can
# compare, and so still decide whether we are addressing the correct volume.
#
# This covers scenarios:
# (A) (b)/(c),
# (B) (c)/(d)/(e)
#
# XXX: we recorded WWID, Device-Bus-ID, Remote WWPN, LUN of the ReIPL target at
# the time; if the latter three don't match the current ReIPL setting
# anymore, we have to assume that someone changed the ReIPL target
# manually, and we cannot use the WWID anymore since we can't possibly
# know whether that stayed the same when the change was done.
PROGRAM!="chreipl-fcp-mpath-is-ipl-vol", GOTO="chreipl_fcp_mpath_end"
# We are here because of scenarios:
# (A) (a)/(b),
# (B) (a)/(b)/(c)/(d)
LABEL="chreipl_fcp_mpath_try_change_ipl_path"
RUN{program}+="chreipl-fcp-mpath-try-change-ipl-path"
LABEL="chreipl_fcp_mpath_end"

View File

@@ -5,7 +5,7 @@ COMMON_INCLUDED = true
# The variable "DISTRELEASE" should be overwritten in rpm spec files with:
# "make DISTRELEASE=%{release}" and "make install DISTRELEASE=%{release}"
VERSION = 2
RELEASE = 18
RELEASE = 19
PATCHLEVEL = 0
DISTRELEASE = build-$(shell date +%Y%m%d)
S390_TOOLS_RELEASE = $(VERSION).$(RELEASE).$(PATCHLEVEL)-$(DISTRELEASE)
@@ -182,6 +182,7 @@ USRSBINDIR = $(INSTALLDIR)/usr/sbin
USRBINDIR = $(INSTALLDIR)/usr/bin
BINDIR = $(INSTALLDIR)/sbin
LIBDIR = $(INSTALLDIR)/lib
USRLIBDIR = $(INSTALLDIR)/usr/lib
USRLIB64DIR = $(INSTALLDIR)/usr/lib64
SYSCONFDIR = $(INSTALLDIR)/etc
MANDIR = $(INSTALLDIR)/usr/share/man
@@ -206,10 +207,10 @@ INSTDIRS = $(USRSBINDIR) $(USRBINDIR) $(BINDIR) $(LIBDIR) $(MANDIR) \
$(TOOLS_LIBDIR) $(TOOLS_DATADIR) \
$(ZFCPDUMP_DIR) $(SYSTEMDSYSTEMUNITDIR) \
$(USRLIB64DIR) $(USRINCLUDEDIR) $(ZKEYKMSPLUGINDIR) \
$(SOINSTALLDIR)
$(SOINSTALLDIR) $(USRLIBDIR)
OWNER = $(shell id -un)
GROUP = $(shell id -gn)
export INSTALLDIR BINDIR LIBDIR USRLIB64DIR MANDIR OWNER GROUP
export INSTALLDIR BINDIR LIBDIR USRLIBDIR USRLIB64DIR MANDIR OWNER GROUP
# Special defines for zfcpdump
ZFCPDUMP_IMAGE = zfcpdump-image

View File

@@ -46,9 +46,9 @@
#define CPUS_POSSIBLE "/sys/devices/system/cpu/possible"
#define CPUS_KERNELMAX "/sys/devices/system/cpu/kernel_max"
#define MAXCTRS 512
#define IOCTLSLEEP 60U
static const unsigned int ioctlsleep = 60;
static unsigned int read_interval = ioctlsleep, cfvn, csvn, authorization;
static unsigned int read_interval = IOCTLSLEEP, cfvn, csvn, authorization;
static unsigned long loop_count = 1;
static unsigned char *ioctlbuffer;
static bool allcpu;

430
hsci/hsci
View File

@@ -11,11 +11,36 @@
hsdev=""
ndev=""
hsci=""
hsdev_mac=""
hscibr=""
hscibp=""
hsci_mac=""
hsif_pnetid=""
netif_pnetid=""
hsci_pnetid=""
##############################################################################
# Concept:
# --------
# | hsci |
# --------
# / \
# -------- --------
# | ndev | | hsdev |
# -------- --------
#
# Detail (bridge w/ bridgeports):
# --------
# | hsci |
# --------
# |
# ----------------------
# | |hsci-bp| |
# | hsci-br --------- |
# | |
# -------- -------- |
# | ndev | | hsdev | |
# ----------------------
function usage {
cat <<-EOD
Usage: hsci COMMAND [OPTION]
@@ -42,50 +67,53 @@ function prereqs_check {
fi
}
function get_pnetid {
local netdev=$1
local pnetid=""
#### ROCE:
if [ -e /sys/class/net/$netdev/device/util_string ]; then
pnetids="$(cat /sys/class/net/$netdev/device/util_string | tr -d '\000' | iconv -f IBM-1047 -t ASCII)"
else
#### OSA /HiperSockets:
if [ -e /sys/class/net/$netdev/device/chpid ]; then
chpid="$(cat /sys/class/net/$netdev/device/chpid | tr [:upper:] [:lower:])"
pnetids="$(cat /sys/devices/css0/chp0.$chpid/util_string | tr -d '\000' | iconv -f IBM-1047 -t ASCII)"
fi
fi
if [ "$pnetids" != "" ]; then
port_if="$(cat /sys/class/net/$netdev/dev_port)"
(( idx=16*$port_if+1 ))
(( end=$idx+15 ))
pnetid="$(echo "$pnetids" | cut -c $idx-$end | tr -d ' ')"
fi
echo $pnetid
}
function check_pnetids {
# get PNETID of the HS
local hsif_pnetids=""
local netif_pnetids=""
if [ -e /sys/class/net/$hsdev/device/util_string ]; then
hsif_pnetids="$(cat /sys/class/net/$hsdev/device/util_string | tr -d '\000' | iconv -f IBM-1047 -t ASCII)"
if [ "$hsdev" != "" ]; then
hsif_pnetid="$(get_pnetid $hsdev)"
else
if [ -e /sys/class/net/$hsdev/device/chpid ]; then
chpid="$(cat /sys/class/net/$hsdev/device/chpid | tr [:upper:] [:lower:])"
hsif_pnetids="$(cat /sys/devices/css0/chp0.$chpid/util_string | tr -d '\000' | iconv -f IBM-1047 -t ASCII)"
fi
hsif_pnetid=""
fi
if [ "$hsif_pnetids" != "" ]; then
port_hsif="$(cat /sys/class/net/$hsdev/dev_port)"
(( idx=16*$port_hsif+1 ))
(( end=$idx+15 ))
hsif_pnetid="$(echo "$hsif_pnetids" | cut -c $idx-$end | tr -d ' ')"
fi
# get PNETID of the NET_DEV
if [ -e /sys/class/net/$ndev/device/util_string ]; then
netif_pnetids="$(cat /sys/class/net/$ndev/device/util_string | tr -d '\000' | iconv -f IBM-1047 -t ASCII)"
if [ "$ndev" != "" ]; then
netif_pnetid="$(get_pnetid $ndev)"
else
if [ -e /sys/class/net/$ndev/device/chpid ]; then
chpid="$(cat /sys/class/net/$ndev/device/chpid | tr [:upper:] [:lower:])"
netif_pnetids="$(cat /sys/devices/css0/chp0.$chpid/util_string | tr -d '\000' | iconv -f IBM-1047 -t ASCII)"
fi
fi
if [ "$netif_pnetids" != "" ]; then
port_netif="$(cat /sys/class/net/$ndev/dev_port)"
(( idx=16*$port_netif+1 ))
(( end=$idx+15 ))
netif_pnetid="$(echo "$netif_pnetids" | cut -c $idx-$end | tr -d ' ')"
netif_pnetid=""
fi
#Check PNETIDs
if [ "$hsif_pnetid" != "" ] && [ "$netif_pnetid" != "" ] && [ "$netif_pnetid" != "$hsif_pnetid" ]; then
echo "Error: $hsdev and $ndev have different PNETIDs! They are $hsif_pnetid and $netif_pnetid respectively" >&2
return 1
fi
if [ "$hsif_pnetid" != "" ] && [ "$netif_pnetid" != "" ] && [ "$netif_pnetid" == "$hsif_pnetid" ]; then
hsci_pnetid=$hsif_pnetid
else
if [ "$hsif_pnetid" != "" ]; then
hsci_pnetid=$hsif_pnetid
else
hsci_pnetid=$netif_pnetid
fi
return 0
fi
}
@@ -96,11 +124,11 @@ function verify_precon {
echo "Error: $hsdev does not exist" >&2
return 1
fi
if [ "$(cat /sys/class/net/$hsdev/device/card_type)" != "HiperSockets" ]; then
if [ "$(cat /sys/class/net/$hsdev/device/card_type 2>/dev/null)" != "HiperSockets" ]; then
echo "Error: $hsdev is not a HiperSockets device" >&2
return 1
fi
if [ "$(cat /sys/class/net/$hsdev/device/layer2)" != "1" ]; then
if [ "$(cat /sys/class/net/$hsdev/device/layer2 2>/dev/null)" != "1" ]; then
echo "Error: $hsdev is not in layer 2 mode" >&2
return 1
fi
@@ -108,7 +136,7 @@ function verify_precon {
echo "Error: Missing vnic-characteristics support" >&2
return 1
fi
if [ "$(cat /sys/class/net/$hsdev/device/vnicc/bridge_invisible)" == "n/a" ]; then
if [ "$(cat /sys/class/net/$hsdev/device/vnicc/bridge_invisible 2>/dev/null)" == "n/a" ]; then
echo "Error: $hsdev does not support vnicc" >&2
return 1
fi
@@ -121,8 +149,10 @@ function verify_precon {
return 1
fi
if [ $(ip link show $hsdev | grep master | wc -l) -ne 0 ]; then
echo "Error: $hsdev is already a bridge port" >&2
return 1
if [ $(ip link show $hsdev | grep "master $hsci" | wc -l) -eq 0 ]; then
echo "Error: $hsdev is already a subordinate to another master" >&2
return 1
fi
fi
#Pre-verify net_dev
@@ -130,17 +160,15 @@ function verify_precon {
echo "Error: $ndev does not exist" >&2
return 1
fi
if [ "$(cat /sys/class/net/$ndev/device/card_type)" == "HiperSockets" ]; then
echo "Error: $ndev is also a HiperSockets device" >&2
return 1
fi
if [ $(ip link show $ndev | grep UP | wc -l) -eq 0 ]; then
echo "Error: $ndev is not in state UP" >&2
return 1
fi
if [ $(ip link show $ndev | grep master | wc -l) -ne 0 ]; then
echo "Error: $ndev is already a bridge port" >&2
return 1
if [ $(ip link show $ndev | grep "master $hsci" | wc -l) -eq 0 ]; then
echo "Error: $ndev is already a subordinate to another master" >&2
return 1
fi
fi
#Check PNETIDs
@@ -153,10 +181,12 @@ function verify_precon {
}
function clean_up {
ip link del $hsci >/dev/null 2>&1
bridge link set dev $hsdev learning_sync off self >/dev/null 2>&1
echo 0 > /sys/class/net/$hsdev/device/vnicc/bridge_invisible >/dev/null 2>&1
bridge fdb del $hsdev_mac dev $ndev >/dev/null 2>&1
ip link del $hsci >/dev/null 2>&1
bridge fdb del $hsci_mac dev $hsdev self local >/dev/null 2>&1
bridge fdb del $hsci_mac dev $ndev self local >/dev/null 2>&1
ip link del $hscibr >/dev/null 2>&1
}
##############################################################################
@@ -172,22 +202,33 @@ function add_hsci {
hsdev=$1
ndev=$2
hsci_postfix="$(readlink /sys/class/net/$hsdev/device/cdev0 | tail -c5)"
hscibr=hsci$hsci_postfix-br
hscibp=hsci$hsci_postfix-bp
hsci=hsci$hsci_postfix
#### Verify preconditions
verify_precon
if [ $? -ne 0 ]; then
return 1
fi
hsci_postfix="$(readlink /sys/class/net/$hsdev/device/cdev0 | tail -c5)"
hsci=hsci$hsci_postfix
echo "Adding $hsci with a HiperSockets dev $hsdev and an external dev $ndev"
#### Create bridge
ip link add name $hsci type bridge stp_state 0 >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Could not create a bridge" >&2
return 1
#### Create bridge (idempotent)
if [ ! -e /sys/class/net/$hscibr ]; then
# ageing_time of $hscibr defaults to 30000 (300 secs)
ip link add name $hscibr type bridge stp_state 0 >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Could not create a bridge" >&2
return 1
fi
else
ip link set dev $hscibr type bridge stp_state 0 >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: $hscibr is not a bridge" >&2
return 1
fi
fi
#### Prepare hsdev
@@ -196,19 +237,24 @@ function add_hsci {
echo 1 > /sys/class/net/$hsdev/device/vnicc/bridge_invisible
#### Create bridge ports
ip link set dev $ndev master $hsci >/dev/null 2>&1
ip link set dev $ndev master $hscibr >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Could not set master for $ndev" >&2
clean_up
return 1
fi
ip link set dev $hsdev master $hsci >/dev/null 2>&1
ip link set dev $hsdev master $hscibr >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Could not set master for $hsdev" >&2
clean_up
return 1
fi
# Do not learn from ndev, but do learn from hsci-bp:
# - First define hsdev and ndev as _isolated_ bridgeports
# - Then turn on learning_sync on self on hsdev
# - Then define hsci-bp as non-isolated veth bridgeport
# no forwarding between ndev and hsdev -> isolated on
# ndev is default for outgoing unknown targets -> flood on
# no need to learn external LAN targets into fdb -> learning off
@@ -230,29 +276,102 @@ function add_hsci {
clean_up
return 1
fi
# NOTE: Although not required, BCs will be sent out on hsdev.
# NOTE: We need to receive BCs on hsdev, as z/OS HSCI does ARP requests on HS.
hsdev_mac="$(cat /sys/class/net/$hsdev/address)"
echo "Set $hsdev MAC $hsdev_mac on $ndev and $hsci"
# set HS MAC on OSA as secondary MAC
bridge fdb add $hsdev_mac dev $ndev >/dev/null 2>&1
ip link set dev $hscibr up >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to set HS MAC on OSA as secondary MAC" >&2
echo "Error: Failed to set $hscibr up" >&2
clean_up
return 1
fi
# set HS MAC (common MAC) on HSCI as primary MAC
ip link set address $hsdev_mac dev $hsci >/dev/null 2>&1
#### Turn on device for bridge notification
#### Toggle is required to learn full list of HS targets,
#### not only future changes.
bridge link set dev $hsdev learning_sync off self >/dev/null 2>&1
bridge link set dev $hsdev learning_sync on self >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to set HiperSockets MAC (common MAC) on HSCI as primary MAC" >&2
echo "Error: Failed to turn on device for bridge notification" >&2
clean_up
return 1
fi
# define veth pair for hsci (idempotent)
if [ ! -e /sys/class/net/$hsci ]; then
ip link add dev $hsci type veth peer name $hscibp >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Could not create veth pair $hsci - $hscibp " >&2
clean_up
return 1
fi
else
if [ $hsci@$hscibp: != "$(ip -o -d link show dev $hsci | awk '/veth/ {print $2}' )" ]; then
echo "Error: $hsci@$hscibp is not a veth" >&2
clean_up
return 1
fi
fi
ip link set dev $hscibp master $hscibr >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to add $hscibp to $hscibr" >&2
clean_up
return 1
fi
bridge link set dev $hscibp isolated off learning on flood on mcast_flood on >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to set bridge parameters for $hscibp" >&2
clean_up
return 1
fi
#### Set a static forwarding rule for hsci MAC, so hsci can be used as a
#### single-MAC network interface without being subject to
#### ageing and re-learning
#### Wait for systemd to change the MAC of hsci, if it wants to:
sleep 1
hsci_mac="$(cat /sys/class/net/$hsci/address)"
#### (idempotent)
if [ $(bridge fdb show dev $hscibp | grep "$hsci_mac master $hscibr static" | wc -l) -eq 0 ]; then
bridge fdb add $hsci_mac dev $hscibp master static
if [ $? -ne 0 ]; then
echo "Error: Failed to set $hsci_mac to $hscibr fdb" >&2
clean_up
return 1
fi
fi
# Bridge-to-device learning will set this MAC on hsdev and ndev.
# Old kernel code doesn't do hsci bridge-to-device learning.
# In this case: Set hsci_mac as local MAC of hsdev and ndev,
# so at least the single-MAC scenario works.
if [ $(bridge fdb show dev $hsdev | grep "$hsci_mac self permanent" | wc -l) -eq 0 ]; then
echo "Warning: $hsci will support only its current static MAC address. Please upgrade your kernel to the latest level." >&2
bridge fdb add $hsci_mac dev $hsdev self local
if [ $? -ne 0 ]; then
echo "Error: Failed to add $hsci_mac to $hsdev" >&2
clean_up
return 1
fi
fi
if [ $(bridge fdb show dev $ndev | grep "$hsci_mac self permanent" | wc -l) -eq 0 ]; then
bridge fdb add $hsci_mac dev $ndev self local
if [ $? -ne 0 ]; then
echo "Error: Failed to add $hsci_mac to $ndev" >&2
clean_up
return 1
fi
fi
#### Set veth pair to UP
ip link set dev $hscibp up >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to set $hscibp up" >&2
clean_up
return 1
fi
ip link set dev $hsci up >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to set $hsci up" >&2
@@ -260,14 +379,7 @@ function add_hsci {
return 1
fi
# Turn on device for bridge notification
bridge link set dev $hsdev learning_sync on self >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to turn on device for bridge notification" >&2
clean_up
return 1
fi
echo "Successfully added HSCI interface $hsci"
echo "Added HSCI interface $hsci"
return 0
}
@@ -286,51 +398,97 @@ function del_hsci {
echo "Error: $hsci does not exit" >&2
return 1
fi
if [ $(ip link show | grep "master $hsci" | wc -l) -eq 0 ]; then
echo "Error: $hsci is not an active HSCI interface" >&2
return 1
hsci_mac="$(cat /sys/class/net/$hsci/address)"
#### Find hscibp and hscibr
hscibp="$(ip -o link show dev $hsci | awk '{print $2}')"
if [[ $hscibp != *@* ]]; then
# $hsci has no HSCI veth peer
echo "Warning: $hsci may have been created by an older version of hsci" >&2
mvp=1
hscibp=""
hscibr=$hsci
else
mvp=0
hscibp=${hscibp##*@}
hscibp=${hscibp%:}
echo "$hsci is paired with $hscibp" >&2
hscibr="$(ip link show dev $hscibp | awk '{for(x=1;x<NF;x++) if($x~/master/) print $(x+1)}')"
fi
bports="$(ip link show | grep "master $hsci" | awk '{print $2}')"
#### Find hsdev and ndev
hsdev=""
ndev=""
bports="$(ip link show | grep "master $hscibr" | awk '{print $2}')"
for bport in $bports; do
bport=${bport%:}
if [[ $bport == *@* ]]; then
bport=${bport%@*}
fi
if [ $(bridge -d link show dev $bport | grep "learning_sync on" | wc -l) -ne 0 ]; then
hsdev=$bport
else
ndev=$bport
bport=${bport%@*}
if [ $(bridge -d link show dev $bport | grep "isolated on" | wc -l) -ne 0 ]; then
if [ $(bridge -d link show dev $bport | grep "learning_sync on" | wc -l) -ne 0 ]; then
hsdev=$bport
else
ndev=$bport
fi
fi
done
if [ "$hsdev" == "" ]; then
echo "Error: $hsci has no active HiperSockets port" >&2
return 1
fi
echo "Deleting HSCI interface $hsci with the HiperSockets $hsdev and the external $ndev"
bridge link set dev $hsdev learning_sync off self >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to turn off learning_sync on $hsdev" >&2
return 1
fi
echo 0 > /sys/class/net/$hsdev/device/vnicc/bridge_invisible
hsdev_mac="$(cat /sys/class/net/$hsdev/address)"
echo "Deleting $hsev MAC $hsdev_mac on $ndev"
bridge fdb del $hsdev_mac dev $ndev >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to delete $hsev MAC $hsdev_mac on $ndev" >&2
return 1
if [ "$ndev" == "" ]; then
echo "Error: $hsci has no active external port" >&2
fi
echo "Deleting HSCI interface $hsci with HiperSockets interface $hsdev and external interface $ndev"
#### Delete veth before resetting learning_sync and deleting bridge,
#### so fdb entries are cleaned up (synced)!
ip link del $hsci >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to delete $hsci" >&2
return 1
fi
echo "Successfully deleted device $hsci"
# Bridge-to-device learning will remove learned MACs from hsdev and ndev.
# Old kernel code doesn't do hsci bridge-to-device learning.
# In this case: Remove the hsci_mac that was added during 'add' from
# hsdev and ndev.
if [ "$hsdev" != "" ] && [ $mvp -eq 0 ]; then
if [ $(bridge fdb show dev $hsdev | grep "$hsci_mac self permanent" | wc -l) -ne 0 ]; then
echo "Warning: It seems your kernel does not support all hsci features, please upgrade." >&2
bridge fdb del $hsci_mac dev $hsdev self local
if [ $? -ne 0 ]; then
echo "Error: Failed to delete $hsci_mac from $hsdev" >&2
fi
fi
fi
# In the mvp case hsci_mac == hs_mac and was only set on ndev.
if [ "$ndev" != "" ]; then
if [ $(bridge fdb show dev $ndev | grep "$hsci_mac self permanent" | wc -l) -ne 0 ]; then
bridge fdb del $hsci_mac dev $ndev self local
if [ $? -ne 0 ]; then
echo "Error: Failed to delete $hsci_mac from $ndev" >&2
fi
fi
fi
#### Reset learning_sync
if [ "$hsdev" != "" ]; then
bridge link set dev $hsdev learning_sync off self >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to turn off learning_sync on $hsdev" >&2
fi
fi
#### Delete bridge
if [ "$hscibr" != "" ]; then
ip link del $hscibr >/dev/null 2>&1
if [ $? -ne 0 ]; then
echo "Error: Failed to delete $hscibr" >&2
fi
fi
if [ "$hsdev" != "" ]; then
echo 0 > /sys/class/net/$hsdev/device/vnicc/bridge_invisible
fi
echo "Deleted $hsci"
return 0
}
@@ -338,19 +496,63 @@ function del_hsci {
## Show HSCI
##############################################################################
function list_active {
hsdev=$1
local ext=""
function print_row {
if [ $mvp -eq 0 ]; then
printf '%-8s %-16s %-15s %-15s\n' "$hsci" "$hsci_pnetid" "$hsdev" "$ndev"
else
printf '%-8s %-16s %-15s %-15s (v1)\n' "$hsci" "$hsci_pnetid" "$hsdev" "$ndev"
fi
}
hsci="$(ip link show dev $hsdev | awk '{for(x=1;x<NF;x++) if($x~/master/) print $(x+1)}')"
ext="$(ip link show | grep "master $hsci" | grep --invert-match $hsdev | awk '{print $2}')"
# remove trailing ':'
ndev="${ext%:}"
function list_active {
hsci=""
hsdev=$1
ndev=""
hscibp=""
mvp=1
hsci_pnetid="$(get_pnetid $hsdev)"
hscibr="$(ip link show dev $hsdev | awk '{for(x=1;x<NF;x++) if($x~/master/) print $(x+1)}')"
if [ "$hscibr" == "" ]; then
print_row
return 0
fi
#### find ndev (the other isolated subordinate)
bports="$(ip link show | grep "master $hscibr" | awk '{print $2}')"
for bport in $bports; do
bport=${bport%:}
bport=${bport%@*}
if [ $(ip -d link show dev $bport | grep "isolated on" | wc -l) -ne 0 ]; then
if [ $bport != $hsdev ]; then
ndev=$bport
fi
else
hscibp=$bport
mvp=0
fi
done
check_pnetids
printf '%-8s %-16s %-15s %-15s\n' "$hsci" "$hsci_pnetid" "$hsdev" "$ndev"
if [ $mvp -ne 0 ]; then
hsci=$hscibr
print_row
return 0
fi
#### find hsci (veth-peer of non-isolated subordinate of hscibr)
hsci="$(ip -o link show dev $hscibp | awk '{print $2}')"
if [[ $hsci != *@* ]]; then
echo "Error: $hscibp has no HSCI veth peer" >&2
print_row
return 1
fi
hsci=${hsci##*@}
hsci=${hsci%:}
print_row
return 0
}
@@ -358,13 +560,14 @@ function print_header {
if [ $header -eq 0 ]; then
echo "HSCI PNET_ID HiperSockets External "
echo "------------------------------------------------------------"
header=1
fi
}
function list_one {
local hsnetdev=$1
if [ $(bridge -d link show dev $hsnetdev | grep "learning_sync on" | wc -l) -ne 0 ]; then
if [ $(bridge -d link show dev $hsnetdev 2>/dev/null | grep "learning_sync on" | wc -l) -ne 0 ]; then
print_header
list_active $hsnetdev
fi
@@ -433,4 +636,3 @@ while true; do
esac
shift
done

View File

@@ -55,7 +55,7 @@ readonly ZDEV_CONF=$(lszdev --configured 2>/dev/null | wc -l)
readonly ZDEV_OFF=$(lszdev --offline 2>/dev/null | wc -l)
readonly ZDEV_ONL=$(lszdev --online 2>/dev/null | wc -l)
paramWORKDIR_BASE="/tmp/" # initial default path
paramWORKDIR_BASE="/tmp" # initial default path
########################################
# print dbginfo.sh version info
@@ -347,6 +347,7 @@ LOGFILES="\
/var/log/openvswitch/ovsdb-server.log\
/run/docker/libcontainerd/containerd/events.log\
/run/containerd/events.log\
/run/udev/chreiplzfcpmp-[0-9][0-9][a-z][a-z][a-z][a-z]-*\
"
########################################
@@ -404,6 +405,7 @@ CONFIGFILES="\
/lib/systemd/system/docker.service\
/usr/lib/systemd/system\
/etc/apparmor.d\
/run/udev/chreiplzfcpmp-ipl-volume-id\
"
########################################
@@ -412,7 +414,6 @@ CMDS="uname -a\
:timedatectl\
:runlevel\
:ulimit -a\
:blockdev --report\
:env\
:df -h\
:df -i\
@@ -1071,7 +1072,7 @@ call_run_command() {
# check if calling command and timeout exist
if which "${raw_cmd}" >/dev/null 2>&1 && which timeout >/dev/null 2>&1; then
eval timeout ${TOS} "${cmd}" >> ${logfile} 2>&1
eval timeout -k ${TOS} ${TOS} "${cmd}" >> ${logfile} 2>&1
rc=$?
# check if command is a builtin (no use of timeout possible)
elif command -v "${raw_cmd}" >/dev/null 2>&1; then

View File

@@ -42,7 +42,7 @@ check_dep_zlib:
all: check_dep_fuse check_dep_zlib zgetdump
OBJECTS = zgetdump.o opts.o zg.o \
dfi.o dfi_vmcoreinfo.o \
dfi.o dfi_mem_chunk.o dfi_vmcoreinfo.o \
dfi_lkcd.o dfi_elf.o \
dfi_s390.o dfi_s390_ext.o\
dfi_s390mv.o dfi_s390mv_ext.o \

View File

@@ -17,7 +17,7 @@
#include <time.h>
#include <unistd.h>
#include "zgetdump.h"
#include "df_s390.h"
/*
* Check, if we can access the lowcore information in the dump

View File

@@ -9,14 +9,19 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <time.h>
#include "lib/util_log.h"
#include "zgetdump.h"
#include "dfi_mem_chunk.h"
#include "dfi_vmcoreinfo.h"
#include "dfi.h"
#define TIME_FMT_STR "%a, %d %b %Y %H:%M:%S %z"
#define PROGRESS_HASH_CNT 50
#define KDUMP_OLDMEM_BASE 0x10418
#define KDUMP_OLDMEM_SIZE 0x10420
/*
* DFI vector - ensure that tape is the first in the list and devmem the second!
*/
@@ -49,17 +54,6 @@ struct cpus {
unsigned int cnt;
};
/*
* Memory information
*/
struct mem {
struct dfi_mem_chunk *chunk_cache;
u64 start_addr;
u64 end_addr;
unsigned int chunk_cnt;
struct util_list chunk_list;
};
/*
* Dump header attribute information
*/
@@ -83,8 +77,6 @@ struct attr {
static struct {
enum dfi_arch arch;
struct attr attr;
struct mem mem_phys;
struct mem mem_virt;
struct cpus cpus;
struct dfi *dfi;
unsigned long kdump_base;
@@ -111,129 +103,6 @@ static void date_print(void)
}
}
/*
* Initialize DFI memory chunks
*/
static void mem_init(struct mem *mem)
{
mem->start_addr = U64_MAX;
mem->end_addr = 0;
util_list_init(&mem->chunk_list, struct dfi_mem_chunk, list);
}
/*
* Memory chunk compare function for list sorting
*/
static int mem_chunk_cmp_fn(void *a, void *b, void *UNUSED(data))
{
struct dfi_mem_chunk *mem_chunk1 = a;
struct dfi_mem_chunk *mem_chunk2 = b;
return mem_chunk1->start < mem_chunk2->start ? -1 : 1;
}
/*
* Update DFI memory chunks
*/
static void mem_update(struct mem *mem)
{
struct dfi_mem_chunk *mem_chunk;
util_list_sort(&mem->chunk_list, mem_chunk_cmp_fn, NULL);
mem->start_addr = U64_MAX;
mem->end_addr = 0;
util_list_iterate(&mem->chunk_list, mem_chunk) {
mem->start_addr = MIN(mem->start_addr, mem_chunk->start);
mem->end_addr = MAX(mem->end_addr, mem_chunk->end);
}
}
/*
* Print memory map
*/
static void mem_map_print(void)
{
struct dfi_mem_chunk *mem_chunk;
u64 print_start = 0, print_end = 0;
const char *zero_str;
u32 volnr = 0;
STDERR("\nMemory map:\n");
/*
* Print each memory chunk if verbose specified
*/
if (g.opts.verbose) {
dfi_mem_chunk_iterate(mem_chunk) {
zero_str = "";
if (mem_chunk->read_fn == dfi_mem_chunk_read_zero)
zero_str = " zeroes";
STDERR(" %016llx - %016llx (%llu MB%s)\n",
mem_chunk->start, mem_chunk->end,
TO_MIB(mem_chunk->size), zero_str);
}
return;
}
/*
* Merge adjacent memory chunks from the same volume
*/
dfi_mem_chunk_iterate(mem_chunk) {
if (print_end == 0) {
print_start = mem_chunk->start;
print_end = mem_chunk->end;
volnr = mem_chunk->volnr;
continue;
}
if (mem_chunk->start != print_end + 1 ||
mem_chunk->volnr != volnr) {
STDERR(" %016llx - %016llx (%llu MB)\n", print_start,
print_end, TO_MIB(print_end - print_start + 1));
print_start = mem_chunk->start;
volnr = mem_chunk->volnr;
}
print_end = mem_chunk->end;
}
STDERR(" %016llx - %016llx (%llu MB)\n", print_start,
print_end, TO_MIB(print_end - print_start + 1));
}
/*
* Is memory range valid?
*/
int dfi_mem_range_valid(u64 addr, u64 len)
{
struct dfi_mem_chunk *mem_chunk;
u64 addr_end = addr + len;
/* check for unsigned wrap */
if (addr_end < addr)
return 0;
do {
mem_chunk = dfi_mem_chunk_find(addr);
if (!mem_chunk)
return 0;
addr += MIN(len, mem_chunk->end - addr + 1);
} while (addr < addr_end);
return 1;
}
/*
* Is memory already mapped at range?
*/
static int mem_range_mapped(u64 start, u64 size)
{
struct dfi_mem_chunk *mem_chunk;
u64 end = start + size - 1;
dfi_mem_chunk_iterate(mem_chunk) {
if (mem_chunk->start > end)
continue;
if (mem_chunk->end < start)
continue;
return 1;
}
return 0;
}
/*
* Print dump information (--info option)
*/
@@ -257,7 +126,7 @@ void dfi_info_print(void)
STDERR(" Volume number......: %u\n", *l.attr.vol_nr);
if (l.attr.build_arch)
STDERR(" Build arch.........: %s\n",
dfi_arch_str(*l.attr.build_arch));
dfi_arch_str(*l.attr.build_arch));
STDERR(" System arch........: %s\n", dfi_arch_str(l.arch));
if (l.cpus.cnt)
STDERR(" CPU count (online).: %u\n", l.cpus.cnt);
@@ -268,255 +137,18 @@ void dfi_info_print(void)
TO_MIB(dfi_mem_range()));
if (l.attr.mem_size_real)
STDERR(" Real memory range..: %lld MB\n",
TO_MIB(*l.attr.mem_size_real));
TO_MIB(*l.attr.mem_size_real));
if (l.attr.file_size)
STDERR(" Dump file size.....: %lld MB\n",
TO_MIB(*l.attr.file_size));
TO_MIB(*l.attr.file_size));
if (dfi_mem_range())
mem_map_print();
dfi_mem_map_print(g.opts.verbose);
if (l.dfi->info_dump) {
STDERR("\nDump device info:\n");
l.dfi->info_dump();
}
}
/*
* Add memory chunk to memory
*/
static void mem_chunk_create(struct mem *mem, u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn)
{
struct dfi_mem_chunk *mem_chunk;
mem_chunk = zg_alloc(sizeof(*mem_chunk));
mem_chunk->start = start;
mem_chunk->end = start + size - 1;
mem_chunk->size = size;
mem_chunk->read_fn = read_fn;
mem_chunk->free_fn = free_fn;
mem_chunk->data = data;
util_list_add_tail(&mem->chunk_list, mem_chunk);
mem->start_addr = MIN(mem->start_addr, mem_chunk->start);
mem->end_addr = MAX(mem->end_addr, mem_chunk->end);
mem->chunk_cache = mem_chunk;
mem->chunk_cnt++;
}
/*
* Check if memory chunk contains address
*/
static int mem_chunk_has_addr(struct dfi_mem_chunk *mem_chunk, u64 addr)
{
return (addr >= mem_chunk->start && addr <= mem_chunk->end);
}
/*
* Find memory chunk that contains address
*/
static struct dfi_mem_chunk *mem_chunk_find(struct mem *mem, u64 addr)
{
struct dfi_mem_chunk *mem_chunk;
if (mem->chunk_cache && mem_chunk_has_addr(mem->chunk_cache, addr))
return mem->chunk_cache;
util_list_iterate(&mem->chunk_list, mem_chunk) {
if (mem_chunk_has_addr(mem_chunk, addr)) {
mem->chunk_cache = mem_chunk;
return mem_chunk;
}
}
return NULL;
}
/*
* Read memory at given address
*/
static void mem_read(struct mem *mem, u64 addr, void *buf, size_t cnt)
{
struct dfi_mem_chunk *mem_chunk;
u64 size, off, copied = 0;
while (copied != cnt) {
mem_chunk = mem_chunk_find(mem, addr);
size = MIN(cnt - copied, mem_chunk->end - addr + 1);
off = addr - mem_chunk->start;
mem_chunk->read_fn(mem_chunk, off, buf + copied, size);
copied += size;
addr += size;
}
}
/*
* Read memory for virtual map memory chunk
*/
static void mem_chunk_map_read_fn(struct dfi_mem_chunk *mem_chunk, u64 off,
void *buf, u64 cnt)
{
u64 *start = mem_chunk->data;
dfi_mem_phys_read(*start + off, buf, cnt);
}
/*
* Check if memory chunk is a virtual mapping
*/
static int mem_chunk_is_map(struct dfi_mem_chunk *mem_chunk)
{
return mem_chunk->read_fn == mem_chunk_map_read_fn;
}
/*
* Return physical start address for memory chunk
*/
static u64 mem_chunk_start_phys(struct dfi_mem_chunk *mem_chunk)
{
if (mem_chunk_is_map(mem_chunk))
return *((u64 *) mem_chunk->data);
else
return mem_chunk->start;
}
/*
* Add virtual memory chunk with simple virtual mapping
*/
static void mem_chunk_map_add(u64 start, u64 size, u64 start_p)
{
u64 *data = zg_alloc(sizeof(*data));
*data = start_p;
dfi_mem_chunk_virt_add(start, size, data, mem_chunk_map_read_fn,
zg_free);
}
/*
* Add virtual memory chunk
*/
void dfi_mem_chunk_virt_add(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn)
{
util_log_print(UTIL_LOG_DEBUG,
"DFI add %svirt mem chunk start 0x%016lx size 0x%016lx\n",
read_fn == dfi_mem_chunk_read_zero ? "zero " : "",
start, size);
if (size == 0)
return;
mem_chunk_create(&l.mem_virt, start, size, data, read_fn, free_fn);
}
/*
* Add memory chunk with volume index
*/
void dfi_mem_chunk_add_vol(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn,
u32 volnr)
{
util_log_print(UTIL_LOG_DEBUG,
"DFI add %svol mem chunk start 0x%016lx size 0x%016lx volnr %u\n",
read_fn == dfi_mem_chunk_read_zero ? "zero " : "",
start, size, volnr);
if (size == 0)
return;
mem_chunk_create(&l.mem_phys, start, size, data, read_fn, free_fn);
mem_chunk_create(&l.mem_virt, start, size, data, read_fn, NULL);
l.mem_virt.chunk_cache->volnr = volnr;
}
/*
* Add memory chunk
*/
void dfi_mem_chunk_add(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn)
{
dfi_mem_chunk_add_vol(start, size, data, read_fn, free_fn, 0);
}
/*
* Read zero pages
*/
void dfi_mem_chunk_read_zero(struct dfi_mem_chunk *UNUSED(mem_chunk),
u64 UNUSED(off), void *buf, u64 cnt)
{
memset(buf, 0, cnt);
}
/*
* Return mem_chunk list head
*/
struct util_list *dfi_mem_chunk_list(void)
{
return &l.mem_virt.chunk_list;
}
/*
* Return number of memory chunks in input dump
*/
unsigned int dfi_mem_chunk_cnt(void)
{
return l.mem_virt.chunk_cnt;
}
/*
* Return maximum memory range
*/
u64 dfi_mem_range(void)
{
if (l.mem_virt.start_addr == U64_MAX)
return 0;
return l.mem_virt.end_addr - l.mem_virt.start_addr + 1;
}
/*
* Return first memory chunk
*/
struct dfi_mem_chunk *dfi_mem_chunk_first(void)
{
if (util_list_is_empty(&l.mem_virt.chunk_list))
return NULL;
return util_list_start(&l.mem_virt.chunk_list);
}
/*
* Return last memory chunk
*/
struct dfi_mem_chunk *dfi_mem_chunk_last(void)
{
if (util_list_is_empty(&l.mem_virt.chunk_list))
return NULL;
return util_list_end(&l.mem_virt.chunk_list);
}
/*
* Return next memory chunk
*/
struct dfi_mem_chunk *dfi_mem_chunk_next(struct dfi_mem_chunk *mem_chunk)
{
return util_list_next(&l.mem_virt.chunk_list, mem_chunk);
}
/*
* Return previous memory chunk
*/
struct dfi_mem_chunk *dfi_mem_chunk_prev(struct dfi_mem_chunk *mem_chunk)
{
return util_list_prev(&l.mem_virt.chunk_list, mem_chunk);
}
/*
* Find memory chunk for given address
*/
struct dfi_mem_chunk *dfi_mem_chunk_find(u64 addr)
{
return mem_chunk_find(&l.mem_virt, addr);
}
/*
* Initialize CPU info
*/
@@ -616,39 +248,6 @@ struct util_list *dfi_cpu_list(void)
return &l.cpus.list;
}
/*
* Read memory at given address and do kdump swap if necessary
*/
void dfi_mem_read(u64 addr, void *buf, size_t cnt)
{
util_log_print(UTIL_LOG_TRACE,
"DFI virt mem read addr 0x%016lx size 0x%016lx\n",
addr, cnt);
mem_read(&l.mem_virt, addr, buf, cnt);
}
/*
* Read physical memory at given address
*/
void dfi_mem_phys_read(u64 addr, void *buf, size_t cnt)
{
util_log_print(UTIL_LOG_TRACE,
"DFI phys mem read addr 0x%016lx size 0x%016lx\n",
addr, cnt);
mem_read(&l.mem_phys, addr, buf, cnt);
}
/*
* Read memory at given address with return code
*/
int dfi_mem_read_rc(u64 addr, void *buf, size_t cnt)
{
if (!dfi_mem_range_valid(addr, cnt))
return -EINVAL;
dfi_mem_read(addr, buf, cnt);
return 0;
}
/*
* Get input dump format name
*/
@@ -908,7 +507,7 @@ static void lc2cpu_64(struct dfi_cpu *cpu, struct dfi_lowcore_64 *lc)
/* Add VX registers if available */
if (!dfi_cpu_lc_has_vx_sa(lc))
return;
if (dfi_mem_read_rc(lc->vector_save_area_addr, &vx_sa, sizeof(vx_sa))) {
if (dfi_mem_virt_read(lc->vector_save_area_addr, &vx_sa, sizeof(vx_sa))) {
STDERR("zgetdump: Vector registers save area is beyond dump memory limit for CPU %d\n", cpu->cpu_id);
return;
}
@@ -952,13 +551,13 @@ int dfi_cpu_add_from_lc(u32 lc_addr)
if (l.arch == DFI_ARCH_32) {
struct dfi_cpu_32 cpu_32;
struct dfi_lowcore_32 lc;
if (dfi_mem_read_rc(lc_addr, &lc, sizeof(lc)))
if (dfi_mem_virt_read(lc_addr, &lc, sizeof(lc)))
return -EINVAL;
lc2cpu_32(&cpu_32, &lc);
cpu_32_to_64(cpu, &cpu_32);
} else {
struct dfi_lowcore_64 lc;
if (dfi_mem_read_rc(lc_addr, &lc, sizeof(lc)))
if (dfi_mem_virt_read(lc_addr, &lc, sizeof(lc)))
return -EINVAL;
lc2cpu_64(cpu, &lc);
}
@@ -1009,89 +608,6 @@ unsigned long dfi_kdump_base(void)
return l.kdump_base;
}
/*
* Unmap memory region
*/
static void mem_unmap(u64 start, u64 size)
{
u64 start_phys, end_phys, addr_phys, addr_virt, size_virt;
struct dfi_mem_chunk *mem_chunk, *tmp;
u64 end = start + size - 1;
util_list_iterate_safe(&l.mem_virt.chunk_list, mem_chunk, tmp) {
/*
* Chunk not hit?
*/
if (mem_chunk->start >= start + size)
continue;
if (mem_chunk->end < start)
continue;
/*
* Chunk completely unmapped
*
* UNMAP: UUUUUUUUU || UUUUUU
* CHUNK: CCCC || CCCCCC
* TO:
*/
if (mem_chunk->start >= start && mem_chunk->end <= end)
goto free;
/*
* Get real start and end addresses
*/
start_phys = mem_chunk_start_phys(mem_chunk);
end_phys = start_phys + mem_chunk->size - 1;
/*
* Chunk hit at start or in the middle?
*
* UNMAP: UUUUUU || UU || UUU
* CHUNK: CCCCC || CCCCCC || CCCC
* TO: NN || NN || NNN
*/
if (mem_chunk->end > end) {
addr_virt = end + 1;
size_virt = mem_chunk->end - end;
addr_phys = end_phys - size_virt + 1;
mem_chunk_map_add(addr_virt, size_virt, addr_phys);
}
/*
* Chunk hit at end or in the middle?
*
* UNMAP: UUUUUU || UU || UUU
* CHUNK: CCCCC || CCCCCC || CCC
* TO: NN || NN || NN
*/
if (mem_chunk->start < start) {
addr_virt = mem_chunk->start;
size_virt = start - addr_virt;
addr_phys = start_phys;
mem_chunk_map_add(addr_virt, size_virt, addr_phys);
}
free:
util_list_remove(&l.mem_virt.chunk_list, mem_chunk);
l.mem_virt.chunk_cnt--;
if (mem_chunk->data && mem_chunk->free_fn)
mem_chunk->free_fn(mem_chunk->data);
zg_free(mem_chunk);
}
mem_update(&l.mem_virt);
}
/*
* Map memory region
*/
static void mem_map(u64 start, u64 size, u64 start_phys)
{
if (mem_range_mapped(start, size)) {
mem_map_print();
ABORT("Map request for already mapped region (%llx/%llx/%llx)",
start, size, start_phys);
}
mem_chunk_map_add(start, size, start_phys);
mem_update(&l.mem_virt);
}
/*
* Check if dump contains a kdump dump and initialize kdump_base and kdump_size
*/
@@ -1101,12 +617,10 @@ static void kdump_init(void)
util_log_print(UTIL_LOG_TRACE, "DFI kdump initialization\n");
if (!dfi_mem_range_valid(0x10418, sizeof(base)))
if (dfi_mem_phys_read(KDUMP_OLDMEM_BASE, &base, sizeof(base)))
return;
if (!dfi_mem_range_valid(0x10420, sizeof(size)))
if (dfi_mem_phys_read(KDUMP_OLDMEM_SIZE, &size, sizeof(size)))
return;
dfi_mem_phys_read(0x10418, &base, sizeof(base));
dfi_mem_phys_read(0x10420, &size, sizeof(size));
if (base == 0 || size == 0)
return;
if (base % MIB || size % MIB)
@@ -1127,11 +641,11 @@ static void kdump_init(void)
if (!g.opts.select_specified)
return;
if (g.opts.select == OPTS_SELECT_PROD) {
mem_unmap(0, size);
mem_unmap(base, size);
mem_map(0, size, base);
dfi_mem_unmap(0, size);
dfi_mem_unmap(base, size);
dfi_mem_map(0, size, base);
} else if (g.opts.select == OPTS_SELECT_KDUMP) {
mem_unmap(l.kdump_size, U64_MAX - l.kdump_size);
dfi_mem_unmap(l.kdump_size, U64_MAX - l.kdump_size);
}
}
@@ -1160,8 +674,8 @@ static void kdump_select_prod_init(void)
}
dfi_cpu_info_init(DFI_CPU_CONTENT_ALL);
for (i = 0; i < count; i++) {
if (dfi_mem_read_rc(ptr + i * sizeof(long), &prefix,
sizeof(prefix)))
if (dfi_mem_virt_read(ptr + i * sizeof(long), &prefix,
sizeof(prefix)))
continue;
if (prefix == 0)
continue;
@@ -1184,7 +698,7 @@ static void utsname_init(void)
if (dfi_vmcoreinfo_symbol(&ptr, "init_uts_ns"))
return;
if (dfi_mem_read_rc(ptr, buf, sizeof(buf)))
if (dfi_mem_virt_read(ptr, buf, sizeof(buf)))
return;
utsname = memchr(buf, 'L', sizeof(buf) - sizeof(*utsname));
if (!utsname)
@@ -1205,7 +719,7 @@ static void livedump_init(void)
util_log_print(UTIL_LOG_TRACE, "DFI livedump initialization\n");
if (dfi_mem_read_rc(0, &magic, sizeof(magic)))
if (dfi_mem_virt_read(0, &magic, sizeof(magic)))
return;
if (magic == dfi_live_dump_magic)
dfi_attr_dump_method_set(DFI_DUMP_METHOD_LIVE);
@@ -1244,8 +758,9 @@ int dfi_init(void)
util_log_print(UTIL_LOG_TRACE, "DFI initialization\n");
l.arch = DFI_ARCH_UNKNOWN;
mem_init(&l.mem_virt);
mem_init(&l.mem_phys);
rc = dfi_mem_chunk_init();
if (rc)
return rc;
attr_init();
dfi_cpu_info_init(DFI_CPU_CONTENT_NONE);
while ((dfi = dfi_vec[i])) {

View File

@@ -12,13 +12,12 @@
#ifndef DFI_H
#define DFI_H
#include <time.h>
#include <linux/utsname.h>
#include "lib/zt_common.h"
#include "lib/util_list.h"
#include "zg.h"
/*
* CPU info functions and definitions
*/
@@ -116,11 +115,11 @@ struct dfi_cpu_32 {
struct dfi_vxrs vxrs_high[16];
};
extern void dfi_cpu_64_to_32(struct dfi_cpu_32 *cpu_32, struct dfi_cpu *cpu_64);
void dfi_cpu_64_to_32(struct dfi_cpu_32 *cpu_32, struct dfi_cpu *cpu_64);
extern enum dfi_arch dfi_arch(void);
extern void dfi_arch_set(enum dfi_arch arch);
extern const char *dfi_arch_str(enum dfi_arch arch);
enum dfi_arch dfi_arch(void);
void dfi_arch_set(enum dfi_arch arch);
const char *dfi_arch_str(enum dfi_arch arch);
enum dfi_cpu_content {
DFI_CPU_CONTENT_NONE, /* No register information available */
@@ -129,113 +128,62 @@ enum dfi_cpu_content {
};
#define DFI_CPU_CONTENT_FAC_VX 0x00000001
extern int dfi_cpu_content_fac_check(int flags);
extern void dfi_cpu_content_fac_add(int flags);
int dfi_cpu_content_fac_check(int flags);
void dfi_cpu_content_fac_add(int flags);
#define dfi_cpu_iterate(cpu) \
util_list_iterate(dfi_cpu_list(), cpu)
extern struct util_list *dfi_cpu_list(void);
extern void dfi_cpu_info_init(enum dfi_cpu_content content);
extern struct dfi_cpu *dfi_cpu_alloc(void);
extern struct dfi_cpu *dfi_cpu(unsigned int cpu_nr);
extern void dfi_cpu_add(struct dfi_cpu *cpu);
extern unsigned int dfi_cpu_cnt(void);
extern enum dfi_cpu_content dfi_cpu_content(void);
extern int dfi_cpu_add_from_lc(u32 lc_addr);
struct util_list *dfi_cpu_list(void);
void dfi_cpu_info_init(enum dfi_cpu_content content);
struct dfi_cpu *dfi_cpu_alloc(void);
struct dfi_cpu *dfi_cpu(unsigned int cpu_nr);
void dfi_cpu_add(struct dfi_cpu *cpu);
unsigned int dfi_cpu_cnt(void);
enum dfi_cpu_content dfi_cpu_content(void);
int dfi_cpu_add_from_lc(u32 lc_addr);
#define DFI_VX_SA_SIZE (32 * 16)
extern int dfi_cpu_lc_has_vx_sa(void *lc);
extern void dfi_cpu_vx_copy(void *buf, struct dfi_cpu *cpu);
/*
* Mem chunk functions and definitions
*/
struct dfi_mem_chunk;
typedef void (*dfi_mem_chunk_read_fn)(struct dfi_mem_chunk *mem_chunk,
u64 off, void *buf, u64 cnt);
typedef void (*dfi_mem_chunk_free_fn)(void *data);
struct dfi_mem_chunk {
struct util_list_node list; /* List */
u64 start; /* Start address in memory */
u64 end; /* End address in memory */
u64 size; /* Size of chunk in dump file */
dfi_mem_chunk_read_fn read_fn; /* Chunk read callback */
dfi_mem_chunk_free_fn free_fn; /* Free data callback */
void *data; /* Data for callback */
u32 volnr; /* Volume id where chunk resides */
};
extern void dfi_mem_chunk_read_zero(struct dfi_mem_chunk *UNUSED(mem_chunk),
u64 UNUSED(off), void *buf, u64 cnt);
extern void dfi_mem_chunk_add_vol(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn,
u32 volnr);
extern void dfi_mem_chunk_add(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn);
extern void dfi_mem_chunk_virt_add(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn);
extern u64 dfi_mem_range(void);
extern int dfi_mem_range_valid(u64 addr, u64 len);
extern unsigned int dfi_mem_chunk_cnt(void);
extern struct dfi_mem_chunk *dfi_mem_chunk_first(void);
extern struct dfi_mem_chunk *dfi_mem_chunk_last(void);
extern struct dfi_mem_chunk *dfi_mem_chunk_next(struct dfi_mem_chunk *chunk);
extern struct dfi_mem_chunk *dfi_mem_chunk_prev(struct dfi_mem_chunk *chunk);
extern struct dfi_mem_chunk *dfi_mem_chunk_find(u64 addr);
extern struct util_list *dfi_mem_chunk_list(void);
#define dfi_mem_chunk_iterate(mem_chunk) \
util_list_iterate(dfi_mem_chunk_list(), mem_chunk)
int dfi_cpu_lc_has_vx_sa(void *lc);
void dfi_cpu_vx_copy(void *buf, struct dfi_cpu *cpu);
/*
* Dump header attribute set/get functions
*/
extern void dfi_attr_time_set(struct timeval *time);
extern struct timeval *dfi_attr_time(void);
void dfi_attr_time_set(struct timeval *time);
struct timeval *dfi_attr_time(void);
extern void dfi_attr_time_end_set(struct timeval *time_end);
extern struct timeval *dfi_attr_time_end(void);
void dfi_attr_time_end_set(struct timeval *time_end);
struct timeval *dfi_attr_time_end(void);
extern void dfi_attr_cpu_id_set(u64 cpu_id);
extern u64 *dfi_attr_cpu_id(void);
void dfi_attr_cpu_id_set(u64 cpu_id);
u64 *dfi_attr_cpu_id(void);
extern void dfi_attr_utsname_set(struct new_utsname *utsname);
extern struct new_utsname *dfi_attr_utsname(void);
void dfi_attr_utsname_set(struct new_utsname *utsname);
struct new_utsname *dfi_attr_utsname(void);
extern void dfi_attr_dump_method_set(char *dump_method);
extern char *dfi_attr_dump_method(void);
void dfi_attr_dump_method_set(char *dump_method);
char *dfi_attr_dump_method(void);
extern void dfi_attr_mem_size_real_set(u64 mem_size_real);
extern u64 *dfi_attr_mem_size_real(void);
void dfi_attr_mem_size_real_set(u64 mem_size_real);
u64 *dfi_attr_mem_size_real(void);
extern void dfi_attr_file_size_set(u64 dump_size);
extern u64 *dfi_attr_file_size(void);
void dfi_attr_file_size_set(u64 dump_size);
u64 *dfi_attr_file_size(void);
extern void dfi_attr_vol_nr_set(unsigned int vol_nr);
extern unsigned int *dfi_attr_vol_nr(void);
void dfi_attr_vol_nr_set(unsigned int vol_nr);
unsigned int *dfi_attr_vol_nr(void);
extern void dfi_attr_version_set(unsigned int dfi_version);
extern unsigned int *dfi_attr_dfi_version(void);
void dfi_attr_version_set(unsigned int dfi_version);
unsigned int *dfi_attr_dfi_version(void);
extern void dfi_attr_build_arch_set(enum dfi_arch build_arch);
extern enum dfi_arch *dfi_attr_build_arch(void);
void dfi_attr_build_arch_set(enum dfi_arch build_arch);
enum dfi_arch *dfi_attr_build_arch(void);
extern void dfi_attr_real_cpu_cnt_set(u32 real_cpu_cnt);
extern u32 *dfi_attr_real_cpu_cnt(void);
void dfi_attr_real_cpu_cnt_set(u32 real_cpu_cnt);
u32 *dfi_attr_real_cpu_cnt(void);
/*
* DFI external functions
*/
extern void dfi_mem_read(u64 addr, void *buf, size_t cnt);
extern int dfi_mem_read_rc(u64 addr, void *buf, size_t cnt);
extern void dfi_mem_phys_read(u64 addr, void *buf, size_t cnt);
extern void dfi_info_print(void);
void dfi_info_print(void);
/*
* DFI feature bits
@@ -243,25 +191,13 @@ extern void dfi_info_print(void);
#define DFI_FEAT_SEEK 0x1 /* Necessary for fuse mount */
#define DFI_FEAT_COPY 0x2 /* Necessary for stdout */
extern int dfi_feat_seek(void);
extern int dfi_feat_copy(void);
int dfi_feat_seek(void);
int dfi_feat_copy(void);
/*
* DFI kdump functions
*/
extern unsigned long dfi_kdump_base(void);
/*
* DFI vmcoreinfo functions
*/
extern void dfi_vmcoreinfo_init(void);
extern char *dfi_vmcoreinfo_get(void);
extern int dfi_vmcoreinfo_tag(char *str, int len, const char *sym);
extern int dfi_vmcoreinfo_symbol(unsigned long *val, const char *sym);
extern int dfi_vmcoreinfo_offset(unsigned long *offs, const char *sym);
extern int dfi_vmcoreinfo_size(unsigned long *size, const char *sym);
extern int dfi_vmcoreinfo_length(unsigned long *len, const char *sym);
extern int dfi_vmcoreinfo_val(unsigned long *val, const char *sym);
unsigned long dfi_kdump_base(void);
/*
* DFI operations
@@ -274,14 +210,14 @@ struct dfi {
int feat_bits;
};
extern const char *dfi_name(void);
extern int dfi_init(void);
extern void dfi_exit(void);
const char *dfi_name(void);
int dfi_init(void);
void dfi_exit(void);
/*
* Dump access
*/
extern struct zg_fh *dfi_dump_open(const char *path);
struct zg_fh *dfi_dump_open(const char *path);
/*
* Live dump memory magic
@@ -293,4 +229,18 @@ extern u64 dfi_live_dump_magic;
*/
#define DFI_DUMP_METHOD_LIVE "live"
/*
* Supported DFI dump formats
*/
extern struct dfi dfi_s390tape;
extern struct dfi dfi_s390mv;
extern struct dfi dfi_s390mv_ext;
extern struct dfi dfi_s390;
extern struct dfi dfi_s390_ext;
extern struct dfi dfi_lkcd;
extern struct dfi dfi_elf;
extern struct dfi dfi_kdump;
extern struct dfi dfi_kdump_flat;
extern struct dfi dfi_devmem;
#endif /* DFI_H */

View File

@@ -21,6 +21,9 @@
#include <unistd.h>
#include "zgetdump.h"
#include "df_s390.h"
#include "dfi.h"
#include "dfi_mem_chunk.h"
/*
* Add live dump magic to buffer

View File

@@ -18,7 +18,12 @@
#include "lib/util_libc.h"
#include "lib/util_log.h"
#include "zgetdump.h"
#include "zg.h"
#include "df_elf.h"
#include "dfi.h"
#include "dfi_mem_chunk.h"
/*
* Read memory for given memory chunk

View File

@@ -9,7 +9,12 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <errno.h>
#include "zgetdump.h"
#include "zg.h"
#include "dfi.h"
#include "dfi_mem_chunk.h"
struct df_kdump_hdr {
char signature[8];

View File

@@ -9,8 +9,15 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <string.h>
#include <errno.h>
#include <zlib.h>
#include "zgetdump.h"
#include "zg.h"
#include "df_lkcd.h"
#include "dfi.h"
#include "dfi_mem_chunk.h"
#define MEM_HOLE_SIZE_MIN (1024 * 1024)
#define IDX_KIB 64 /* One index entry per IDX_KIB */

527
zdump/dfi_mem_chunk.c Normal file
View File

@@ -0,0 +1,527 @@
/*
* Copyright IBM Corp. 2001, 2018
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <stdlib.h>
#include <string.h>
#include <errno.h>
#include "lib/util_libc.h"
#include "lib/util_log.h"
#include "zg.h"
#include "dfi_mem_chunk.h"
/*
* Memory information
*/
struct mem {
struct dfi_mem_chunk *chunk_cache;
u64 start_addr;
u64 end_addr;
unsigned int chunk_cnt;
struct util_list chunk_list;
};
/*
* File local static data
*/
static struct {
struct mem mem_phys;
struct mem mem_virt;
} l;
/*
* Initialize DFI memory chunks
*/
static void mem_init(struct mem *mem)
{
mem->start_addr = U64_MAX;
mem->end_addr = 0;
util_list_init(&mem->chunk_list, struct dfi_mem_chunk, list);
}
/*
* Memory chunk compare function for list sorting
*/
static int mem_chunk_cmp_fn(void *a, void *b, void *UNUSED(data))
{
struct dfi_mem_chunk *mem_chunk1 = a;
struct dfi_mem_chunk *mem_chunk2 = b;
return mem_chunk1->start < mem_chunk2->start ? -1 : 1;
}
/*
* Update DFI memory chunks
*/
static void mem_update(struct mem *mem)
{
struct dfi_mem_chunk *mem_chunk;
util_list_sort(&mem->chunk_list, mem_chunk_cmp_fn, NULL);
mem->start_addr = U64_MAX;
mem->end_addr = 0;
util_list_iterate(&mem->chunk_list, mem_chunk) {
mem->start_addr = MIN(mem->start_addr, mem_chunk->start);
mem->end_addr = MAX(mem->end_addr, mem_chunk->end);
}
}
/*
* Print memory map
*/
void dfi_mem_map_print(bool verbose)
{
struct dfi_mem_chunk *mem_chunk;
u64 print_start = 0, print_end = 0;
const char *zero_str;
u32 volnr = 0;
STDERR("\nMemory map:\n");
/*
* Print each memory chunk if verbose specified
*/
if (verbose) {
dfi_mem_chunk_iterate(mem_chunk) {
zero_str = "";
if (mem_chunk->read_fn == dfi_mem_chunk_read_zero)
zero_str = " zeroes";
STDERR(" %016llx - %016llx (%llu MB%s)\n",
mem_chunk->start, mem_chunk->end,
TO_MIB(mem_chunk->size), zero_str);
}
return;
}
/*
* Merge adjacent memory chunks from the same volume
*/
dfi_mem_chunk_iterate(mem_chunk) {
if (print_end == 0) {
print_start = mem_chunk->start;
print_end = mem_chunk->end;
volnr = mem_chunk->volnr;
continue;
}
if (mem_chunk->start != print_end + 1 ||
mem_chunk->volnr != volnr) {
STDERR(" %016llx - %016llx (%llu MB)\n", print_start,
print_end, TO_MIB(print_end - print_start + 1));
print_start = mem_chunk->start;
volnr = mem_chunk->volnr;
}
print_end = mem_chunk->end;
}
STDERR(" %016llx - %016llx (%llu MB)\n", print_start,
print_end, TO_MIB(print_end - print_start + 1));
}
/*
* Check if memory chunk contains address
*/
static int mem_chunk_has_addr(struct dfi_mem_chunk *mem_chunk, u64 addr)
{
return (addr >= mem_chunk->start && addr <= mem_chunk->end);
}
/*
* Find memory chunk that contains address
*/
static struct dfi_mem_chunk *mem_chunk_find(struct mem *mem, u64 addr)
{
struct dfi_mem_chunk *mem_chunk;
if (mem->chunk_cache && mem_chunk_has_addr(mem->chunk_cache, addr))
return mem->chunk_cache;
util_list_iterate(&mem->chunk_list, mem_chunk) {
if (mem_chunk_has_addr(mem_chunk, addr)) {
mem->chunk_cache = mem_chunk;
return mem_chunk;
}
}
return NULL;
}
/*
* Is memory range valid?
*/
static int mem_range_valid(struct mem *mem, u64 addr, u64 len)
{
struct dfi_mem_chunk *mem_chunk;
u64 addr_end = addr + len;
/* check for unsigned wrap */
if (addr_end < addr)
return 0;
do {
mem_chunk = mem_chunk_find(mem, addr);
if (!mem_chunk)
return 0;
addr += MIN(len, mem_chunk->end - addr + 1);
} while (addr < addr_end);
return 1;
}
/*
* Is memory already mapped at range?
*/
static int mem_range_mapped(u64 start, u64 size)
{
struct dfi_mem_chunk *mem_chunk;
u64 end = start + size - 1;
dfi_mem_chunk_iterate(mem_chunk) {
if (mem_chunk->start > end)
continue;
if (mem_chunk->end < start)
continue;
return 1;
}
return 0;
}
/*
* Add memory chunk to memory
*/
static void mem_chunk_create(struct mem *mem, u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn)
{
struct dfi_mem_chunk *mem_chunk;
mem_chunk = util_malloc(sizeof(*mem_chunk));
mem_chunk->start = start;
mem_chunk->end = start + size - 1;
mem_chunk->size = size;
mem_chunk->read_fn = read_fn;
mem_chunk->free_fn = free_fn;
mem_chunk->data = data;
util_list_add_tail(&mem->chunk_list, mem_chunk);
mem->start_addr = MIN(mem->start_addr, mem_chunk->start);
mem->end_addr = MAX(mem->end_addr, mem_chunk->end);
mem->chunk_cache = mem_chunk;
mem->chunk_cnt++;
}
/*
* Read memory at given address
*/
static void mem_read(struct mem *mem, u64 addr, void *buf, size_t cnt)
{
struct dfi_mem_chunk *mem_chunk;
u64 size, off, copied = 0;
while (copied != cnt) {
mem_chunk = mem_chunk_find(mem, addr);
size = MIN(cnt - copied, mem_chunk->end - addr + 1);
off = addr - mem_chunk->start;
mem_chunk->read_fn(mem_chunk, off, buf + copied, size);
copied += size;
addr += size;
}
}
/*
* Read memory for virtual map memory chunk
*/
static void mem_chunk_map_read_fn(struct dfi_mem_chunk *mem_chunk, u64 off,
void *buf, u64 cnt)
{
u64 *start = mem_chunk->data;
dfi_mem_phys_read(*start + off, buf, cnt);
}
/*
* Check if memory chunk is a virtual mapping
*/
static int mem_chunk_is_map(struct dfi_mem_chunk *mem_chunk)
{
return mem_chunk->read_fn == mem_chunk_map_read_fn;
}
/*
* Return physical start address for memory chunk
*/
static u64 mem_chunk_start_phys(struct dfi_mem_chunk *mem_chunk)
{
if (mem_chunk_is_map(mem_chunk))
return *((u64 *) mem_chunk->data);
else
return mem_chunk->start;
}
/*
* Add virtual memory chunk
*/
static void mem_chunk_virt_add(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn)
{
util_log_print(UTIL_LOG_DEBUG,
"DFI add %svirt mem chunk start 0x%016lx size 0x%016lx\n",
read_fn == dfi_mem_chunk_read_zero ? "zero " : "",
start, size);
if (size == 0)
return;
mem_chunk_create(&l.mem_virt, start, size, data, read_fn, free_fn);
}
/*
* Add virtual memory chunk with simple virtual mapping
*/
static void mem_chunk_map_add(u64 start, u64 size, u64 start_p)
{
u64 *data = util_malloc(sizeof(*data));
*data = start_p;
mem_chunk_virt_add(start, size, data, mem_chunk_map_read_fn, free);
}
/*
* Add memory chunk with volume index
*/
void dfi_mem_chunk_add_vol(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn,
u32 volnr)
{
util_log_print(UTIL_LOG_DEBUG,
"DFI add %svol mem chunk start 0x%016lx size 0x%016lx volnr %u\n",
read_fn == dfi_mem_chunk_read_zero ? "zero " : "",
start, size, volnr);
if (size == 0)
return;
mem_chunk_create(&l.mem_phys, start, size, data, read_fn, free_fn);
mem_chunk_create(&l.mem_virt, start, size, data, read_fn, NULL);
l.mem_virt.chunk_cache->volnr = volnr;
}
/*
* Add memory chunk
*/
void dfi_mem_chunk_add(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn)
{
dfi_mem_chunk_add_vol(start, size, data, read_fn, free_fn, 0);
}
/*
* Read zero pages
*/
void dfi_mem_chunk_read_zero(struct dfi_mem_chunk *UNUSED(mem_chunk),
u64 UNUSED(off), void *buf, u64 cnt)
{
memset(buf, 0, cnt);
}
/*
* Return mem_chunk list head
*/
struct util_list *dfi_mem_chunk_list(void)
{
return &l.mem_virt.chunk_list;
}
/*
* Return number of memory chunks in input dump
*/
unsigned int dfi_mem_chunk_cnt(void)
{
return l.mem_virt.chunk_cnt;
}
/*
* Return maximum memory range
*/
u64 dfi_mem_range(void)
{
if (l.mem_virt.start_addr == U64_MAX)
return 0;
return l.mem_virt.end_addr - l.mem_virt.start_addr + 1;
}
/*
* Is memory range valid?
*/
int dfi_mem_range_valid(u64 addr, u64 len)
{
return mem_range_valid(&l.mem_virt, addr, len);
}
/*
* Return first memory chunk
*/
struct dfi_mem_chunk *dfi_mem_chunk_first(void)
{
if (util_list_is_empty(&l.mem_virt.chunk_list))
return NULL;
return util_list_start(&l.mem_virt.chunk_list);
}
/*
* Return last memory chunk
*/
struct dfi_mem_chunk *dfi_mem_chunk_last(void)
{
if (util_list_is_empty(&l.mem_virt.chunk_list))
return NULL;
return util_list_end(&l.mem_virt.chunk_list);
}
/*
* Return next memory chunk
*/
struct dfi_mem_chunk *dfi_mem_chunk_next(struct dfi_mem_chunk *mem_chunk)
{
return util_list_next(&l.mem_virt.chunk_list, mem_chunk);
}
/*
* Return previous memory chunk
*/
struct dfi_mem_chunk *dfi_mem_chunk_prev(struct dfi_mem_chunk *mem_chunk)
{
return util_list_prev(&l.mem_virt.chunk_list, mem_chunk);
}
/*
* Find memory chunk for given address
*/
struct dfi_mem_chunk *dfi_mem_chunk_find(u64 addr)
{
return mem_chunk_find(&l.mem_virt, addr);
}
/*
* Read physical memory at given address
*/
int dfi_mem_phys_read(u64 addr, void *buf, size_t cnt)
{
util_log_print(UTIL_LOG_TRACE,
"DFI phys mem read addr 0x%016lx size 0x%016lx\n",
addr, cnt);
if (!mem_range_valid(&l.mem_phys, addr, cnt))
return -EINVAL;
mem_read(&l.mem_phys, addr, buf, cnt);
return 0;
}
/*
* Read virtual memory at given address
*/
int dfi_mem_virt_read(u64 addr, void *buf, size_t cnt)
{
util_log_print(UTIL_LOG_TRACE,
"DFI virt mem read addr 0x%016lx size 0x%016lx\n",
addr, cnt);
if (!mem_range_valid(&l.mem_virt, addr, cnt))
return -EINVAL;
mem_read(&l.mem_virt, addr, buf, cnt);
return 0;
}
/*
* Unmap memory region
*/
void dfi_mem_unmap(u64 start, u64 size)
{
u64 start_phys, end_phys, addr_phys, addr_virt, size_virt;
struct dfi_mem_chunk *mem_chunk, *tmp;
u64 end = start + size - 1;
util_list_iterate_safe(&l.mem_virt.chunk_list, mem_chunk, tmp) {
/*
* Chunk not hit?
*/
if (mem_chunk->start >= start + size)
continue;
if (mem_chunk->end < start)
continue;
/*
* Chunk completely unmapped
*
* UNMAP: UUUUUUUUU || UUUUUU
* CHUNK: CCCC || CCCCCC
* TO:
*/
if (mem_chunk->start >= start && mem_chunk->end <= end)
goto free;
/*
* Get real start and end addresses
*/
start_phys = mem_chunk_start_phys(mem_chunk);
end_phys = start_phys + mem_chunk->size - 1;
/*
* Chunk hit at start or in the middle?
*
* UNMAP: UUUUUU || UU || UUU
* CHUNK: CCCCC || CCCCCC || CCCC
* TO: NN || NN || NNN
*/
if (mem_chunk->end > end) {
addr_virt = end + 1;
size_virt = mem_chunk->end - end;
addr_phys = end_phys - size_virt + 1;
mem_chunk_map_add(addr_virt, size_virt, addr_phys);
}
/*
* Chunk hit at end or in the middle?
*
* UNMAP: UUUUUU || UU || UUU
* CHUNK: CCCCC || CCCCCC || CCC
* TO: NN || NN || NN
*/
if (mem_chunk->start < start) {
addr_virt = mem_chunk->start;
size_virt = start - addr_virt;
addr_phys = start_phys;
mem_chunk_map_add(addr_virt, size_virt, addr_phys);
}
free:
util_list_remove(&l.mem_virt.chunk_list, mem_chunk);
l.mem_virt.chunk_cnt--;
if (mem_chunk->data && mem_chunk->free_fn)
mem_chunk->free_fn(mem_chunk->data);
free(mem_chunk);
}
mem_update(&l.mem_virt);
}
/*
* Map memory region
*/
void dfi_mem_map(u64 start, u64 size, u64 start_phys)
{
if (mem_range_mapped(start, size)) {
dfi_mem_map_print(false);
ABORT("Map request for already mapped region (%llx/%llx/%llx)",
start, size, start_phys);
}
mem_chunk_map_add(start, size, start_phys);
mem_update(&l.mem_virt);
}
int dfi_mem_chunk_init(void)
{
mem_init(&l.mem_virt);
mem_init(&l.mem_phys);
return 0;
}
void dfi_mem_chunk_deinit(void)
{
memset(&l, 0, sizeof(l));
}

69
zdump/dfi_mem_chunk.h Normal file
View File

@@ -0,0 +1,69 @@
/*
* Copyright IBM Corp. 2001, 2018
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef DFI_MEM_CHUNK_H
#define DFI_MEM_CHUNK_H
#include <stdbool.h>
#include "lib/zt_common.h"
#include "lib/util_list.h"
/*
* Mem chunk functions and definitions
*/
struct dfi_mem_chunk;
typedef void (*dfi_mem_chunk_read_fn)(struct dfi_mem_chunk *mem_chunk,
u64 off, void *buf, u64 cnt);
typedef void (*dfi_mem_chunk_free_fn)(void *data);
struct dfi_mem_chunk {
struct util_list_node list; /* List */
u64 start; /* Start address in memory */
u64 end; /* End address in memory */
u64 size; /* Size of chunk in dump file */
dfi_mem_chunk_read_fn read_fn; /* Chunk read callback */
dfi_mem_chunk_free_fn free_fn; /* Free data callback */
void *data; /* Data for callback */
u32 volnr; /* Volume id where chunk resides */
};
void dfi_mem_chunk_read_zero(struct dfi_mem_chunk *UNUSED(mem_chunk),
u64 UNUSED(off), void *buf, u64 cnt);
void dfi_mem_chunk_add_vol(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn,
u32 volnr);
void dfi_mem_chunk_add(u64 start, u64 size, void *data,
dfi_mem_chunk_read_fn read_fn,
dfi_mem_chunk_free_fn free_fn);
u64 dfi_mem_range(void);
int dfi_mem_range_valid(u64 addr, u64 len);
unsigned int dfi_mem_chunk_cnt(void);
struct dfi_mem_chunk *dfi_mem_chunk_first(void);
struct dfi_mem_chunk *dfi_mem_chunk_last(void);
struct dfi_mem_chunk *dfi_mem_chunk_next(struct dfi_mem_chunk *chunk);
struct dfi_mem_chunk *dfi_mem_chunk_prev(struct dfi_mem_chunk *chunk);
struct dfi_mem_chunk *dfi_mem_chunk_find(u64 addr);
struct util_list *dfi_mem_chunk_list(void);
#define dfi_mem_chunk_iterate(mem_chunk) \
util_list_iterate(dfi_mem_chunk_list(), mem_chunk)
int dfi_mem_virt_read(u64 addr, void *buf, size_t cnt);
int dfi_mem_phys_read(u64 addr, void *buf, size_t cnt);
void dfi_mem_map_print(bool verbose);
void dfi_mem_unmap(u64 start, u64 size);
void dfi_mem_map(u64 start, u64 size, u64 start_phys);
int dfi_mem_chunk_init(void);
void dfi_mem_chunk_deinit(void);
#endif /* DFI_MEM_CHUNK_H */

View File

@@ -17,9 +17,16 @@
#include <sys/types.h>
#include <time.h>
#include <unistd.h>
#include <errno.h>
#include "lib/util_log.h"
#include "zgetdump.h"
#include "zg.h"
#include "dfi.h"
#include "dfi_mem_chunk.h"
#include "df_elf.h"
#include "df_s390.h"
/*
* File local static data

View File

@@ -9,7 +9,9 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include "zgetdump.h"
#include "dt.h"
#include "df_s390.h"
#include "dfi.h"
/*
* Initialize s390 input dump format (extended)

View File

@@ -20,12 +20,17 @@
#include <sys/sysmacros.h>
#include <time.h>
#include <unistd.h>
#include <errno.h>
#include "lib/util_file.h"
#include "lib/util_path.h"
#include "zgetdump.h"
#include "dt.h"
#include "df_s390.h"
#include "dfi.h"
#include "dfi_s390mv.h"
#include "dfi_mem_chunk.h"
/*
* Volume information

View File

@@ -10,6 +10,9 @@
*/
#include "zgetdump.h"
#include "dt.h"
#include "df_s390.h"
#include "dfi.h"
/*
* Initialize s390 multi-volume input dump format (extedend)

View File

@@ -17,8 +17,12 @@
#include <sys/types.h>
#include <time.h>
#include <unistd.h>
#include <errno.h>
#include "zgetdump.h"
#include "df_s390.h"
#include "dfi.h"
#include "dfi_mem_chunk.h"
#define TAPE_BLK_SIZE 32768 /* Defined by zipl tape dumper */

View File

@@ -9,12 +9,16 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <stdio.h>
#include <string.h>
#include <elf.h>
#include "lib/zt_common.h"
#include "lib/util_log.h"
#include "zgetdump.h"
#include "zg.h"
#include "dfi_mem_chunk.h"
#include "dfi_vmcoreinfo.h"
#ifdef __s390x__
#define LC_VMCORE_INFO 0xe0c
@@ -23,7 +27,7 @@
#endif
#define LC_OS_INFO 0xe18
#define OS_INFO_MAGIC 0x4f53494e464f535aULL /* OSINFOSZ */
#define OS_INFO_MAGIC 0x4f53494e464f535aULL /* OSINFOSZ */
struct os_info {
u64 magic;
@@ -65,11 +69,11 @@ static struct os_info *os_info_get(void)
util_log_print(UTIL_LOG_TRACE, "DFI get osinfo\n");
if (dfi_mem_read_rc(LC_OS_INFO, &addr, sizeof(addr)))
if (dfi_mem_virt_read(LC_OS_INFO, &addr, sizeof(addr)))
return NULL;
if (addr % 0x1000)
return NULL;
if (dfi_mem_read_rc(addr, &os_info, sizeof(os_info)))
if (dfi_mem_virt_read(addr, &os_info, sizeof(os_info)))
return NULL;
if (os_info.magic != OS_INFO_MAGIC)
return NULL;
@@ -96,16 +100,16 @@ void dfi_vmcoreinfo_init(void)
addr = l.os_info->vmcoreinfo_addr;
size = l.os_info->vmcoreinfo_size;
} else {
if (dfi_mem_read_rc(LC_VMCORE_INFO, &addr, sizeof(addr)))
if (dfi_mem_virt_read(LC_VMCORE_INFO, &addr, sizeof(addr)))
return;
if (addr == 0)
return;
if (dfi_mem_read_rc(addr, &note, sizeof(note)))
if (dfi_mem_virt_read(addr, &note, sizeof(note)))
return;
if (note.n_namesz == 0 || note.n_namesz > sizeof(str))
return;
memset(str, 0, sizeof(str));
if (dfi_mem_read_rc(addr + sizeof(note), str, note.n_namesz))
if (dfi_mem_virt_read(addr + sizeof(note), str, note.n_namesz))
return;
if (memcmp(str, "VMCOREINFO", sizeof("VMCOREINFO")) != 0)
return;
@@ -116,7 +120,7 @@ void dfi_vmcoreinfo_init(void)
"DFI vmcoreinfo addr 0x%016lx size 0x%016lx\n",
addr, size);
l.vmcoreinfo = zg_alloc(size + 1);
if (dfi_mem_read_rc(addr, l.vmcoreinfo, size)) {
if (dfi_mem_virt_read(addr, l.vmcoreinfo, size)) {
zg_free(l.vmcoreinfo);
l.vmcoreinfo = NULL;
return;

20
zdump/dfi_vmcoreinfo.h Normal file
View File

@@ -0,0 +1,20 @@
/*
* Copyright IBM Corp. 2001, 2018
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef DFI_VMCOREINFO_H
#define DFI_VMCOREINFO_H
void dfi_vmcoreinfo_init(void);
char *dfi_vmcoreinfo_get(void);
int dfi_vmcoreinfo_tag(char *str, int len, const char *sym);
int dfi_vmcoreinfo_symbol(unsigned long *val, const char *sym);
int dfi_vmcoreinfo_offset(unsigned long *offs, const char *sym);
int dfi_vmcoreinfo_size(unsigned long *size, const char *sym);
int dfi_vmcoreinfo_length(unsigned long *len, const char *sym);
int dfi_vmcoreinfo_val(unsigned long *val, const char *sym);
#endif /* DFI_VMCOREINFO_H */

View File

@@ -10,7 +10,11 @@
*/
#include <time.h>
#include "zgetdump.h"
#include "lib/util_list.h"
#include "dfi_mem_chunk.h"
#include "dfo.h"
#define dfo_chunk_iterate(dfo_chunk) \
util_list_iterate(&l.dump.chunk_list, dfo_chunk)

View File

@@ -52,4 +52,10 @@ struct dfo {
void (*init)(void);
};
/*
* Supported DFO dump formats
*/
extern struct dfo dfo_s390;
extern struct dfo dfo_elf;
#endif /* DFO_H */

View File

@@ -16,7 +16,11 @@
#include <string.h>
#include <unistd.h>
#include "zgetdump.h"
#include "df_elf.h"
#include "dfi.h"
#include "dfi_mem_chunk.h"
#include "dfi_vmcoreinfo.h"
#include "dfo.h"
#define HDR_PER_CPU_SIZE 0x4a0
#define HDR_PER_MEMC_SIZE 0x100

View File

@@ -17,7 +17,9 @@
#include <time.h>
#include <unistd.h>
#include "zgetdump.h"
#include "df_s390.h"
#include "dfi_mem_chunk.h"
#include "dfo.h"
/*
* File local static data
@@ -98,7 +100,8 @@ static void dfo_s390_dump_chunk_lc_fn(struct dfo_chunk *dump_chunk,
struct dfi_cpu *cpu = dump_chunk->data;
char lc[0x2000];
dfi_mem_read(cpu->prefix + off, &lc[off], cnt);
if (dfi_mem_virt_read(cpu->prefix + off, &lc[off], cnt))
return;
if (dfi_arch() == DFI_ARCH_64)
cpu2lc_64(lc, cpu);
else
@@ -142,7 +145,8 @@ static void add_cpu_to_dfo(struct dfi_cpu *cpu)
return;
if (!dfi_cpu_content_fac_check(DFI_CPU_CONTENT_FAC_VX))
return;
dfi_mem_read(cpu->prefix, &lc, sizeof(lc));
if (dfi_mem_virt_read(cpu->prefix, &lc, sizeof(lc)))
return;
if (!dfi_cpu_lc_has_vx_sa(&lc))
return;
vx_regs = zg_alloc(DFI_VX_SA_SIZE);
@@ -181,8 +185,8 @@ static void dump_chunks_init(void)
dfi_cpu_iterate(cpu)
add_cpu_to_dfo(cpu);
dfo_chunk_add(dfi_mem_range() + DF_S390_HDR_SIZE,
DF_S390_EM_SIZE,
&l.em, dfo_chunk_buf_fn);
DF_S390_EM_SIZE,
&l.em, dfo_chunk_buf_fn);
}
/*

View File

@@ -10,6 +10,8 @@
*/
#include "zgetdump.h"
#include "zg.h"
#include "dt.h"
/*
* Supported dump tools

View File

@@ -12,9 +12,10 @@
#ifndef DT_H
#define DT_H
#include "dfi.h"
#include <stdbool.h>
#include "dfi.h"
#define DUMP_EXTENDED true
#define DUMP_NON_EXTENDED false
@@ -24,12 +25,21 @@ struct dt {
void (*info)(void);
};
extern void dt_init(void);
extern void dt_info_print(void);
extern void dt_arch_set(enum dfi_arch arch);
extern void dt_version_set(int version);
extern void dt_attr_mem_limit_set(u64 mem_limit);
extern void dt_attr_force_set(int value);
extern void dt_attr_dasd_type_set(const char *dasd_type);
void dt_init(void);
void dt_info_print(void);
void dt_arch_set(enum dfi_arch arch);
void dt_version_set(int version);
void dt_attr_mem_limit_set(u64 mem_limit);
void dt_attr_force_set(int value);
void dt_attr_dasd_type_set(const char *dasd_type);
/*
* Supported s390 dumpers
*/
extern struct dt dt_s390mv;
extern struct dt dt_s390mv_ext;
extern struct dt dt_s390sv;
extern struct dt dt_s390sv_ext;
extern struct dt dt_scsi;
#endif

View File

@@ -9,7 +9,8 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include "zgetdump.h"
#include "dt.h"
#include "df_s390.h"
/*
* Initialize s390 multi-volume dump tool (for -d option)

View File

@@ -9,7 +9,8 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include "zgetdump.h"
#include "dt.h"
#include "df_s390.h"
/*
* Initialize s390 multi-volume dump tool (extedend) for -d option

View File

@@ -9,8 +9,13 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <errno.h>
#include <linux/fs.h>
#include "zgetdump.h"
#include "zg.h"
#include "dt.h"
#include "df_s390.h"
/*
* File local static data

View File

@@ -10,7 +10,9 @@
*/
#include <linux/fs.h>
#include "zgetdump.h"
#include "dt.h"
#include "df_s390.h"
/*
* Initialize s390 single-volume extended dump tool (for -d option)

View File

@@ -9,13 +9,16 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include <linux/fs.h>
#include <stdio.h>
#include <stdint.h>
#include <linux/fs.h>
#include "lib/zt_common.h"
#include "lib/util_part.h"
#include "zgetdump.h"
#include "zg.h"
#include "dt.h"
/*
* Single volume SCSI dump superblock

View File

@@ -16,42 +16,59 @@
#include "lib/zt_common.h"
#include "lib/util_log.h"
#include "zgetdump.h"
#include "opts.h"
static struct option long_opts[] = {
{"help", no_argument, NULL, 'h'},
{"version", no_argument, NULL, 'v'},
{"info", no_argument, NULL, 'i'},
{"device", no_argument, NULL, 'd'},
{"mount", no_argument, NULL, 'm'},
{"umount", no_argument, NULL, 'u'},
{"fmt", required_argument, NULL, 'f'},
{"select", required_argument, NULL, 's'},
{"debug", no_argument, NULL, 'X'},
{"verbose", no_argument, NULL, 'V'},
{NULL, 0, NULL, 0 }
};
static const char optstr[] = "hvVidmus:f:X";
/*
* Text for --help option
*/
static char help_text[] =
"Usage: zgetdump DUMP [-s SYS] [-f FMT] > DUMP_FILE\n"
" -m DUMP [-s SYS] [-f FMT] DIR\n"
" -i DUMP [-s SYS]\n"
" -d DUMPDEV\n"
" -u DIR\n"
"\n"
"The zgetdump tool can read different dump formats from a dump device or from\n"
"a dump file. You can use zgetdump to:\n"
"\n"
" - Write the dump content to standard output or to a file\n"
" - Mount the dump content to a Linux directory\n"
" - Convert a dump to a different dump format\n"
" - Check if a dump is valid\n"
" - Check if a DASD contains a valid dump tool.\n"
"\n"
"In the syntax description, DUMP specifies a dump device or dump file to be\n"
"read. The following options are available:\n"
"\n"
"-m, --mount Mount DUMP to mount point DIR\n"
"-u, --umount Unmount dump from mount point DIR\n"
"-i, --info Print DUMP information\n"
"-f, --fmt Specify target dump format FMT (\"elf\" or \"s390\")\n"
"-s, --select Select system data SYS (\"kdump\", \"prod\", or \"all\")\n"
"-d, --device Print DUMPDEV (dump device) information\n"
"-v, --version Print version information, then exit\n"
"-V, --verbose Print verbose messages to stdout. Repeat this option\n"
" for increased verbosity from just error messages to\n"
" also include warning, information, debug, and trace\n"
" messages. This option is intended for debugging\n"
"-h, --help Print this help, then exit\n";
static const char help_text[] =
"Usage: zgetdump DUMP [-s SYS] [-f FMT] > DUMP_FILE\n"
" -m DUMP [-s SYS] [-f FMT] DIR\n"
" -i DUMP [-s SYS]\n"
" -d DUMPDEV\n"
" -u DIR\n"
"\n"
"The zgetdump tool can read different dump formats from a dump device or from\n"
"a dump file. You can use zgetdump to:\n"
"\n"
" - Write the dump content to standard output or to a file\n"
" - Mount the dump content to a Linux directory\n"
" - Convert a dump to a different dump format\n"
" - Check if a dump is valid\n"
" - Check if a DASD contains a valid dump tool.\n"
"\n"
"In the syntax description, DUMP specifies a dump device or dump file to be\n"
"read. The following options are available:\n"
"\n"
"-m, --mount Mount DUMP to mount point DIR\n"
"-u, --umount Unmount dump from mount point DIR\n"
"-i, --info Print DUMP information\n"
"-f, --fmt Specify target dump format FMT (\"elf\" or \"s390\")\n"
"-s, --select Select system data SYS (\"kdump\", \"prod\", or \"all\")\n"
"-d, --device Print DUMPDEV (dump device) information\n"
"-v, --version Print version information, then exit\n"
"-V, --verbose Print verbose messages to stdout. Repeat this option\n"
" for increased verbosity from just error messages to\n"
" also include warning, information, debug, and trace\n"
" messages. This option is intended for debugging\n"
"-h, --help Print this help, then exit\n";
static const char copyright_str[] = "Copyright IBM Corp. 2001, 2018";
@@ -65,27 +82,26 @@ const char *OPTS_SELECT_ALL = "all";
/*
* Initialize default settings
*/
static void init_defaults(void)
static void init_defaults(struct options *opts)
{
g.prog_name = "zgetdump";
g.opts.action = ZG_ACTION_STDOUT;
opts->prog_name = "zgetdump";
opts->action = ZG_ACTION_COPY;
#ifdef __s390x__
g.opts.fmt = "elf";
opts->fmt = "elf";
#else
g.opts.fmt = "s390";
opts->fmt = "s390";
#endif
dfo_set(g.opts.fmt);
/* Verbose logging */
g.opts.verbose = UTIL_LOG_ERROR;
util_log_set_level(g.opts.verbose);
opts->verbose = UTIL_LOG_ERROR;
util_log_set_level(opts->verbose);
}
/*
* Print "help" hint
*/
static void __noreturn print_usage_exit(void)
static void __noreturn print_usage_exit(const char *prog_name)
{
STDERR("Try '%s --help' for more information.\n", g.prog_name);
STDERR("Try '%s --help' for more information.\n", prog_name);
zg_exit(1);
}
@@ -101,10 +117,10 @@ static void __noreturn print_help_exit(void)
/*
* Print version information
*/
static void __noreturn print_version_exit(void)
static void __noreturn print_version_exit(const char *prog_name)
{
STDOUT("%s: Tool for copying and converting dumps version %s\n",
g.prog_name, RELEASE_STRING);
prog_name, RELEASE_STRING);
STDOUT("%s\n", copyright_str);
zg_exit(0);
}
@@ -112,96 +128,94 @@ static void __noreturn print_version_exit(void)
/*
* Set "--fmt" option
*/
static void fmt_set(const char *fmt)
static void fmt_set(struct options *opts, const char *fmt)
{
if (dfo_set(fmt) != 0)
ERR_EXIT("Invalid target format \"%s\" specified", fmt);
g.opts.fmt_specified = 1;
g.opts.fmt = fmt;
opts->fmt_specified = 1;
opts->fmt = fmt;
}
/*
* Set "--select" option
*/
static void select_set(const char *select)
static void select_set(struct options *opts, const char *select)
{
if (strcmp(select, OPTS_SELECT_KDUMP) == 0)
g.opts.select = OPTS_SELECT_KDUMP;
opts->select = OPTS_SELECT_KDUMP;
else if (strcmp(select, OPTS_SELECT_PROD) == 0)
g.opts.select = OPTS_SELECT_PROD;
opts->select = OPTS_SELECT_PROD;
else if (strcmp(select, OPTS_SELECT_ALL) == 0)
g.opts.select = OPTS_SELECT_ALL;
opts->select = OPTS_SELECT_ALL;
else
ERR_EXIT("Invalid select argument \"%s\" specified", select);
g.opts.select_specified = 1;
opts->select_specified = 1;
}
/*
* Set mount point
*/
static void mount_point_set(const char *mount_point)
static void mount_point_set(struct options *opts, const char *mount_point)
{
g.opts.mount_point = zg_strdup(mount_point);
opts->mount_point = zg_strdup(mount_point);
}
/*
* Set device
*/
static void device_set(const char *path)
static void device_set(struct options *opts, const char *path)
{
g.opts.device = zg_strdup(path);
opts->device = zg_strdup(path);
}
/*
* Set FUSE debug options
*/
static void argv_fuse_set(char **argv, int argc)
static void argv_fuse_set(struct options *opts, char **argv, int argc)
{
int i;
g.opts.argv_fuse = argv;
g.opts.argc_fuse = argc;
opts->argv_fuse = argv;
opts->argc_fuse = argc;
STDERR_PR("Fuse Options: ");
for (i = 0; i < argc; i++)
STDERR("%s ", g.opts.argv_fuse[i]);
STDERR("%s ", opts->argv_fuse[i]);
STDERR("\n");
}
/*
* Set action
*/
static void action_set(enum zg_action action)
static void action_set(struct options *opts, enum zg_action action)
{
if (g.opts.action_specified)
if (opts->action_specified)
ERR_EXIT("Please specify only one of the \"-i\", \"-d\", "
"\"-m\" or \"-u\" option");
g.opts.action = action;
g.opts.action_specified = 1;
opts->action = action;
opts->action_specified = 1;
}
/*
* Verify option combinations
*/
static void verify_opts(void)
static void verify_opts(struct options *opts)
{
if (g.opts.select_specified) {
if (g.opts.action != ZG_ACTION_MOUNT &&
g.opts.action != ZG_ACTION_STDOUT &&
g.opts.action != ZG_ACTION_DUMP_INFO)
if (opts->select_specified) {
if (opts->action != ZG_ACTION_MOUNT &&
opts->action != ZG_ACTION_COPY &&
opts->action != ZG_ACTION_DUMP_INFO)
ERR_EXIT("The \"--select\" option can only be "
"specified for info, mount, or copy");
}
if (!g.opts.fmt_specified)
if (!opts->fmt_specified)
return;
if (g.opts.action == ZG_ACTION_DUMP_INFO)
if (opts->action == ZG_ACTION_DUMP_INFO)
ERR_EXIT("The \"--fmt\" option cannot be specified "
"together with \"--info\"");
if (g.opts.action == ZG_ACTION_DEVICE_INFO)
if (opts->action == ZG_ACTION_DEVICE_INFO)
ERR_EXIT("The \"--fmt\" option cannot be specified "
"together with \"--device\"");
if (g.opts.action == ZG_ACTION_UMOUNT)
if (opts->action == ZG_ACTION_UMOUNT)
ERR_EXIT("The \"--fmt\" option cannot be specified "
"together with \"--umount\"");
}
@@ -209,36 +223,38 @@ static void verify_opts(void)
/*
* Parse positional arguments
*/
static void parse_pos_args(char *argv[], int argc)
static void parse_pos_args(struct options *opts, char *argv[], int argc)
{
int pos_args = argc - optind;
switch (g.opts.action) {
case ZG_ACTION_STDOUT:
switch (opts->action) {
case ZG_ACTION_COPY:
case ZG_ACTION_DUMP_INFO:
case ZG_ACTION_DEVICE_INFO:
if (pos_args == 0)
ERR_EXIT("No device or dump specified");
if (pos_args > 1 && !g.opts.debug_specified)
if (pos_args > 1)
ERR_EXIT("Too many positional parameters specified");
device_set(argv[optind]);
device_set(opts, argv[optind]);
break;
case ZG_ACTION_MOUNT:
if (pos_args == 0)
ERR_EXIT("No dump specified");
if (pos_args == 1)
ERR_EXIT("No mount point specified");
if (pos_args > 2 && !g.opts.debug_specified)
if (pos_args > 2 && !opts->debug_specified)
ERR_EXIT("Too many positional parameters specified");
device_set(argv[optind]);
mount_point_set(argv[optind + 1]);
if (g.opts.debug_specified && pos_args > 2)
argv_fuse_set(&argv[optind + 2], pos_args - 2);
device_set(opts, argv[optind]);
mount_point_set(opts, argv[optind + 1]);
if (opts->debug_specified && pos_args > 2)
argv_fuse_set(opts, &argv[optind + 2], pos_args - 2);
break;
case ZG_ACTION_UMOUNT:
if (pos_args == 0)
ERR_EXIT("No mount point specified");
mount_point_set(argv[optind]);
if (pos_args > 1)
ERR_EXIT("Too many positional parameters specified");
mount_point_set(opts, argv[optind]);
break;
}
}
@@ -246,60 +262,46 @@ static void parse_pos_args(char *argv[], int argc)
/*
* Main command line parsing function
*/
void opts_parse(int argc, char *argv[])
void opts_parse(int argc, char *argv[], struct options *opts)
{
int opt, idx;
static struct option long_opts[] = {
{"help", no_argument, NULL, 'h'},
{"version", no_argument, NULL, 'v'},
{"info", no_argument, NULL, 'i'},
{"device", no_argument, NULL, 'd'},
{"mount", no_argument, NULL, 'm'},
{"umount", no_argument, NULL, 'u'},
{"fmt", required_argument, NULL, 'f'},
{"select", required_argument, NULL, 's'},
{"debug", no_argument, NULL, 'X'},
{"verbose", no_argument, NULL, 'V'},
{NULL, 0, NULL, 0 }
};
static const char optstr[] = "hvVidmus:f:X";
init_defaults();
init_defaults(opts);
while ((opt = getopt_long(argc, argv, optstr, long_opts, &idx)) != -1) {
switch (opt) {
case 'h':
print_help_exit();
case 'v':
print_version_exit();
print_version_exit(opts->prog_name);
case 'V':
g.opts.verbose++;
util_log_set_level(g.opts.verbose);
opts->verbose++;
util_log_set_level(opts->verbose);
break;
case 'i':
action_set(ZG_ACTION_DUMP_INFO);
action_set(opts, ZG_ACTION_DUMP_INFO);
break;
case 'd':
action_set(ZG_ACTION_DEVICE_INFO);
action_set(opts, ZG_ACTION_DEVICE_INFO);
break;
case 'm':
action_set(ZG_ACTION_MOUNT);
action_set(opts, ZG_ACTION_MOUNT);
break;
case 'u':
action_set(ZG_ACTION_UMOUNT);
action_set(opts, ZG_ACTION_UMOUNT);
break;
case 'f':
fmt_set(optarg);
fmt_set(opts, optarg);
break;
case 's':
select_set(optarg);
select_set(opts, optarg);
break;
case 'X':
g.opts.debug_specified = 1;
opts->debug_specified = 1;
break;
default:
print_usage_exit();
print_usage_exit(opts->prog_name);
}
}
parse_pos_args(argv, argc);
verify_opts();
parse_pos_args(opts, argv, argc);
verify_opts(opts);
}

38
zdump/opts.h Normal file
View File

@@ -0,0 +1,38 @@
/*
* Copyright IBM Corp. 2001, 2018
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef OPTS_H
#define OPTS_H
#include "zg.h"
/*
* zgetdump options
*/
struct options {
const char *prog_name;
int action_specified;
enum zg_action action;
char *device;
char *mount_point;
int fmt_specified;
const char *fmt;
int debug_specified;
char **argv_fuse;
int argc_fuse;
const char *select;
int select_specified;
int verbose;
};
extern const char *OPTS_SELECT_KDUMP;
extern const char *OPTS_SELECT_PROD;
extern const char *OPTS_SELECT_ALL;
void opts_parse(int argc, char *argv[], struct options *opts);
#endif /* OPTS_H */

View File

@@ -9,13 +9,16 @@
* it under the terms of the MIT license. See LICENSE for details.
*/
#include "zgetdump.h"
#include "zg.h"
#include "dfi.h"
#include "dfo.h"
#include "stdout.h"
int stdout_write_dump(void)
{
u64 cnt, written = 0;
char buf[32768];
ssize_t rc;
const u64 output_size = dfo_size();
char buf[8UL * PAGE_SIZE];
u64 written = 0;
if (!dfi_feat_copy())
ERR_EXIT("Copying not possible for %s dumps", dfi_name());
@@ -23,8 +26,11 @@ int stdout_write_dump(void)
STDERR(" Source: %s\n", dfi_name());
STDERR(" Target: %s\n", dfo_name());
STDERR("\n");
zg_progress_init("Copying dump", dfo_size());
do {
zg_progress_init("Copying dump", output_size);
while (written != output_size) {
ssize_t rc;
u64 cnt;
cnt = dfo_read(buf, sizeof(buf));
rc = write(STDOUT_FILENO, buf, cnt);
if (rc == -1)
@@ -33,7 +39,7 @@ int stdout_write_dump(void)
ERR_EXIT("Error: Could not write full block");
written += cnt;
zg_progress(written);
} while (written != dfo_size());
};
STDERR("\n");
STDERR("Success: Dump has been copied\n");
return 0;

13
zdump/stdout.h Normal file
View File

@@ -0,0 +1,13 @@
/*
* Copyright IBM Corp. 2001, 2018
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef STDOUT_H
#define STDOUT_H
int stdout_write_dump(void);
#endif /* STDOUT_H */

View File

@@ -16,6 +16,7 @@
#include <fuse.h>
#include <limits.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <sys/stat.h>
#include <sys/types.h>
@@ -23,6 +24,10 @@
#include <unistd.h>
#include "zgetdump.h"
#include "zg.h"
#include "dfi.h"
#include "dfo.h"
#include "zfuse.h"
#define DUMP_PATH_MAX 100

25
zdump/zfuse.h Normal file
View File

@@ -0,0 +1,25 @@
/*
* Copyright IBM Corp. 2001, 2018
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
*/
#ifndef ZFUSE_H
#define ZFUSE_H
#if HAVE_FUSE == 0
static inline int zfuse_mount_dump(void)
{
ERR_EXIT("Program compiled without fuse support");
}
static inline void zfuse_umount(void)
{
ERR_EXIT("Program compiled without fuse support");
}
#else
int zfuse_mount_dump(void);
void zfuse_umount(void);
#endif
#endif /* ZFUSE_H */

View File

@@ -188,7 +188,7 @@ extern enum zg_type zg_type(struct zg_fh *zg_fh);
* zgetdump actions
*/
enum zg_action {
ZG_ACTION_STDOUT,
ZG_ACTION_COPY,
ZG_ACTION_DUMP_INFO,
ZG_ACTION_DEVICE_INFO,
ZG_ACTION_MOUNT,

View File

@@ -27,6 +27,11 @@
#include <unistd.h>
#include "zgetdump.h"
#include "dt.h"
#include "dfi.h"
#include "dfo.h"
#include "stdout.h"
#include "zfuse.h"
/*
* Globals
@@ -154,7 +159,7 @@ static int do_mount(void)
/*
* Run "copy to stdout" action
*/
static int do_stdout(void)
static int do_copy(void)
{
int rc;
@@ -173,11 +178,14 @@ static int do_stdout(void)
int main(int argc, char *argv[])
{
sig_handler_init();
opts_parse(argc, argv);
opts_parse(argc, argv, &g.opts);
if (dfo_set(g.opts.fmt) != 0)
ERR_EXIT("Invalid target format \"%s\" specified", g.opts.fmt);
switch (g.opts.action) {
case ZG_ACTION_STDOUT:
return do_stdout();
case ZG_ACTION_COPY:
return do_copy();
case ZG_ACTION_DUMP_INFO:
return do_dump_info();
case ZG_ACTION_DEVICE_INFO:

View File

@@ -1,8 +1,6 @@
/*
* zgetdump - Tool for copying and converting System z dumps
*
* Main include file - Should be included by all source files
*
* Copyright IBM Corp. 2001, 2018
*
* s390-tools is free software; you can redistribute it and/or modify
@@ -12,92 +10,17 @@
#ifndef ZGETDUMP_H
#define ZGETDUMP_H
#include "df_elf.h"
#include "df_lkcd.h"
#include "df_s390.h"
#include "dfi.h"
#include "dfo.h"
#include "dt.h"
#include "zg.h"
/*
* zgetdump options
*/
struct options {
int action_specified;
enum zg_action action;
char *device;
char *mount_point;
int fmt_specified;
const char *fmt;
int debug_specified;
char **argv_fuse;
int argc_fuse;
const char *select;
int select_specified;
int verbose;
};
extern const char *OPTS_SELECT_KDUMP;
extern const char *OPTS_SELECT_PROD;
extern const char *OPTS_SELECT_ALL;
#include "opts.h"
/*
* zgetdump globals
*/
extern struct zgetdump_globals {
struct zgetdump_globals {
struct zg_fh *fh;
const char *prog_name;
struct options opts;
} g;
};
/*
* Misc fuctions
*/
extern void opts_parse(int argc, char *argv[]);
extern int stdout_write_dump(void);
#if HAVE_FUSE == 0
static inline int zfuse_mount_dump(void)
{
ERR_EXIT("Program compiled without fuse support");
}
static inline void zfuse_umount(void)
{
ERR_EXIT("Program compiled without fuse support");
}
#else
int zfuse_mount_dump(void);
void zfuse_umount(void);
#endif
/*
* Supported DFI dump formats
*/
extern struct dfi dfi_s390tape;
extern struct dfi dfi_s390mv;
extern struct dfi dfi_s390mv_ext;
extern struct dfi dfi_s390;
extern struct dfi dfi_s390_ext;
extern struct dfi dfi_lkcd;
extern struct dfi dfi_elf;
extern struct dfi dfi_kdump;
extern struct dfi dfi_kdump_flat;
extern struct dfi dfi_devmem;
/*
* Supported DFO dump formats
*/
extern struct dfo dfo_s390;
extern struct dfo dfo_elf;
/*
* Supported s390 dumpers
*/
extern struct dt dt_s390mv;
extern struct dt dt_s390mv_ext;
extern struct dt dt_s390sv;
extern struct dt dt_s390sv_ext;
extern struct dt dt_scsi;
extern struct zgetdump_globals g;
#endif /* ZGETDUMP_H */

View File

@@ -3,7 +3,7 @@
*
* Utility classes to print framsets
*
* Copyright IBM Corp. 2008, 2017
* Copyright IBM Corp. 2008, 2021
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
@@ -478,21 +478,22 @@ void VirtAdapterPrinter::print_failures(FILE *fp, const struct ioerr_cnt *cnt)
}
void VirtAdapterPrinter::print_throughput(FILE *fp,
const struct blkiomon_stat *stat)
const struct blkiomon_stat *stat,
const __u64 interval)
{
double tmp;
if (!stat || stat->d2c_r.num <= 0 || stat->size_r.num <= 0)
if (!stat || interval == 0 || stat->size_r.num <= 0)
tmp = 0;
else
tmp = stat->size_r.sum/(double)stat->d2c_r.sum;
tmp = calc_avg(stat->size_r.sum, interval);
print_delimiter(fp);
print_abbrev_num(fp, tmp);
if (!stat || stat->d2c_w.num <= 0 || stat->size_w.num <= 0)
if (!stat || interval == 0 || stat->size_w.num <= 0)
tmp = 0;
else
tmp = stat->size_w.sum/(double)stat->d2c_w.sum;
tmp = calc_avg(stat->size_w.sum, interval);
print_delimiter(fp);
print_abbrev_num(fp, tmp);
}
@@ -536,6 +537,12 @@ int VirtAdapterPrinter::print_frame(FILE *fp,
const struct blkiomon_stat *blk_stat;
const struct zfcpdd_dstat *zfcp_stat;
/**
* Receive Frameset interval and convert it
* from seconds to microseconds
*/
__u64 interval = frameset.get_duration() * 1000000;
list<__u32> devnos;
devnos = ((StagedDeviceFilter*)&dev_filt)->get_filter_devnos();
@@ -557,7 +564,7 @@ int VirtAdapterPrinter::print_frame(FILE *fp,
print_queue_fill(fp, zfcp_stat, util);
print_queue_full(fp, util);
print_failures(fp, ioerr);
print_throughput(fp, blk_stat);
print_throughput(fp, blk_stat, interval);
print_num_requests(fp, blk_stat);
if (lrc) {
fprintf(stderr, "%s: Did not find matching data in"
@@ -810,6 +817,12 @@ int TrafficPrinter::print_frame(FILE *fp, const Frameset &frameset,
const struct zfcpdd_dstat *zfcp_stat = NULL;
const AggregationCollapser *agg_col;
/**
* Receive Frameset interval and convert it
* from seconds to microseconds
*/
__u64 interval = frameset.get_duration() * 1000000;
switch (m_agg_crit) {
case none:
get_device_list(lst_32, dev_filt);
@@ -847,7 +860,7 @@ int TrafficPrinter::print_frame(FILE *fp, const Frameset &frameset,
blk_stat = frameset.get_blkiomon_stat_by_wwpn(*i);
zfcp_stat = frameset.get_zfcpdd_stat_by_wwpn(*i);
print_device_wwpn(fp, *i);
print_data_row(fp, blk_stat, zfcp_stat);
print_data_row(fp, blk_stat, zfcp_stat, interval);
}
}
else if (m_agg_crit == all) {
@@ -856,7 +869,7 @@ int TrafficPrinter::print_frame(FILE *fp, const Frameset &frameset,
blk_stat = frameset.get_first_blkiomon_stat();
zfcp_stat = frameset.get_first_zfcpdd_stat();
print_device_all(fp);
print_data_row(fp, blk_stat, zfcp_stat);
print_data_row(fp, blk_stat, zfcp_stat, interval);
}
else {
for (list<__u32>::const_iterator i = lst_32.begin();
@@ -894,7 +907,7 @@ int TrafficPrinter::print_frame(FILE *fp, const Frameset &frameset,
}
if (rc )
return -1;
print_data_row(fp, blk_stat, zfcp_stat);
print_data_row(fp, blk_stat, zfcp_stat, interval);
}
}
@@ -930,7 +943,8 @@ void SummaryTrafficPrinter::print_topline(FILE *fp)
}
void SummaryTrafficPrinter::print_throughput(FILE *fp, const struct blkiomon_stat *stat)
void SummaryTrafficPrinter::print_throughput(FILE *fp, const struct blkiomon_stat *stat,
const __u64 interval)
{
struct minmax thrp_data, total_size, total_latency;
double tmp;
@@ -968,7 +982,7 @@ void SummaryTrafficPrinter::print_throughput(FILE *fp, const struct blkiomon_sta
tmp = 0;
if (stat && total_size.sum > 0)
tmp = calc_avg(total_size.sum, total_latency.sum);
tmp = calc_avg(total_size.sum, interval);
print_delimiter(fp);
print_abbrev_num(fp, tmp);
@@ -1101,14 +1115,15 @@ void SummaryTrafficPrinter::print_fabric_latency(FILE *fp, const struct zfcpdd_d
void SummaryTrafficPrinter::print_data_row(FILE *fp,
const struct blkiomon_stat *blk_stat,
const struct zfcpdd_dstat *zfcp_stat)
const struct zfcpdd_dstat *zfcp_stat,
const __u64 interval)
{
if (!blk_stat)
blk_stat = get_empty_blkiomon_stat();
if (!zfcp_stat)
zfcp_stat = get_empty_zfcpdd_dstat();
print_throughput(fp, blk_stat);
print_throughput(fp, blk_stat, interval);
print_request_stats(fp, blk_stat);
print_io_subsystem_latency(fp, blk_stat);
print_channel_latency(fp, zfcp_stat);
@@ -1177,7 +1192,8 @@ void DetailedTrafficPrinter::print_topline(FILE *fp)
void DetailedTrafficPrinter::print_data_row(FILE *fp,
const struct blkiomon_stat *blk_stat,
const struct zfcpdd_dstat *zfcp_stat)
const struct zfcpdd_dstat *zfcp_stat,
const __u64 interval __attribute__ ((unused)))
{
if (!blk_stat)
blk_stat = get_empty_blkiomon_stat();

View File

@@ -3,7 +3,7 @@
*
* Utility classes to print framsets
*
* Copyright IBM Corp. 2008, 2017
* Copyright IBM Corp. 2008, 2021
*
* s390-tools is free software; you can redistribute it and/or modify
* it under the terms of the MIT license. See LICENSE for details.
@@ -128,7 +128,8 @@ private:
const struct adapter_utilization *res);
void print_queue_full(FILE *fp, const struct adapter_utilization *res);
void print_failures(FILE *fp, const struct ioerr_cnt *cnt);
void print_throughput(FILE *fp, const struct blkiomon_stat *stat);
void print_throughput(FILE *fp, const struct blkiomon_stat *stat,
const __u64 interval);
void print_num_requests(FILE *fp, const struct blkiomon_stat *stat);
};
@@ -147,7 +148,8 @@ protected:
* Print the actual row, excluding the first column */
virtual void print_data_row(FILE *fp,
const struct blkiomon_stat *blk_stat,
const struct zfcpdd_dstat *zfcp_stat) = 0;
const struct zfcpdd_dstat *zfcp_stat,
const __u64 interval) = 0;
void print_topline_prefix1(FILE *fp);
void print_topline_prefix2(FILE *fp);
@@ -185,8 +187,10 @@ public:
private:
virtual void print_data_row(FILE *fp,
const struct blkiomon_stat *blk_stat,
const struct zfcpdd_dstat *zfcp_stat);
void print_throughput(FILE *fp, const struct blkiomon_stat *stat);
const struct zfcpdd_dstat *zfcp_stat,
const __u64 interval);
void print_throughput(FILE *fp, const struct blkiomon_stat *stat,
const __u64 interval);
void print_request_stats(FILE *fp, const struct blkiomon_stat *stat);
void print_io_subsystem_latency(FILE *fp, const struct blkiomon_stat *stat);
void print_channel_latency(FILE *fp, const struct zfcpdd_dstat *stat);
@@ -207,7 +211,8 @@ public:
private:
virtual void print_data_row(FILE *fp,
const struct blkiomon_stat *blk_stat,
const struct zfcpdd_dstat *zfcp_stat);
const struct zfcpdd_dstat *zfcp_stat,
const __u64 interval);
void print_histogram_io_reqs(FILE *fp,
const struct blkiomon_stat *stat);
void print_histogram_io_subs_lat(FILE *fp,

View File

@@ -143,12 +143,13 @@ int menu(void)
int rc;
cmd_line_extra = (char *)COMMAND_LINE_EXTRA;
memset(cmd_line_extra, 0, COMMAND_LINE_EXTRA_SIZE);
rc = sclp_setup(SCLP_INIT);
if (rc)
/* sclp setup failed boot default */
goto boot;
memset(cmd_line_extra, 0, COMMAND_LINE_SIZE);
rc = menu_param(&value);
if (rc == 0) {
/* got number from loadparm, boot it */

View File

@@ -675,7 +675,8 @@ static int add_ipl_program(int fd, char *filename,
IMAGE_ENTRY,
(info->type == disk_type_scsi) ? 0 : 1,
flags, ipl->image_addr, image_size,
ipl->envblk_addr, envblk->size);
ipl->envblk_addr,
add_envblk ? envblk->size : 0);
if (rc) {
free(table);
return rc;