The primary goal of this change is to split DFO ELF module into
multiple smaller ones for better unit testability.
The change doesn't introduce any functionality changes, just shuffling code
around.
Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Reviewed-by: Marc Hartmayer <mhartmay@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
The primary goal is to separate DFO code into multiple smaller modules
and make it unit testable.
This refactoring only moved the code around w/o changing any functionality.
Signed-off-by: Alexander Egorenkov <egorenar@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Without the fix, readers of syslog / systemd-journal can erroneously
think that dbginfo.sh was incomplete because the last logged step is
"17 of 18".
The fix also makes the missing line appear in dbginfo.log inside the
generated DBGINFO....tgz.
s390-tools v2.17.0 commit b7807d0195 ("dbginfo.sh: Add KVM commands
and rework data collection") turned create_package into a numbered
step.
Snippet of terminal output:
...
17 of 18: Postprocessing
18 of 18: Finalizing: Creating archive with collected data
Collected data was saved to:
>> /tmp/DBGINFO-....tgz <<
Please review all collected data before sending to your service organization.
Syslog before:
...
dbginfo.sh[...]: 17 of 18: Postprocessing
dbginfo.sh[...]: Data collection completed
Syslog after:
...
dbginfo.sh[...]: 17 of 18: Postprocessing
dbginfo.sh[...]: 18 of 18: Finalizing: Creating archive with collected data
dbginfo.sh[...]: Data collection completed
Fixes: b7807d0195 ("dbginfo.sh: Add KVM commands and rework data collection")
Signed-off-by: Steffen Maier <maier@linux.ibm.com>
Signed-off-by: Jan Höppner <hoeppner@linux.ibm.com>
Remove the call to `gzip` before installing the manpage during the
`make install` call. What and if compression is done should be handled by
the distribution tooling.
This also removes a dependency for the build process.
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>
`chreipl-fcp-mpath-common.sh.in` is never executed, only used as argument
for `source` in the udev helper scripts, so the shebang is unnecessary, and
might be confusing.
Also, tools like `rpmlint` from the rpm software management will complain
about this; e.g.:
s390-tools-chreipl-fcp-mpath.s390x: W: non-executable-script /usr/lib/chreipl-fcp-mpath/chreipl-fcp-mpath-common.sh 644 /bin/bash
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>
...to make clear why `off` is initialized to 0 and that all calls to
`dfo_chunk_add` occur for the same reason. For consistency reasons adapt
the `off` value calculation as well.
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>
Rename the parameter `ptr` of `notes_init` to something meaningful. Since
it points to the start of the notes segment rename it to `segment_start`.
In addition, improve the documentation for the function.
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>
Use `mem_chunk->size` instead of recalculate it. In addition, assign
`p_vaddr` to `p_paddr` to make clear that they must be equal and rename the
parameter `load_offset` to `elf_offset` to make clear that the offset in
the ELF vmcore file is meant. Finally rename `loads_init` to
`load_phdrs_init`.
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>
This cast makes the macros easier to use without getting compiler warnings.
e.g.
dfo_elf.c: In function 'dfo_elf_init':
dfo_elf.c:343:20: warning: assignment to 'Elf64_Phdr *' from incompatible pointer type 'char *' [-Wincompatible-pointer-types]
343 | phdrs_load = PTR_ADD(phdr_notes, sizeof(Elf64_Phdr) * phdrs_notes_count);
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
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>
