Previously when user tried to allocate to big bio he recieved NULL. As the
behavior was changed to `BUG()` passing out of range value is no longer legal.
Limiting the requested bio size to BIO_MAX_PAGES (on the older kernels) or to
BIO_MAX_VECS (on the newer ones) eliminates the problem as the macros represent
the max possible value.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
This commit fixes the case of using unintialized variable in error
handling path after failed core addition.
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
This check will always return 0, as none of kernel versions supported
by Open CAS sets a barrier flag in for bio. Those flags are exclusive to
struct request, which is no longer used by Open CAS.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
Rework for RHEL8.5 compatibility to avoid potential bug of exiting queue
after IO completion on a different CPU.
Targeted RHEL kernel version corrected.
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
This patch adds support for RHEL8.5 with kernel 4.18.0.348.2.1.el8 and
later.
Additional request queue operations added for these kernels.
Configure scripts added/updated.
Fixes#1278
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
In specific circumstances, IO could be completed twice.
This patch fixes this problem.
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
Null pointer dereference occurs when terminating cache is standby detached state.
This fixes this problem.
Fixes#1235
Signed-off-by: Krzysztof Majzerowicz-Jaszcz <krzysztof.majzerowicz-jaszcz@intel.com>
When dir is ignored, and 0 is passed instead, each flush request will
appear as READ request, which is not supported by some block device
drivers.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
One of the steps of discarding data in cache is invalidating OCF metadata.
If a cache line which is supposed to be discarded is dirty, invalidating
it will require flushing metadata. Unfortunately, OCF allocates flushing
requests with the exactly the same flags as the original IO (in this case
discard flag is set) so the page on the disk is discarded instead of being
flushed. In case of power failure occurring before the metadata is flushed
to the disk, the data may be corrupted even if recovery will succeed.
Disabling propagation of original I/O flags for discard requests solves
this problem.
Signed-off-by: Michal Mielewczyk <michal.mielewczyk@intel.com>
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
The bio_alloc_bioset() function now BUG() if trying to allocate a bio
with more than BIO_MAX_VECS vectors.
A no-limit value (-1) is defined in order not to change old kernels'
behaviour.
Signed-off-by: Gal Hammer <gal.hammer@huawei.com>
Signed-off-by: Shai Fultheim <shai.fultheim@huawei.com>
This is a suboptimal solution to CAS on top of MD RAID1 device. If using
only submit_bio API RAID1 would process all IOs in single thread.
Plugging bypasses this thread and processess IOs in blk_finish_plug
caller context improving performance drastically.
Testing showed no negative impact to other usecases and it's a thing
that Linux does in AIO, so it's vetted and proven to work.
Signed-off-by: Jan Musial <jan.musial@intel.com>
OCF cannot allocate request map bigger than 4MiB (due to kmalloc
limitations), thus we need to split bigger IOs into series smaller
ones to reduce request map size.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
It's defined on every single supported kernel, so there is actually no need
for this define at all.
Signed-off-by: Robert Baldyga <robert.baldyga@intel.com>
