Hi,
since kernel 4.11 (sorry it took so long to report) I have a box failing to
boot with a NULL pointer dereference (the box is stuck there afterwards).
The bug has also been reported to the Debian BTS (https://bugs.debian.org/cgi-
bin/bugreport.cgi?bug=882414) and a suggestion to revert 9096576
Signed-off-by: Shivasharan S
---
drivers/scsi/megaraid/megaraid_sas_base.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/scsi/megaraid/megaraid_sas_base.c
b/drivers/scsi/megaraid/megaraid_sas_base.c
index 38120917c609..424858874d57 100644
--- a/drivers/scsi/megaraid/megaraid_sas_ba
Memory allocated for IOC_INIT command and stream detection array
are not zero'd before using. Use kzalloc instead of kmalloc to zero out
the memory allocated.
Signed-off-by: Sumit Saxena
Signed-off-by: Shivasharan S
---
drivers/scsi/megaraid/megaraid_sas_base.c | 2 +-
drivers/scsi/megaraid/m
Currently driver does not validate ldcount provided by firmware.
If the value is invalid, fail RAID map validation accordingly.
This issue is rare to hit in field and is fixed as part of code review.
Signed-off-by: Sumit Saxena
Signed-off-by: Shivasharan S
---
drivers/scsi/megaraid/megaraid_sas
Shivasharan S (14):
megaraid_sas: zero out IOC INIT and stream detection memory
megaraid_sas: memset IOC INIT frame using correct size
megaraid_sas: Return the DCMD status from megasas_get_seq_num
megaraid_sas: Reset ldio_outstanding in megasas_resume
megaraid_sas: Error handling for inva
In megasas_get_seq_num, the status of the DCMD fired to FW is
not returned, always success is returned.
We could end up registering AEN request with incorrect sequence
number if the DCMD failed.
Return the DCMD status back to caller.
This was discovered during code review and very rare to see issue
Driver needs to avoid PCI writes while OCR is in progress.
Use reset_mutex to synchronize between firing DCMDs MR_DCMD_PD_GET_INFO and
MR_DCMD_DRV_GET_TARGET_PROP while OCR is triggered.
Without this fix, if Device/VD add/creation is in progress and at the same
time MR Firmware is going through OCR
Issue - Driver returns DID_NO_CONNECT when unload is in progress, indicated
using instance->unload flag. In case of dynamic unload of driver, this flag
is set before calling scsi_remove_host(). While doing manual driver unload,
user will see lots of prints for Sync Cache command with DID_NO_CONNECT
Commit b9637d14dc00 ("scsi: megaraid_sas: Resize MFA frame used for
IOC INIT to 4k") increased the size of IOC INIT frame to 4k.
Need to use updated size when memsetting init_frame.
Signed-off-by: Shivasharan S
---
drivers/scsi/megaraid/megaraid_sas_fusion.c | 2 +-
1 file changed, 1 insertion(+
Performance improvement-
Current driver calls stream detection unconditionally for all IOs.
Stream Detection logic is not required for most of the fast path IO.
To improve performance, avoid stream detection logic and do it only if
required.
Below are the cases where stream detection is required in
Expose FW outstanding commands (fw_outstanding) through sysfs interface.
This helps in debugging certain performance issues in the field.
Signed-off-by: Sasikumar Chandrasekaran
Signed-off-by: Shivasharan S
---
drivers/scsi/megaraid/megaraid_sas_base.c | 13 +
1 file changed, 13 ins
Issue –
There may be some IO accessing incorrect raid map, but driver has checks in
IO path to handle those cases. It is always better to move to new raid map only
once raid map is populated and validated.
No functional defect. Fix is provided as part of review.
Fix –
Update instance->map_id after
In IOCTL path, re-use megasas_wait_for_adapter_operational API to detect
controller state. This will make driver to use this API uniformly in all cases
where we need to wait for adapter to become operational.
Signed-off-by: Sumit Saxena
Signed-off-by: Shivasharan S
---
drivers/scsi/megaraid/meg
Feature - NVME pass-through via MFI interface.
Current MR product support different encapsulation via MFI frame work.
NVME native command should be framed by application and it should be embedded
in MFI as payload.
Driver will provide interface to send MFI frame along with payload
(In this case, p
No functional changes.
This patch is a re-work of DCMD refire code to better manage
all the different cases to decide whether to REFIRE or SKIP
or COMPLETE certain DCMD.
Signed-off-by: Sumit Saxena
Signed-off-by: Shivasharan S
---
drivers/scsi/megaraid/megaraid_sas_base.c | 6 ++--
drivers/s
Signed-off-by: Shivasharan S
---
drivers/scsi/megaraid/megaraid_sas.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/megaraid/megaraid_sas.h
b/drivers/scsi/megaraid/megaraid_sas.h
index d557ff03eb1d..ba6503f37756 100644
--- a/drivers/scsi/megaraid/megaraid_s
Feature - NVME pass-through via MFI interface.
Current MR product support different encapsulation via MFI frame work.
NVME native command should be framed by application and it should be embedded
in MFI as payload.
Driver will provide interface to send MFI frame along with payload
(In this case, p
Feature - NVME pass-through via MFI interface.
Current MR product support different encapsulation via MFI frame work.
NVME native command should be framed by application and it should be embedded
in MFI as payload.
Driver will provide interface to send MFI frame along with payload
(In this case, p
From: Colin Ian King
A previous commit no longer stores the contents of c, so we now have
a situation where c is being updated but the value is never read. Clean
up the code by removing the now redundant setting of variable c.
Cleans up clang warning:
drivers/scsi/aacraid/aachba.c:943:3: warning
From: Colin Ian King
The initialization of d is redundant as this value is never read
and it is overwritten inside the subsequent for-loop. Remove this
redundant assignment.
Cleans up clang warning:
drivers/scsi/qla2xxx/qla_gs.c:3985:29: warning: Value stored to 'd'
during its initialization is
On Thu, Dec 21 2017, Damien Le Moal wrote:
> This series, formerly titled "scsi-mq support for ZBC disks", implements
> support for ZBC disks for system using the scsi-mq I/O path.
>
> The current scsi level support of ZBC disks guarantees write request ordering
> using a per-zone write lock which
Use the sgl_alloc() and sgl_free() functions instead of open coding
these functions.
Signed-off-by: Bart Van Assche
Acked-by: Ard Biesheuvel
Cc: Herbert Xu
---
crypto/Kconfig | 1 +
crypto/scompress.c | 51 ++-
2 files changed, 3 insertions(
Use the sgl_alloc() and sgl_free() functions instead of open coding
these functions.
Signed-off-by: Bart Van Assche
Reviewed-by: Johannes Thumshirn
Reviewed-by: Hannes Reinecke
Cc: Keith Busch
Cc: Christoph Hellwig
Cc: Sagi Grimberg
---
drivers/nvme/target/Kconfig | 1 +
drivers/nvme/targe
Use the sgl_alloc() and sgl_free() functions instead of open coding
these functions.
Signed-off-by: Bart Van Assche
Reviewed-by: Johannes Thumshirn
Reviewed-by: Hannes Reinecke
Cc: Keith Busch
Cc: Christoph Hellwig
Cc: James Smart
Cc: Sagi Grimberg
---
drivers/nvme/target/Kconfig | 1 +
d
Use the sgl_alloc_order() and sgl_free() functions instead of open
coding these functions.
Signed-off-by: Bart Van Assche
Acked-by: Nicholas A. Bellinger
Reviewed-by: Hannes Reinecke
Cc: Christoph Hellwig
Cc: Sagi Grimberg
---
drivers/target/Kconfig | 1 +
drivers/target/tar
Hello Jens,
As you know there are multiple drivers that both allocate a scatter/gather
list and populate that list with pages. This patch series moves the code for
allocating and freeing such scatterlists from several drivers into
lib/scatterlist.c. Please consider this patch series for kernel v4.
Many kernel drivers contain code that allocates and frees both a
scatterlist and the pages that populate that scatterlist.
Introduce functions in lib/scatterlist.c that perform these tasks
instead of duplicating this functionality in multiple drivers.
Only include these functions in the build if CO
On Thu, 2018-01-04 at 14:32 -0800, Vinson Lee wrote:
> HP ProLiant DL360p Gen8 with Smart Array P420i boots to the login
> prompt and hangs with Linux 4.13 or later. I cannot log in on console
> or SSH into the machine. Linux 4.12 and older boot fine.
>
> I see these messages on the console.
>
>
On 1/5/2018 8:26 AM, Bart Van Assche wrote:
Use the sgl_alloc() and sgl_free() functions instead of open coding
these functions.
Signed-off-by: Bart Van Assche
Reviewed-by: Johannes Thumshirn
Reviewed-by: Hannes Reinecke
Cc: Keith Busch
Cc: Christoph Hellwig
Cc: James Smart
Cc: Sagi Grimbe
> -Original Message-
> From: Colin King [mailto:colin.k...@canonical.com]
> Sent: Friday, January 5, 2018 7:31 AM
> To: dl-esc-Aacraid Linux Driver ; James E . J .
> Bottomley ; Martin K . Petersen
> ; linux-scsi@vger.kernel.org
> Cc: kernel-janit...@vger.kernel.org; linux-ker...@vger.kern
Hello Martin,
Recently it was discovered that the v4.15 rework of suspend/resume handling
in the SCSI subsystem introduced a race condition between SPI domain validation
and system suspend. The two patches in this series fix that race. Please
consider these two patches for kernel v4.15.
Thanks,
Since pm_mutex is not exported using lock/unlock_system_sleep() from
inside a kernel module causes a "pm_mutex undefined" linker error.
Hence move lock/unlock_system_sleep() into kernel/power/main.c and
export these.
Signed-off-by: Bart Van Assche
Cc: Rafael J. Wysocki
---
include/linux/suspend
Avoid that the following warning is reported when suspending a system
that is using the mptspi driver:
WARNING: CPU: 0 PID: 4187 at drivers/scsi/scsi_lib.c:2960
scsi_device_quiesce+0x20/0xb0
EIP: scsi_device_quiesce+0x20/0xb0
Call Trace:
spi_dv_device+0x65/0x5f0 [scsi_transport_spi]
mptspi_dv_d
On 1/5/18 9:26 AM, Bart Van Assche wrote:
> Hello Jens,
>
> As you know there are multiple drivers that both allocate a scatter/gather
> list and populate that list with pages. This patch series moves the code for
> allocating and freeing such scatterlists from several drivers into
> lib/scatterli
On Fri, Jan 5, 2018 at 6:19 PM, Bart Van Assche wrote:
> Since pm_mutex is not exported using lock/unlock_system_sleep() from
> inside a kernel module causes a "pm_mutex undefined" linker error.
> Hence move lock/unlock_system_sleep() into kernel/power/main.c and
> export these.
>
> Signed-off-by:
On Sat, 2018-01-06 at 00:00 +0100, Rafael J. Wysocki wrote:
> The changes are fine by me, but I really would prefer them to go in
> via the PM tree which I guess means that the second patch would need
> to go in this way too.
Hello Rafael,
If I can get a Tested-by from Woody and an Acked-by from
On Sat, Jan 6, 2018 at 12:32 AM, Bart Van Assche wrote:
> On Sat, 2018-01-06 at 00:00 +0100, Rafael J. Wysocki wrote:
>> The changes are fine by me, but I really would prefer them to go in
>> via the PM tree which I guess means that the second patch would need
>> to go in this way too.
>
> Hello R
Static analysis reports that 'handle' may be a user controlled value
that is used as a data dependency to read 'sp' from the
'req->outstanding_cmds' array. In order to avoid potential leaks of
kernel memory values, block speculative execution of the instruction
stream that could issue reads based
Quoting Mark's original RFC:
"Recently, Google Project Zero discovered several classes of attack
against speculative execution. One of these, known as variant-1, allows
explicit bounds checks to be bypassed under speculation, providing an
arbitrary read gadget. Further details can be found on the
Dan Williams writes:
> Quoting Mark's original RFC:
>
> "Recently, Google Project Zero discovered several classes of attack
> against speculative execution. One of these, known as variant-1, allows
> explicit bounds checks to be bypassed under speculation, providing an
> arbitrary read gadget. Fu
On Fri, Jan 5, 2018 at 6:22 PM, Eric W. Biederman wrote:
> Dan Williams writes:
>
>> Quoting Mark's original RFC:
>>
>> "Recently, Google Project Zero discovered several classes of attack
>> against speculative execution. One of these, known as variant-1, allows
>> explicit bounds checks to be by
41 matches
Mail list logo
