Hey MNC,
On Fri, 2017-05-26 at 22:14 -0500, Mike Christie wrote:
> Thanks for the patch.
>
Btw, after running DATERA's internal longevity and scale tests across
~20 racks on v4.1.y with this patch over the long weekend, there haven't
been any additional regressions.
> On 05/26/2017 12:32 AM,
The driver may sleep under a write spin lock, the function call path is:
qla4_82xx_wr_32 (acquire the lock)
qla4_82xx_crb_win_lock
schedule or cpu_relax
To fix it, the lock is released before "schedule" and "cpu_relax",
and the lock is acquired again after "schedule" and "cpu_relax".
The driver may sleep under a spin lock, and the function call path is:
iscsit_tpg_enable_portal_group (acquire the lock by spin_lock)
iscsi_update_param_value
kstrdup(GFP_KERNEL) --> may sleep
To fix it, the "GFP_KERNEL" is replaced with "GFP_ATOMIC".
Signed-off-by: Jia-Ju Bai
Hi Jia-Ju,
[auto build test ERROR on scsi/for-next]
[also build test ERROR on v4.12-rc3 next-20170530]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system]
url:
https://github.com/0day-ci/linux/commits/Jia-Ju-Bai/qla4xxx-Fix-a-sleep-in-atomic
The driver may sleep under a spin lock, and the function call path is:
mraid_mm_attach_buf (acquire the lock by spin_lock_irqsave)
pci_pool_alloc(GFP_KERNEL) --> may sleep
To fix it, the "GFP_KERNEL" is replaced with "GFP_ATOMIC".
Signed-off-by: Jia-Ju Bai
---
On Tue, May 30, 2017 at 03:12:41PM +, Bart Van Assche wrote:
> On Sat, 2017-05-27 at 22:21 +0800, Ming Lei wrote:
> > --- a/drivers/scsi/scsi_lib.c
> > +++ b/drivers/scsi/scsi_lib.c
> > @@ -3030,7 +3030,10 @@ scsi_internal_device_unblock(struct scsi_device
> > *sdev,
> > return
On Tue, May 30, 2017 at 12:04:02PM -0700, Eduardo Valentin wrote:
> On Sat, May 27, 2017 at 10:21:21PM +0800, Ming Lei wrote:
> > blk_mq_unquiesce_queue() is used for unquiescing the
> > queue explicitly, so replace blk_mq_start_stopped_hw_queues()
> > with it.
> >
> > Cc:
The driver may sleep under a write spin lock, the function call path is:
qla4_82xx_wr_32 (acquire the lock)
qla4_82xx_crb_win_lock
schedule or cpu_relax
To fixed it, the lock is released before "schedule" and "cpu_relax",
and the lock is acquired again after "schedule" and "cpu_relax".
Hi Bryant,
[auto build test ERROR on target/master]
[also build test ERROR on v4.12-rc3 next-20170530]
[if your patch is applied to the wrong git tree, please drop us a note to help
improve the system]
url:
https://github.com/0day-ci/linux/commits/Bryant-G-Ly/TCMU-Enable-Reconfiguration
On Sat, May 27, 2017 at 10:21:21PM +0800, Ming Lei wrote:
> blk_mq_unquiesce_queue() is used for unquiescing the
> queue explicitly, so replace blk_mq_start_stopped_hw_queues()
> with it.
>
> Cc: linux-n...@lists.infradead.org
> Cc: linux-scsi@vger.kernel.org
> Cc: dm-de...@redhat.com
>
This will enable the toggling of write_cache in tcmu through targetcli-fb
Signed-off-by: Bryant G. Ly
Reviewed-By: Mike Christie
---
drivers/target/target_core_user.c | 30 ++
1 file changed, 30 insertions(+)
diff
This patch adds more info about the attribute being changed,
so that usersapce can easily figure out what is happening.
Signed-off-by: Bryant G. Ly
---
drivers/target/target_core_user.c | 27 +--
include/uapi/linux/target_core_user.h | 1
This patch consists of adding a netlink to allow for reconfiguration
of a device in tcmu.
It also changes and adds some attributes that are reconfigurable:
write_cache, device size, and device path.
Bryant G. Ly (5):
tcmu: Support emulate_write_cache
tcmu: Add netlink for device
This gives tcmu the ability to handle events that can cause
reconfiguration, such as resize, path changes, write_cache, etc...
Signed-off-by: Bryant G. Ly
Reviewed-By: Mike Christie
---
drivers/target/target_core_user.c | 12
This allows for userspace to change the device path after
it has been created. Thus giving the user the ability to change
the path. The use case for this is to allow for virtual optical
to have media change.
Signed-off-by: Bryant G. Ly
---
Allow tcmu backstores to be able to set the device size
after it has been configured via set attribute.
Part of support in userspace to support certain backstores
changing device size.
Signed-off-by: Bryant G. Ly
---
drivers/target/target_core_user.c | 59
From: Quinn Tran
Accelerate generation of SCSI busy to let initiators slow
down when target is running low in resources.
Signed-off-by: Quinn Tran
Signed-off-by: Himanshu Madhani
---
drivers/scsi/qla2xxx/qla_init.c |
From: Quinn Tran
Remove extra register read in each interrupt processing
to improve performance.
Signed-off-by: Quinn Tran
Signed-off-by: Himanshu Madhani
---
drivers/scsi/qla2xxx/qla_iocb.c | 4 +++-
From: Quinn Tran
Previously, the ql2xexchoffld module parameter was used to control
the max number of exchanges to be offload onto host memory.
Module parameter ql_dm_tgt_ex_pct was used to control the percentage
of exchanges allocated to the Target side.
With this
From: Quinn Tran
Assign unique id to all traces and logs for debug purpose.
Signed-off-by: Quinn Tran
Signed-off-by: Himanshu Madhani
---
drivers/scsi/qla2xxx/qla_attr.c | 2 +-
drivers/scsi/qla2xxx/qla_bsg.c|
From: Quinn Tran
Tell FW to track exchange/cmd state to prevent driver
from using stale exchange or exchange that is not meant
for this command.
Signed-off-by: Quinn Tran
Signed-off-by: Himanshu Madhani
---
From: Quinn Tran
Current code already destroy all target sessions when
target Mode is stopped. Target core would waits for
all commands that belong to each session to purge.
The extra wait for interrupts to settle down is not
relevant.
Signed-off-by: Quinn Tran
From: Quinn Tran
During ABTS or Abort task, qla2xxx does a pre-search for
the se_cmd, based on command's tag. The same search is
performed by TCM. Remove the extra search from qla2xxx.
The reason for hard coding LUN ID to 0 is that, from the FC
protocol perspective, ABTS
From: Quinn Tran
When driver is unloaded, all sessions are torn down, all
commmands are flushed, chip is reset to ensure there is
no knowledge of target mode in ISP. The irq_cmd_count field
was used to make sure all commands are processed on top of that.
The irq_cmd_count
From: Quinn Tran
If the port is moved/changed, current code would trigger
a deletion. If the port is already deleted, then do relogin.
Signed-off-by: Quinn Tran
Signed-off-by: Himanshu Madhani
---
From: Sawan Chandak
There is already flag defined PFLG_DISCONNECTED, which is set
for PCI or register disconnect error condition. There is no need to have
flag PCI_ERR, which has same purpose. Remove use of PCI_ERR flag and use
PFLG_DISCONNECTED flag during error
From: Quinn Tran
Convert usage of spin_lock to spin_lock_irqsave because qla2xxx driver
accesses all the data structures in an interrupt context.
Signed-off-by: Quinn Tran
Signed-off-by: Himanshu Madhani
---
Hi Martin,
This series contains patches that were dropped from 4.12.0-rc3 inclusion, since
they can go to 4.13 merge window.
Changes from v1 --> v2
o addressed 0-day kernel warning.
o addressed cleanups and updates as per Bart's comments.
o added Acked-by tag from Nicholas to applicable patches.
From: Quinn Tran
Name server login is normally handle by FW. In some
rare case where one of the switches is being updated,
name server login could get affected. Trigger relogin
to name server when driver detects this condition.
Signed-off-by: Quinn Tran
From: Joe Carnuccio
Signed-off-by: Joe Carnuccio
Signed-off-by: Himanshu Madhani
---
drivers/scsi/qla2xxx/qla_init.c | 8
drivers/scsi/qla2xxx/qla_tmpl.c | 16 +---
2 files changed, 13
From: Quinn Tran
In case of hardware queue full, commands can loop between
TCM stack and tcm_qla2xx shim layers for retry. While command
is waiting for retry, task mgmt can get ahead and abort the
cmmand that encountered queue full condition. Fix this by
dropping the
From: Quinn Tran
Driver added mechanism to move ABTS/PUREX/RIDA mailbox to
ATIO queue as part of commit id 41dc529a4602ac737020f423f84686a81de38e6d
("qla2xxx: Improve RSCN handling in driver").
This patch addes a check to only allow ABTS/PURX/RIDA
to be moved to ATIO
From: Quinn Tran
Acked-by: Nicholas Bellinger
Signed-off-by: Quinn Tran
Signed-off-by: Himanshu Madhani
---
drivers/scsi/qla2xxx/qla_target.c | 32 +++-
From: Quinn Tran
Add missing memory dump of Exchange Offload and Extended
login into FW dump.
Cc:
Signed-off-by: Quinn Tran
Signed-off-by: Himanshu Madhani
---
drivers/scsi/qla2xxx/qla_dbg.c
If a discovery is already running (ie if the 'pending' flags is set)
there is no need to restart discovery from the timeout handler.
Signed-off-by: Hannes Reinecke
---
drivers/scsi/libfc/fc_disc.c | 8 +---
1 file changed, 5 insertions(+), 3 deletions(-)
diff --git
fc_disc_stop() calls fc_disc_stop_rports(), which requires the
disc_mutex to be held.
And we need to ensure that no fc_disc_timeout functions are queued
after this function, so set the callback to NULL and ensure that
all functions terminate early when the callback is not set.
Signed-off-by:
Move 'pending' and 'requested' setting out of fc_disc_gpn_fc_req()
into the calling function.
No functional change.
Signed-off-by: Hannes Reinecke
---
drivers/scsi/libfc/fc_disc.c | 9 ++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git
Hi all,
we've seen reports for a crash with an invalid timer_list->function,
which turned out to be an unsafe usage of libfc discovery callbacks.
This patchset fixes up the problem
As usual, comments and reviews are welcome.
Hannes Reinecke (3):
libfc: move 'pending' and 'requested' setting
On Sat, 2017-05-27 at 22:21 +0800, Ming Lei wrote:
> --- a/drivers/scsi/scsi_lib.c
> +++ b/drivers/scsi/scsi_lib.c
> @@ -3030,7 +3030,10 @@ scsi_internal_device_unblock(struct scsi_device *sdev,
> return -EINVAL;
>
> if (q->mq_ops) {
> -
39 matches
Mail list logo