On 25/04/17 00:15, Martin K. Petersen wrote:
>
> Juergen,
>
>> On 22/04/17 03:21, Geliang Tang wrote:
>>> Use offset_in_page() macro instead of open-coding.
>>>
>>> Signed-off-by: Geliang Tang
>>
>> Reviewed-by: Juergen Gross
>
> Taking this through the
Jens,
On 4/25/17 13:32, Jens Axboe wrote:
>> I just realized that this macro is defined in linux/device-mapper.h,
>> which does not seem like to best place to have it. Why not blkdev.h ?
>> Any particular reason ? This leads to some strange include dependencies,
>> like many nfs/blocklayout/
On 04/24/2017 09:27 PM, Damien Le Moal wrote:
> Jon,
>
> On 4/25/17 13:00, Jens Axboe wrote:
>> On Mon, Apr 24 2017, Jon Derrick wrote:
>>> The current command submission code uses a sector-based value when
>>> considering the maximum number of blocks per command. With a
>>> 4k-formatted
Jon,
On 4/25/17 13:00, Jens Axboe wrote:
> On Mon, Apr 24 2017, Jon Derrick wrote:
>> The current command submission code uses a sector-based value when
>> considering the maximum number of blocks per command. With a
>> 4k-formatted namespace and a command exceeding max hardware limits, this
>>
On Mon, Apr 24 2017, Jon Derrick wrote:
> The current command submission code uses a sector-based value when
> considering the maximum number of blocks per command. With a
> 4k-formatted namespace and a command exceeding max hardware limits, this
> calculation doesn't split IOs which should be
From: Wei Yongjun
free_irq() expects the same device identity that was passed to
corresponding request_irq(), otherwise the IRQ is not freed.
Signed-off-by: Wei Yongjun
---
drivers/scsi/mac_esp.c | 2 +-
1 file changed, 1 insertion(+), 1
Martin,
looks good to me, so:
Reviewed-By: Michael Schmitz
Am 25.04.2017 um 10:29 schrieb Martin K. Petersen:
>
> Finn,
>
>> Commit da244654c66e ("[SCSI] mac_esp: fix for quadras with two esp chips")
>> added mac_scsi_esp_intr() to handle the IRQ lines from a pair of
Martin,
On 4/25/17 08:11, Martin K. Petersen wrote:
>
> Christoph,
>
>> Use a switch for the sense key, and remove two pointless variables
>> that are only used once.
>
>> -if (unmap)
>
> The rationale behind the unmap variable was clarity and avoiding magic
>
The current command submission code uses a sector-based value when
considering the maximum number of blocks per command. With a
4k-formatted namespace and a command exceeding max hardware limits, this
calculation doesn't split IOs which should be split and fails in the
nvme layer. This patch fixes
On Mon, 2017-04-24 at 19:33 -0400, Martin K. Petersen wrote:
> > The debugfs infrastructure is already there, and it already supports
> > showing requests. Bart is just exposing the SCSI information.
>
> That's fine.
>
> I was merely objecting to the fact that we already have umpteen existing
>
Omar,
> The debugfs infrastructure is already there, and it already supports
> showing requests. Bart is just exposing the SCSI information.
That's fine.
I was merely objecting to the fact that we already have umpteen existing
interfaces for displaying SCSI command information.
--
Martin K.
Michael,
> I must have missed that one - where was it posted to?
Usual bat channel:
https://patchwork.kernel.org/patch/9658369/
--
Martin K. Petersen Oracle Linux Engineering
Hi Martin,
I must have missed that one - where was it posted to?
Cheers,
Michael
Am 25.04.2017 um 10:29 schrieb Martin K. Petersen:
>
> Finn,
>
>> Commit da244654c66e ("[SCSI] mac_esp: fix for quadras with two esp chips")
>> added mac_scsi_esp_intr() to handle the IRQ lines from a
On Mon, Apr 24, 2017 at 07:19:50PM -0400, Martin K. Petersen wrote:
>
> Bart,
>
> > SCSI tracing has to be enabled before a test is started, produces a
> > huge amount of data, and deriving state information from a huge trace
> > is far from easy. The information in debugfs provides an easy to
Bart,
> SCSI tracing has to be enabled before a test is started, produces a
> huge amount of data, and deriving state information from a huge trace
> is far from easy. The information in debugfs provides an easy to read
> overview of the current state without having to analyze megabytes of
>
Christoph,
> Use a switch for the sense key, and remove two pointless variables
> that are only used once.
> - if (unmap)
The rationale behind the unmap variable was clarity and avoiding magic
values.
I'm OK with this, however:
> + if
> From: Tomohiro Kusumi
>
> MASK_EE_STATUS added by 66ec6d59 was unused, but it seems to have been
> defined to do this.
>
> Signed-off-by: Tomohiro Kusumi
Subhash: Please review these two. Thanks!
--
Martin K. Petersen Oracle Linux Engineering
Sinan,
> Due to relaxed ordering requirements on multiple architectures,
> drivers are required to use wmb/rmb/mb combinations when they need to
> guarantee observability between the memory and the HW.
Applied to 4.12/scsi-queue, thanks!
--
Martin K. Petersen Oracle Linux Engineering
> There were pci_alloc_consistent() failures on ARM64 platform.
> Use dma_alloc_coherent() with GFP_KERNEL flag DMA memory allocations.
Somebody please review!
--
Martin K. Petersen Oracle Linux Engineering
Finn,
> Commit da244654c66e ("[SCSI] mac_esp: fix for quadras with two esp chips")
> added mac_scsi_esp_intr() to handle the IRQ lines from a pair of on-board
> ESP chips (a normal shared IRQ did not work).
>
> Proper mutual exclusion was missing from that patch. This patch fixes
> race
Hannes,
> Commit c5ce0ab ("scsi: sas: move scsi_remove_host call...")
> moved the call to scsi_remove_host() into sas_remove_host(),
> but forgot to modify the mpt drivers.
Applied to 4.12/scsi-queue.
--
Martin K. Petersen Oracle Linux Engineering
Hannes,
> Once the reserved page array is unused we can reset the 'res_in_use'
> state; here we can do a lazy update without holding the mutex as we
> only need to check against concurrent access, not concurrent release.
Applied to 4.12/scsi-queue.
--
Martin K. Petersen Oracle Linux
Alexey,
> As Christoph Hellwig noted, SCSI commands that transfer data
> always have a SG entry. The patch removes dead code in
> mvumi_make_sgl(), mvumi_complete_cmd() and mvumi_timed_out()
> that handle zero scsi_sg_count(scmd) case.
>
> Also the patch adds pci_unmap_sg() on failure path in
Juergen,
> On 22/04/17 03:21, Geliang Tang wrote:
>> Use offset_in_page() macro instead of open-coding.
>>
>> Signed-off-by: Geliang Tang
>
> Reviewed-by: Juergen Gross
Taking this through the Xen tree or should I queue it?
--
Martin K. Petersen
Arnd,
> As pointed out by Al Viro for my previous series, the driver has no need
> to call access_ok() and __copy_from_user()/__copy_to_user(). Changing
> it to regular copy_from_user()/copy_to_user() simplifies the code without
> any real downsides, making it less error-prone at best.
>
> This
Christoph,
> sparse found a bug that has always been present since the driver was
> merged:
>
> drivers/scsi/pmcraid.c:2353:12: warning: context imbalance in
> 'pmcraid_reset_reload' - different lock contexts for basic block
>
> Fix this by using a common unlock goto label, and also reduce the
On Mon, 2017-04-24 at 17:35 -0400, Martin K. Petersen wrote:
> > Show the SCSI CDB, .eh_eflags and .result for pending SCSI commands
> > in /sys/kernel/debug/block/*/mq/*/dispatch and */rq_list.
>
> Why not use SCSI tracing if you are interested in these?
Hello Martin,
SCSI tracing has to be
On Sun, 2017-04-23 at 12:28 -0500, James Bottomley wrote:
> --- a/drivers/scsi/scsi_lib.c
> +++ b/drivers/scsi/scsi_lib.c
> @@ -1250,6 +1250,12 @@ scsi_prep_state_check(struct scsi_device *sdev, struct
> request *req)
> break;
> case SDEV_BLOCK:
>
Bart,
> Show the SCSI CDB, .eh_eflags and .result for pending SCSI commands
> in /sys/kernel/debug/block/*/mq/*/dispatch and */rq_list.
Why not use SCSI tracing if you are interested in these?
--
Martin K. Petersen Oracle Linux Engineering
Dear Beloved, Sorry for the inconvenience, I am getting in touch with you
regarding an extremely important and urgent matter, Please I need your urgent
assistance and idea to finish up a project (Orphanage Home) Due to my health
condition, Everything is available to finish up the project, All I
From: Colin Ian King
These module parameter variables don't need global scope, make them static
Signed-off-by: Colin Ian King
---
drivers/scsi/fcoe/fcoe.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git
On 24/04/17 01:36 AM, Knut Omang wrote:
> My first reflex when reading this thread was to think that this whole domain
> lends it self excellently to testing via Qemu. Could it be that doing this in
> the opposite direction might be a safer approach in the long run even though
> (significant)
Please disregard this patch. there is an issue in the fix
-- james
On 4/21/2017 5:24 PM, jsmart2...@gmail.com wrote:
From: James Smart
To select the appropriate shost template, the driver is issuing
a mailbox command to retrieve the wwn. Turns out the sending of
the
Looks fine,
Reviewed-by: Christoph Hellwig
Looks fine,
Reviewed-by: Christoph Hellwig
Looks fine,
Reviewed-by: Christoph Hellwig
Looks fine,
Reviewed-by: Christoph Hellwig
I would just use a switch and kill the op and unmap variables entirely,
e.g. something like the patch below:
---
From: Christoph Hellwig
Subject: sd: Cleanup sd_done sense data handling
Use a switch for the sense key, and remove two pointless variables
that are only used once.
On Mon, Apr 24, 2017 at 04:51:10PM +0900, damien.lem...@wdc.com wrote:
> From: Damien Le Moal
>
> Re-shuffle the code to be more efficient by not initializing
> variables upfront (i.e. do it only when necessary).
> Also replace the do_div calls with calls to
Looks good,
Reviewed-by: Christoph Hellwig
Looks fine,
Reviewed-by: Christoph Hellwig
Looks good,
Reviewed-by: Christoph Hellwig
On Fri, Apr 21, 2017 at 02:13:02PM -0700, Song Liu wrote:
> When a device is deleted through sysfs handle "delete", the code
> locks shost->scan_mutex. If multiple devices are deleted at the
> same time, these deletes will be handled in series.
>
> On the other hand, some devices do long latency
Our final fix before the 4.12 release (hopefully). It's an error leg
again: the fix to not bug on empty DMA transfers is returning the wrong
code and confusing the block layer.
The patch is available here:
git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi.git scsi-fixes
The short
On Mon, Apr 24, 2017 at 02:38:40PM +0100, John Garry wrote:
> On 24/04/2017 11:09, John Garry wrote:
> >On 21/04/2017 13:11, Johannes Thumshirn wrote:
> >>Move scsi_remove_host call into sas_remove_host and remove it from SAS
> >>HBA
> >>drivers, so we don't mess up the ordering. This solves an
On 24/04/2017 11:09, John Garry wrote:
On 21/04/2017 13:11, Johannes Thumshirn wrote:
Move scsi_remove_host call into sas_remove_host and remove it from SAS
HBA
drivers, so we don't mess up the ordering. This solves an issue with
double
deleting sysfs entries that was introduced by the change
On 04/20/2017 04:14 AM, Martin K. Petersen wrote:
> Bart Van Assche writes:
>
>> On Thu, 2017-04-06 at 15:36 +0200, Hannes Reinecke wrote:
>>> The block layer always calls the timeout function from a workqueue
>>> context, so there is no need to have yet another
On 21/04/2017 13:11, Johannes Thumshirn wrote:
Move scsi_remove_host call into sas_remove_host and remove it from SAS HBA
drivers, so we don't mess up the ordering. This solves an issue with double
deleting sysfs entries that was introduced by the change of sysfs behaviour
from commit bcdde7e
Commit c5ce0ab ("scsi: sas: move scsi_remove_host call...")
moved the call to scsi_remove_host() into sas_remove_host(),
but forgot to modify the mpt drivers.
Fixes: c5ce0ab ("scsi: sas: move scsi_remove_host call into sas_remove_host")
Signed-off-by: Hannes Reinecke
---
On 04/21/2017 02:11 PM, Johannes Thumshirn wrote:
> Move scsi_remove_host call into sas_remove_host and remove it from SAS HBA
> drivers, so we don't mess up the ordering. This solves an issue with double
> deleting sysfs entries that was introduced by the change of sysfs behaviour
> from commit
On Mon, Apr 24, 2017 at 10:26:36AM +0200, Hannes Reinecke wrote:
> Once the reserved page array is unused we can reset the
> 'res_in_use' state; here we can do a lazy update without
> holding the mutex as we only need to check against
> concurrent access, not concurrent release.
>
> Fixes:
Once the reserved page array is unused we can reset the
'res_in_use' state; here we can do a lazy update without
holding the mutex as we only need to check against
concurrent access, not concurrent release.
Fixes: 1bc0eb0 ("scsi: sg: protect accesses to 'reserved' page array")
Signed-off-by:
From: Bart Van Assche
A value is assigned to the variable 'capacity' in sd_zbc_read_zones()
but that value is never used. Hence remove the variable 'capacity'.
[Damien: There is no need to initialize to 0 the variable 'ret'
in sd_zbc_read_zones()]
Signed-off-by:
From: Damien Le Moal
Resetting a zone write pointer is equivalent to discarding sectors:
after a reset, the zone sectors will contain zeros (or the format
pattern). So there is no need for mutual exclusion between a zone reset
and write. Similarly to discard, make it the
From: Damien Le Moal
Rename sd_zbc_setup_write_cmnd() to sd_zbc_write_lock_zone() to be
clear about what the function actually does. To be consistent, also
rename sd_zbc_cancel_write_cmnd() to sd_zbc_write_unlock_zone().
No functional change is introduced by this patch.
From: Damien Le Moal
In sd_done(), for the ILLEGAL REQUEST sense key case, add an 'else if'
after the first 'if (sshdr.asc == 0x10)' test to avoid the second test
(the values tested are different).
Still for the same ILLEGAL REQUEST case, move the declarations of the
From: Damien Le Moal
Re-shuffle the code to be more efficient by not initializing
variables upfront (i.e. do it only when necessary).
Also replace the do_div calls with calls to sectors_to_logical().
No functional change is introduced by this patch.
Signed-off-by: Damien
From: Damien Le Moal
Fix argument names and description of functions documentation comments.
No functional change is introduced by this patch.
Signed-off-by: Damien Le Moal
---
drivers/scsi/sd.c | 29 ++---
1 file changed,
From: Damien Le Moal
This series of small patches cleanup code mainly in sd.c and sd_zbc.c, with
another small improvement in scsi_error.c.
Only the last patch of this series introduces a functional change. All other
patches do not change any functionality.
Changes from
From: Damien Le Moal
Use get_unaligned_be32 and get_unaligned_be64 to obtain values from
the sense buffer instead of open coding the operations.
Also change the function return value to a bool and fix the function
signature declaration to remove spaces triggering
On Mon, 2017-04-17 at 08:31 +1000, Benjamin Herrenschmidt wrote:
> On Sun, 2017-04-16 at 10:34 -0600, Logan Gunthorpe wrote:
> >
> > On 16/04/17 09:53 AM, Dan Williams wrote:
> > > ZONE_DEVICE allows you to redirect via get_dev_pagemap() to retrieve
> > > context about the physical address in
On 04/22/2017 01:40 AM, Bart Van Assche wrote:
> Show the SCSI CDB, .eh_eflags and .result for pending SCSI commands
> in /sys/kernel/debug/block/*/mq/*/dispatch and */rq_list.
>
> Signed-off-by: Bart Van Assche
> Cc: Martin K. Petersen
>
Applie to the nvme-4.12 branch. Any praying we're not going to see
any conflicts with the SCSI tree.
FYI, we noticed the following commit:
commit: ab1218235c09d0fabdea31cbd4f099e7cabccb27 ("sd: Make synchronize cache
upon shutdown asynchronous")
url:
https://github.com/0day-ci/linux/commits/Bart-Van-Assche/Avoid-that-__scsi_remove_device-hangs/20170418-013958
base:
Looks fine,
Reviewed-by: Christoph Hellwig
Note that there is plenty opportunity for cleanup even in the moved
code section, but let's get this issue sorted out only for now.
65 matches
Mail list logo