Update driver version from 27.100.00.00 to 27.101.00.00.
Signed-off-by: Suganath Prabu
---
drivers/scsi/mpt3sas/mpt3sas_base.h | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.h
b/drivers/scsi/mpt3sas/mpt3sas_base.h
index 3a294b9..8003519
Use ioc->base_readl to restrict the readl retries to only
Aero controllers.
Signed-off-by: Suganath Prabu
---
drivers/scsi/mpt3sas/mpt3sas_base.c | 39 +++--
1 file changed, 20 insertions(+), 19 deletions(-)
diff --git a/drivers/scsi/mpt3sas/mpt3sas_base.c
b/dri
Sometimes Aero controllers appears to be returning bad data (0)
for doorbell register read and if retries are performed immediately
after the bad read, they return good data.
Workaround is added to retry read from doorbell registers
for maximum three times if driver get the zero.
Added functions b
Problem statement:
Sometimes aero controllers appears to be returning bad data (0)
for doorbell register read and if retries are performed immediately
after the bad read, they return good data.
Fix:
In below four patches added workaround to retry read operation from
controller doorbell registers f
Adding flag "is_aero_ioc" to differentiate aero based controllers
from other gen35 controllers.
Signed-off-by: Suganath Prabu
---
drivers/scsi/mpt3sas/mpt3sas_base.h | 1 +
drivers/scsi/mpt3sas/mpt3sas_scsih.c | 14 --
2 files changed, 9 insertions(+), 6 deletions(-)
diff --git a/
> On 12/5/18 10:45 PM, Kashyap Desai wrote:
> >>
> >> If the 'tag' passed to scsi_host_find_tag() is valid, I think there
> >> shouldn't have such issue.
> >>
> >> If you want to find outstanding IOs, maybe you can try
> >> blk_mq_queue_tag_busy_iter()
> >> or blk_mq_tagset_busy_iter(), because you
This reverts commit db186382af21e926e90df19499475f2552192b77.
This commit introduced regression with FCP discovery so revert
it back to fix discovery for FCP luns
Cc:
Signed-off-by: Himanshu Madhani
---
Hi Martin,
This patch has introduced regression for LUN discovery with FC.
Please apply
John,
> "DIX is supported in mpt3sas , but not by default and hence while
> loading mpt3sas driver set module parameter "prot_mask=0x7F" to enable
> DIX support. "
The plumbing is there but it doesn't work. I tried earlier today and I
still can't get a single I/O to complete. I'll try to hack a
The intent of invoking configfs_depend_item in commit 7474f52a82d51
("tcm_qla2xxx: Perform configfs depend/undepend for base_tpg")
was to prevent a physical Fibre Channel port removal when
virtual (NPIV) ports announced through that physical port are active.
The change does not work as expected: it
On Thu, 6 Dec 2018, Christoph Hellwig wrote:
> This function is a huge mess with duplicated error handling. Split out
> a few useful helpers and use goto labels to untangle the error handling
> and no-data ioctl handling.
>
> Signed-off-by: Christoph Hellwig
> ---
> drivers/scsi/gdth.c | 244
On Thu, Dec 06, 2018 at 07:52:55AM -0800, Christoph Hellwig wrote:
> mesh has no limitations on crossing pages for segments. Just make
> the 65535 byte segment size limit explicit, even if it matches the
> current block layer limit.
>
> Signed-off-by: Christoph Hellwig
Tested-by: Paul Mackerras
On Thu, Dec 06, 2018 at 07:52:56AM -0800, Christoph Hellwig wrote:
> mac53c94 has no limitations on crossing pages for segments. Just make
> the 65535 byte segment size limit explicit, even if it matches the
> current block layer limit.
>
> Signed-off-by: Christoph Hellwig
Tested-by: Paul Macke
Would be nice to be pointed to the correct place to report major
regressions, if not here.
Note that the same error occurs:
* on 4.18.19
* on 4.19.8-rc1
* with the latest firmware (16.00.01)
* on a number of other peoples' powerpc64/power9 hardware.
Note that both 4.18.X and 4.19.Y will, in
UBSAN reported those with MegaRAID SAS-3 3108,
[ 77.467308] UBSAN: Undefined behaviour in
drivers/scsi/megaraid/megaraid_sas_fp.c:117:32
[ 77.475402] index 255 is out of range for type 'MR_LD_SPAN_MAP [1]'
[ 77.481677] CPU: 16 PID: 333 Comm: kworker/16:1 Not tainted 4.20.0-rc5+ #1
[ 77.48
Hey everybody,
I'm the arch lead for Gentoo Linux on alpha (yes, the arch still
exists). Starting with 4.20-rc*, I get SCSI reset loops shortly
after boot:
aboot: starting kernel boot/vmlinuz-4.20.0-rc3 with arguments ro root=/dev/sda2
console=ttyS0 loglevel=6
[0.00] Linux version 4.20.
On 06/12/2018 04:22, Martin K. Petersen wrote:
Hi Martin,
John,
I'll have to go revisit the archives for your exact fio bits.
Your fio test case ran to completion on all my test systems.
But I successfully ran our DIX/T10 PI qualification tooling without any
errors on the latest SCSI t
https://bugzilla.kernel.org/show_bug.cgi?id=199435
--- Comment #27 from Gaetan Trellu (gaetan.tre...@incloudus.com) ---
By compiling the hpsa kernel module from SourceForge on Ubuntu 16.04 with
kernel 4.4 solved the issue for us.
Steps:
# apt-get install dkms build-essential
# tar xjvf hpsa-3.4.2
On Thu, 2018-12-06 at 07:52 -0800, Christoph Hellwig wrote:
> This allows the host driver to indicate the maximum supported
> segment size in a nice an easy way, so that the driver doesn't
> have to worry about DMA-layer imposed limitations.
Reviewed-by: Bart Van Assche
On Thu, 2018-12-06 at 07:52 -0800, Christoph Hellwig wrote:
> The same effects can be archived by setting the dma_boundary to
achieved?
> PAGE_SIZE - 1 and the max_segment_size to PAGE_SIZE, so shift those
> settings into the drivers. No
On Thu, 2018-12-06 at 07:52 -0800, Christoph Hellwig wrote:
> Most SCSI drivers want to enable "clustering", that is merging of
> segments so that they might span more than a single page. Remove the
> ENABLE_CLUSTERING define, and require drivers to explicitly set
> DISABLE_CLUSTERING to disable t
On 06/12/2018 17:50, Johannes Thumshirn wrote:
> Why not calling dma_alloc_coherent() directly instead of using the
> pci_alloc_consistent() wrapper?
Ah should've read the whole series
--
Johannes ThumshirnSUSE Labs Filesystems
jthumsh...@suse.de
On 06/12/2018 16:57, Christoph Hellwig wrote:
> Out of the three callers once insists on the scratch buffer, and the
> others are fine with a new allocation. Switch those two to juse use
> pci_alloc_consistent directly, and open code the scratch buffer
> allocation in the remaining one. This avoi
Complements
v2.6.35 commit 64deb6efdc55
("[SCSI] zfcp: Use status_read_buf_num provided by FCP channel")
which replaced the hardcoded 16 with a variable value
Also complements already existing fixups for above commit
v2.6.35 commit 8d88cf3f3b9a
("[SCSI] zfcp: Update status read mempool")
v3.10
Suppose adapter (open) recovery is between opened QDIO queues and before
(the end of) initial posting of status read buffers (SRBs). This time
window can be seconds long due to FSF_PROT_HOST_CONNECTION_INITIALIZING
causing by design looping with exponential increase sleeps in the function
performin
James, Martin,
One new recovery fix, which is not urgent, for an old bug.
It's sufficient to apply it on top of the previously sent
23 zfcp updates for the v4.21 merge window
[https://www.spinics.net/lists/linux-scsi/msg125211.html].
The 2 new patches apply to Martin's 4.21/scsi-queue
and to James
On 06/12/2018 16:37, John Garry wrote:
> which is important. Let me know if you disagree and I can change it.
Sure, it's your driver. It was just because the patch is even titled
"Relocate some code to reduce complexity", so I thought of reducing the
complexity for readers even further (like you d
This code has been under a never defined ifdef since the beginning
of time (or at least history), and has just bitrotted. Nuke it.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/gdth.c | 151
1 file changed, 12 insertions(+), 139 deletions(-)
dif
The non-PCI code has bitrotted for quite a while and will just oops
on load because it passes a NULL pointer to the PCI DMA routines.
Lets kill it for good - if someone really wants to use one of these
cards I'll help mentoring them to write a proper driver glue.
Signed-off-by: Christoph Hellwig
Out of the three callers once insists on the scratch buffer, and the
others are fine with a new allocation. Switch those two to juse use
pci_alloc_consistent directly, and open code the scratch buffer
allocation in the remaining one. This avoids a case where we might
be doing a memory allocation
gdth_show_info currently allocs and frees a dma buffer four times,
which isn't very efficient. Reuse a single allocation instead.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/gdth_proc.c | 20 +---
1 file changed, 5 insertions(+), 15 deletions(-)
diff --git a/drivers/scsi/g
This function is a huge mess with duplicated error handling. Split out
a few useful helpers and use goto labels to untangle the error handling
and no-data ioctl handling.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/gdth.c | 244 +++-
1 file changed,
Switch from the legacy PCI DMA API to the generic DMA API. Also switch
to dma_map_single from pci_map_page in one case where this makes the code
simpler.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/gdth.c | 59 +++-
drivers/scsi/gdth_proc.c | 4 +-
This code can't be built into the kernel without editing the source
file and is not generall useful.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/gdth.c | 18 --
drivers/scsi/gdth_proc.c | 8
2 files changed, 26 deletions(-)
diff --git a/drivers/scsi/gdth.c b
This can't ever be compiled into the kernel, so remove it.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/gdth_ioctl.h | 89 ---
drivers/scsi/gdth_proc.c | 18
2 files changed, 107 deletions(-)
diff --git a/drivers/scsi/gdth_ioctl.h b/drivers/scs
Remove never compile in support for sending debug traces straight to
the serial port using direct port access.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/gdth.c | 70 -
1 file changed, 70 deletions(-)
diff --git a/drivers/scsi/gdth.c b/drivers/
This code has been under the never defined GDTH_RTC ifdef forever,
nuke it.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/gdth.c | 32
1 file changed, 32 deletions(-)
diff --git a/drivers/scsi/gdth.c b/drivers/scsi/gdth.c
index e8121b80233c..63d704301875 100
Cleans up various oddities found during a code audit, and drops the
legacy ISA support which hasn't had a chance to actually work for a
long time.
UBSAN reported those with MegaRAID SAS-3 3108,
[ 77.467308] UBSAN: Undefined behaviour in
drivers/scsi/megaraid/megaraid_sas_fp.c:117:32
[ 77.475402] index 255 is out of range for type 'MR_LD_SPAN_MAP [1]'
[ 77.481677] CPU: 16 PID: 333 Comm: kworker/16:1 Not tainted 4.20.0-rc5+ #1
[ 77.48
This allows the host driver to indicate the maximum supported
segment size in a nice an easy way, so that the driver doesn't
have to worry about DMA-layer imposed limitations.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/be2iscsi/be_main.c | 8 +---
drivers/scsi/hosts.c| 5 +
The old DAC960 drivers was fine with merging over segment bondaries,
so this new driver should be to.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/myrs.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/scsi/myrs.c b/drivers/scsi/myrs.c
index f47b36382afa..0264a2e2bc19 100644
---
mesh has no limitations on crossing pages for segments. Just make
the 65535 byte segment size limit explicit, even if it matches the
current block layer limit.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/mesh.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/sc
There is no such limitation in the protocol or implementation, so
remove it.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/xen-scsifront.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/scsi/xen-scsifront.c b/drivers/scsi/xen-scsifront.c
index bb76d0d2022b..f0068e96a177 100644
--
mac53c94 has no limitations on crossing pages for segments. Just make
the 65535 byte segment size limit explicit, even if it matches the
current block layer limit.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/mac53c94.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/dr
Most SCSI drivers want to enable "clustering", that is merging of
segments so that they might span more than a single page. Remove the
ENABLE_CLUSTERING define, and require drivers to explicitly set
DISABLE_CLUSTERING to disable this feature. All drivers not setting
use_clustering at all are swit
The same effects can be archived by setting the dma_boundary to
PAGE_SIZE - 1 and the max_segment_size to PAGE_SIZE, so shift those
settings into the drivers. Note that in many cases the setting might
be bogus, but this keeps the status quo.
Signed-off-by: Christoph Hellwig
---
Documentation/sc
Now that the the SCSI layer replaced the use of the cluster flag with
segment size limits and the DMA boundary we can remove the cluster
flag from the block layer.
Signed-off-by: Christoph Hellwig
---
block/blk-merge.c | 18 +++---
block/blk-settings.c | 3 ---
block/blk-sysf
The block layer has been carrying around a magic clustering flag,
which is an artefact of prehistoric scsi layer design decisions.
Originally it controller if any segment merging happens at all for a
given LLLD, but with merging moving to the block layer and massively
more complex now it now essen
The old DAC960 drivers was fine with merging over segment bondaries,
so this new driver should be to.
Signed-off-by: Christoph Hellwig
---
drivers/scsi/myrb.c | 1 -
1 file changed, 1 deletion(-)
diff --git a/drivers/scsi/myrb.c b/drivers/scsi/myrb.c
index f1abe38e3b3a..aeb282f617c5 100644
---
On 06/12/2018 14:17, Johannes Thumshirn wrote:
On 06/12/2018 14:34, John Garry wrote:
[...]
+static void hisi_sas_dma_unmap(struct hisi_hba *hisi_hba,
+ struct sas_task *task, int n_elem,
+ int n_elem_req, int n_elem_resp)
+{
+ stru
On 12/5/18 10:45 PM, Kashyap Desai wrote:
>>
>> If the 'tag' passed to scsi_host_find_tag() is valid, I think there
>> shouldn't have such issue.
>>
>> If you want to find outstanding IOs, maybe you can try
>> blk_mq_queue_tag_busy_iter()
>> or blk_mq_tagset_busy_iter(), because you may not know if
On 06/12/2018 14:34, John Garry wrote:
[...]
> +static void hisi_sas_dma_unmap(struct hisi_hba *hisi_hba,
> +struct sas_task *task, int n_elem,
> +int n_elem_req, int n_elem_resp)
> +{
> + struct device *dev = hisi_hba->dev;
> +
> + if
On 23/10/2018 06:35, Can Guo wrote:
> From: Dov Levenglick
>
> Enables core reset support. Add full initialization of the PHY and the
> controller before initializing UFS PHY and during link recovery.
>
> Signed-off-by: Dov Levenglick
> Signed-off-by: Amit Nischal
> Signed-off-by: Subhash Jad
From: Xiang Chen
Sht->sg_tablesize is set in the driver, and it will be assigned to
shost->sg_tablesize in SCSI mid-layer. So it is not necessary to
assign shost->sg_table one more time in the driver.
In addition to the change, change each scsi_host_template.sg_tablesize
to HISI_SAS_SGE_PAGE_CNT
From: Xiang Chen
This patchset adds support for DIX to v3 hw driver.
As discussed in the following thread, DIX seems to be in conflict with
SCSI MQ, so this is why we mark the support for this driver as
experimental:
https://marc.info/?l=linux-scsi&m=154392687627603&w=2,
As for cause of the iss
From: Xiang Chen
Relocate the codes related to dma_map/unmap in hisi_sas_task_prep()
to reduce complexity, with a view to add DIF/DIX support.
Signed-off-by: Xiang Chen
Signed-off-by: John Garry
---
drivers/scsi/hisi_sas/hisi_sas_main.c | 146 +-
1 file changed
This patchset fixes some warnings detected by the sparse tool, like these:
drivers/scsi/hisi_sas/hisi_sas_main.c:1469:52: warning: incorrect type in
assignment (different base types)
drivers/scsi/hisi_sas/hisi_sas_main.c:1469:52:expected unsigned short
[unsigned] [assigned] [usertype] tag_of_
From: Xiang Chen
For v3 hw, we support DIF operation for SAS, but not SATA.
This patchset adds the SW support for the described features. The main
components are as follows:
- Get DIF enablement from module param
- Fill PI fields
- Fill related to DIF in DQ and protection iu memories
Signed-off
This patchset introduces support to the driver for DIF/DIX.
We are making DIX support as RFC for now, based on issues discussed in the
following:
https://marc.info/?l=linux-scsi&m=154357719329297&w=4 and
https://marc.info/?l=linux-scsi&m=154406987703456&w=2
We will only support PI in v3 hw at the
David,
> Indeed, the comment should refer to page 0x83.
> @Martin: all patches in this series have now been reviewed+acked. Can
> you fix the above comment (s/0x80/0x83) if/when you merge, or
>should I resend the series with this fixed?
I'll fix it up.
--
Martin K. Petersen
kmemdup has implemented the function that kmalloc() + memcpy().
We prefer to kmemdup rather than code opened implementation.
This issue was detected with the help of coccinelle.
Signed-off-by: Wen Yang
CC: Hannes Reinecke
CC: "James E.J. Bottomley"
CC: "Martin K. Petersen"
CC: linux-scsi@vger
kmemdup has implemented the function that kmalloc() + memcpy().
We prefer to kmemdup rather than code opened implementation.
This issue was detected with the help of coccinelle.
Signed-off-by: Wen Yang
CC: Hannes Reinecke
CC: "James E.J. Bottomley"
CC: "Martin K. Petersen"
CC: linux-scsi@vger
On Thu, 6 Dec 2018 15:25:42 +0300, Roman Bolshakov wrote:
> > /*
> > + * STANDARD and VPD page 0x80 T10 Vendor Identification
>
> Perhaps you meant 0x83 (Device Identification VPD page, T10 vendor ID
> based designator). INQUIRY page 0x80 is Unit Serial Number.
Indeed, the comment should refe
On Wed, Dec 05, 2018 at 01:18:35PM +0100, David Disseldorp wrote:
> In preparation for supporting user provided vendor strings, add an extra
> byte to the vendor, model and revision arrays in struct t10_wwn. This
> ensures that the full INQUIRY data can be carried in the arrays along
> with a null-
kmemdup has implemented the function that kmalloc() + memcpy().
We prefer to kmemdup rather than code opened implementation.
This issue was detected with the help of coccinelle.
Signed-off-by: Wen Yang
CC: Sathya Prakash
CC: Chaitra P B
CC: Suganath Prabu Subramani
CC: mpt-fusionlinux@bro
On Wed, Dec 05, 2018 at 01:18:36PM +0100, David Disseldorp wrote:
> The vendor_id attribute will allow for the modification of the T10
> Vendor Identification string returned in inquiry responses. Its value
> can be viewed and modified via the ConfigFS path at:
> target/core/$backstore/$name/wwn/ve
Hi,
在 2018/12/6 20:04, John Garry 写道:
On 06/12/2018 04:17, Martin K. Petersen wrote:
+
Bart,
Had you considered to use lower_32_bits() instead of "0x"?
That would to avoid that reviewers have to count the 'f'-s to verify
correctness of t10_pi_ref_tag().
I hadn't. I guess I tend
On 05/12/2018 08:01, Vivek Gautam wrote:
> On Tue, Oct 23, 2018 at 10:07 AM Can Guo wrote:
>>
>> This patch series adds support for UFS QMP PHY on SDM845 and the
>> compatible string for it. This patch series depends on the current
>> proposed QMP V3 USB3 UNI PHY support for sdm845 driver [1], on
On 06/12/2018 04:17, Martin K. Petersen wrote:
+
Bart,
Had you considered to use lower_32_bits() instead of "0x"?
That would to avoid that reviewers have to count the 'f'-s to verify
correctness of t10_pi_ref_tag().
I hadn't. I guess I tend to think of lower_32_bits() as something
Dne 6.12.2018 v 11:34 Maurizio Lombardi napsal(a):
> This is what I see when a cd burn operation completes:
>
This is the complete blktrace log:
11,034 0.81759 11653 D W 63488 (2a 00 00 03 3c 29 00 00
1f 00 ..) [wodim]
11,034 0.81759 11653 D W
Hi Jens,
Dne 20.6.2018 v 16:09 Jens Axboe napsal(a):
> On 6/20/18 5:52 AM, Maurizio Lombardi wrote:
>> Hi Jens,
>>
>> Dne 23.5.2018 v 16:42 Jens Axboe napsal(a):
>>> On 5/23/18 3:19 AM, Maurizio Lombardi wrote:
Dne 22.5.2018 v 16:47 Jens Axboe napsal(a):
> It's been many years,
70 matches
Mail list logo
