Hi Martin,
On Fri, Sep 15, 2017 at 6:37 AM, Martin K. Petersen
wrote:
>
> Suganath,
>
>> Is there any update on the submitted mpt3sas patches.
>
> We are waiting for you to report back your findings on PRP vs. SGL.
We are working on this, since there is h/w
Hello, I decided to widen the coverage of my kernel testbed and put some
FC cards into servers. This one is a PCI-X QLA2340 in HP Proliant DL 380
G4 (first 64-bit generation of Proliants). I got a UBSAN warning from
qla2xxx before probing for the firmware.
This is reproducible with or without
On Mon, Sep 18, 2017 at 03:03:30PM +0800, 陈华才 wrote:
> I don't think dma_get_cache_alignment is the "absolute minimum alignment" in
> all cases. At least on MIPS/Loongson, if we use I/O coherent mode (Cached DMA
> mode), align block queue to 4Bytes is enough. If we align block queue to
>
Oh, I missed log_to_span. Well, in that case log_to_span is
_the_ candidate for moving into a separate allocation.
And in fact you're probably better off by using a sensible data
structure for it, e.g. a radix tree.
> -Original Message-
> From: Christoph Hellwig [mailto:h...@infradead.org]
> Sent: Friday, September 15, 2017 11:30 PM
> To: shuw...@redhat.com
> Cc: kashyap.de...@broadcom.com; sumit.sax...@broadcom.com;
> shivasharan.srikanteshw...@broadcom.com; j...@linux.vnet.ibm.com;
>
Looks ok,
although I really wish we could come up with some common FC code,
including making the existing FC drivers use more infrastructure
from libfc..
Reviewed-by: Christoph Hellwig
On Sun, 2017-09-17 at 20:40 +0800, Ming Lei wrote:
> "if no request has completed before the delay has expired" can't be a
> reason to rerun the queue, because the queue can still be busy.
That statement of you shows that there are important aspects of the SCSI
core and dm-mpath driver that you
On 06/09/2017 10:15, Jason Yan wrote:
Hello all, Yijing Wang handed over this topic to me. We are working
on it the last two months. We have tested the patchset for a long
time. Here is the new version.
Now the libsas hotplug has some issues, Dan Williams report
a similar bug here before
Each creation of a FCoE device increases counter which is used as a suffix
in a FCoE device name in sysfs (i.e. /sys/bus/fcoe/devices/ctlr_1).
Once this counter reaches the value of two digits (10 and larger),
get_ctlr_num() stopped working properly and a command like `fcoeadm -i`
which depends on
> Despite the first one being a false positive, her I am reporting another
> one, from a older dual P2 server with Adaptec aic7896/97 Ultra2 SCSI
> adapter and QUANTUM VIKING II 4.5WSE HDD-s.
The same happens in other machines, like IBM xSeries 305, and is still
present in 2017, in 4.14-rc1.
> Fixing this reveals another UBSAN warning from the same driver, will fix
> that too.
This is the warning - devinfo->target_offset is unsigned but
ahc_reset_channel() calls it with signed -1 for the parameter.
The patch below silences it but is it a good way?
[0.394215]
Hi,
The current SCSI quiesce isn't safe and easy to trigger I/O deadlock.
Once SCSI device is put into QUIESCE, no new request except for
RQF_PREEMPT can be dispatched to SCSI successfully, and
scsi_device_quiesce() just simply waits for completion of I/Os
dispatched to SCSI stack. It isn't
We need to pass PREEMPT flags to blk_queue_enter()
for allocating request with RQF_PREEMPT in the
following patch.
Tested-by: Cathy Avery
Tested-by: Oleksandr Natalenko
Signed-off-by: Ming Lei
---
block/blk-core.c | 10
Both two are used for legacy and blk-mq, so rename them
as .freeze_wq and .freeze_depth for avoiding to confuse
people.
No functional change.
Tested-by: Cathy Avery
Tested-by: Oleksandr Natalenko
Signed-off-by: Ming Lei
---
The only change on legacy is that blk_drain_queue() is run
from blk_freeze_queue(), which is called in blk_cleanup_queue().
So this patch removes the explicit call of __blk_drain_queue() in
blk_cleanup_queue().
Tested-by: Cathy Avery
Tested-by: Oleksandr Natalenko
The two APIs are required to allow request allocation of
RQF_PREEMPT when queue is preempt frozen.
We have to guarantee that normal freeze and preempt freeze
are run exclusive. Because for normal freezing, once
blk_freeze_queue_wait() is returned, no request can enter
queue any more.
Another
We have to preempt freeze queue in scsi_device_quiesce(),
and unfreeze in scsi_device_resume(), so call scsi_device_resume()
for the device which is quiesced by scsi_device_quiesce().
Tested-by: Cathy Avery
Tested-by: Oleksandr Natalenko
We will support to freeze queue on block legacy path too.
No functional change.
Tested-by: Cathy Avery
Tested-by: Oleksandr Natalenko
Signed-off-by: Ming Lei
---
block/bfq-iosched.c | 2 +-
block/blk-cgroup.c | 8
This usage is basically same with blk-mq, so that we can
support to freeze legacy queue easily.
Also 'wake_up_all(>mq_freeze_wq)' has to be moved
into blk_set_queue_dying() since both legacy and blk-mq
may wait on the wait queue of .mq_freeze_wq.
Tested-by: Cathy Avery
This patch just makes it explicitely.
Tested-by: Cathy Avery
Tested-by: Oleksandr Natalenko
Reviewed-by: Johannes Thumshirn
Signed-off-by: Ming Lei
---
block/blk-mq.c | 3 ++-
1 file changed, 2
Simply quiesing SCSI device and waiting for completeion of IO
dispatched to SCSI queue isn't safe, it is easy to use up
request pool because all allocated requests before can't
be dispatched when device is put in QIUESCE. Then no request
can be allocated for RQF_PREEMPT, and system may hang
REQF_PREEMPT is a bit special because the request is required
to be dispatched to lld even when SCSI device is quiesced.
So this patch introduces __blk_get_request() to allow block
layer to allocate request when queue is preempt frozen, since we
will preempt freeze queue before quiescing SCSI
Hi, Christoph,
I don't think dma_get_cache_alignment is the "absolute minimum alignment" in
all cases. At least on MIPS/Loongson, if we use I/O coherent mode (Cached DMA
mode), align block queue to 4Bytes is enough. If we align block queue to
dma_get_cache_alignment in I/O coherent mode,
On Mon, Sep 18, 2017 at 03:18:16PM +, Bart Van Assche wrote:
> On Sun, 2017-09-17 at 20:40 +0800, Ming Lei wrote:
> > "if no request has completed before the delay has expired" can't be a
> > reason to rerun the queue, because the queue can still be busy.
>
> That statement of you shows that
24 matches
Mail list logo