t se_cmd->t_prot_sg needs
to be allocated or use pre-allocated protection information by scsi
mid-layer.
Signed-off-by: Akinobu Mita
Cc: Nicholas Bellinger
Cc: Sagi Grimberg
Cc: "Martin K. Petersen"
Cc: Christoph Hellwig
Cc: "James E.J. Bottomley"
Cc: target-de...
On 4/26/2015 12:26 PM, Sagi Grimberg wrote:
On 4/25/2015 5:33 PM, Akinobu Mita wrote:
Even if the device backend is initialized with protection info is
enabled, some requests don't have the protection info attached for
WRITE SAME command issued by block device helpers, WRITE command
f-tag)?
Signed-off-by: Akinobu Mita
Cc: Nicholas Bellinger
Cc: Sagi Grimberg
Cc: "Martin K. Petersen"
Cc: Christoph Hellwig
Cc: "James E.J. Bottomley"
Cc: target-de...@vger.kernel.org
Cc: linux-scsi@vger.kernel.org
---
* Changes from v2:
- Handle odd SG mapping corre
On 4/25/2015 5:33 PM, Akinobu Mita wrote:
Now we can generate correct PI for WRITE SAME command, so it is
unnecessary to disallow WRITE SAME when protection info is enabled.
Signed-off-by: Akinobu Mita
Cc: Nicholas Bellinger
Cc: Sagi Grimberg
Cc: "Martin K. Petersen"
Cc: Christo
: Tim Chen
Cc: Herbert Xu
Cc: "David S. Miller"
Cc: linux-cry...@vger.kernel.org
Cc: Nicholas Bellinger
Cc: Sagi Grimberg
Cc: "Martin K. Petersen"
Cc: Christoph Hellwig
Cc: "James E.J. Bottomley"
Cc: target-de...@vger.kernel.org
Cc: linux-scsi@vger.kernel.or
On 4/27/2015 3:57 PM, Akinobu Mita wrote:
2015-04-26 18:44 GMT+09:00 Sagi Grimberg :
@@ -2181,6 +2182,12 @@ static inline void
transport_reset_sgl_orig(struct se_cmd *cmd)
static inline void transport_free_pages(struct se_cmd *cmd)
{
+if (!(cmd->se_cmd_fl
]
[] process_one_work+0x1db/0x780
[] worker_thread+0x11b/0x450
[] kthread+0xe4/0x100
[] ret_from_fork+0x7c/0xb0
See also patch "scsi_transport_srp: Add transport layer error
handling" (commit ID 29c17324803c).
Signed-off-by: Bart Van Assche
Cc: James Bottomley
Cc: Sagi Grimberg
Cc: Sebastian
On 4/30/2015 2:25 PM, Bart Van Assche wrote:
On 04/30/15 11:51, Sagi Grimberg wrote:
On 4/30/2015 11:58 AM, Bart Van Assche wrote:
-srp_change_conn_state(target, false);
+ch->connected = false;
Shouldn't this be protected by the channel lock (like the target)?
On
Lyakas
Signed-off-by: Sagi Grimberg
---
drivers/scsi/scsi_transport_iscsi.c |2 ++
1 files changed, 2 insertions(+), 0 deletions(-)
diff --git a/drivers/scsi/scsi_transport_iscsi.c
b/drivers/scsi/scsi_transport_iscsi.c
index 67d43e3..55647aa 100644
--- a/drivers/scsi/scsi_transport_iscsi.c
On 5/20/2015 12:38 PM, Sagi Grimberg wrote:
When creating a new endpoint, we look for a free id
for the new endpoint. We baisically loop on possible ids
and use the first id that class_find_device() returns NULL.
However, we are missing a reference put when class_find_device()
does find an
On 6/4/2015 10:06 AM, Nicholas A. Bellinger wrote:
On Wed, 2015-06-03 at 14:57 +0200, Christoph Hellwig wrote:
This makes lockdep very unhappy, rightly so. If you execute
one end_io function inside another you basŃ–cally nest every possible
lock taken in the I/O completion path. Also adding mor
On 6/9/2015 10:27 AM, Christoph Hellwig wrote:
diff --git a/drivers/target/target_core_tpg.c b/drivers/target/target_core_tpg.c
index 3fbb0d4..aa08d6b 100644
--- a/drivers/target/target_core_tpg.c
+++ b/drivers/target/target_core_tpg.c
@@ -37,6 +37,7 @@
#include
#include
+#include
#incl
Instead of open coding the sense buffer construction, use
scsi scsi_build_sense_buffer() and scsi_set_sense_information()
helpers.
This patch also fixes wrong setting of descriptor format sense data
for t10-pi integrity errors.
Signed-off-by: Sagi Grimberg
---
drivers/target/target_core_spc.c
his issue by checking the return value of iscsi_send_nopout.
If it fails set the next_timeout to one second later.
Signed-off-by: Ariel Nahum
Signed-off-by: Sagi Grimberg
---
drivers/scsi/libiscsi.c | 15 ++-
1 files changed, 10 insertions(+), 5 deletions(-)
diff --git a/drivers/s
From: Bart Van Assche
Move the code for translating a sense_reason_t code into a SCSI status
ASC and ASCQ codes from transport_send_check_condition_and_sense() into
the new function translate_sense_reason(). Convert the switch statement
that performs the translation into table-driven code.
Signe
Instead of open coding the sense buffer construction, use
scsi scsi_build_sense_buffer() and scsi_set_sense_information()
helpers which moved to scsi_common.
This patch also fixes wrong setting of descriptor format sense data
for t10-pi integrity errors.
Signed-off-by: Sagi Grimberg
Split transport_send_check_condition_and_sense()
Sagi Grimberg (1):
target: Use scsi helpers to build the sense data correctly
drivers/scsi/scsi_common.c | 58 +
drivers/scsi/scsi_error.c | 58 -
drivers/target/target_core_spc.c | 31 +--
drivers/t
From: Bart Van Assche
Inline this function in its call site since it performs a trivial
task and since it is only called once.
Signed-off-by: Bart Van Assche
---
drivers/target/target_core_transport.c | 16 ++--
1 file changed, 2 insertions(+), 14 deletions(-)
diff --git a/drivers
On 7/3/2015 7:12 PM, Christoph Hellwig wrote:
On Mon, Jun 29, 2015 at 06:05:25PM +0300, Sagi Grimberg wrote:
Instead of open coding the sense buffer construction, use
scsi scsi_build_sense_buffer() and scsi_set_sense_information()
helpers.
This patch also fixes wrong setting of descriptor
his issue by checking the return value of iscsi_send_nopout.
If it fails set the next_timeout to one second later.
Signed-off-by: Ariel Nahum
Signed-off-by: Sagi Grimberg
---
Changes from v0:
- Fixed coding style comment
drivers/scsi/libiscsi.c | 17 +++--
1 file changed, 11 inserti
On 7/4/2015 6:48 PM, Bart Van Assche wrote:
Bart, any chance you could resend them?
Bart, would you prefer to re-send your changes on top of v1 of this one?
It should be easy enough.
The modification needed is that struct sense_info needs a desc_format
bool to indicate if the sense data is c
On 7/2/2015 11:11 PM, Mike Christie wrote:
On 6/30/15, 9:55 AM, Sagi Grimberg wrote:
From: Ariel Nahum
Connection last_ping is not being updated when iscsi_send_nopout fails.
Not updating the last_ping will cause firing a timer to a past time
(last_ping + ping_tmo < current_time) wh
also moved include to scsi_common.h
Changes from v0:
- Added Bart's patches and converted my patch to apply over his
- Moved scsi sense helpers to scsi_common
Bart Van Assche (2):
target: Inline transport_get_sense_codes()
target: Split transport_send_check_condition_and_sense()
Instead of open coding the sense buffer construction, use
scsi scsi_build_sense_buffer() and scsi_set_sense_information()
helpers which moved to scsi_common.
This patch also fixes wrong setting of descriptor format sense data
for t10-pi integrity errors.
Signed-off-by: Sagi Grimberg
From: Bart Van Assche
Move the code for translating a sense_reason_t code into a SCSI status
ASC and ASCQ codes from transport_send_check_condition_and_sense() into
the new function translate_sense_reason(). Convert the switch statement
that performs the translation into table-driven code.
Signe
From: Bart Van Assche
Inline this function in its call site since it performs a trivial
task and since it is only called once.
Signed-off-by: Bart Van Assche
Reviewed-by: Hannes Reinecke
---
drivers/target/target_core_transport.c | 16 ++--
1 file changed, 2 insertions(+), 14 dele
On 7/6/2015 11:59 AM, Christoph Hellwig wrote:
On Mon, Jul 06, 2015 at 11:02:25AM +0300, Sagi Grimberg wrote:
From: Bart Van Assche
Inline this function in its call site since it performs a trivial
task and since it is only called once.
Signed-off-by: Bart Van Assche
Reviewed-by: Hannes
On 7/6/2015 12:06 PM, Christoph Hellwig wrote:
On Mon, Jul 06, 2015 at 11:02:27AM +0300, Sagi Grimberg wrote:
Instead of open coding the sense buffer construction, use
scsi scsi_build_sense_buffer() and scsi_set_sense_information()
helpers which moved to scsi_common.
This patch also fixes
.
Signed-off-by: Bart Van Assche
Signed-off-by: Sagi Grimberg
Reviewed-by: Hannes Reinecke
---
drivers/target/target_core_transport.c | 383 +
1 file changed, 148 insertions(+), 235 deletions(-)
diff --git a/drivers/target/target_core_transport.c
b/drivers/target
ed my patch to apply over his
- Moved scsi sense helpers to scsi_common
Bart Van Assche (2):
target: Inline transport_get_sense_codes()
target: Split transport_send_check_condition_and_sense()
Sagi Grimberg (3):
scsi: Move sense handling routines to scsi_common
target: Use scsi helpers to buil
PI errors should be reported in a descriptor format sense data.
Fix that by adding a desc_format flag to struct sense_info and pass
it to scsi_build_sense_buffer() to do the right thing.
Signed-off-by: Sagi Grimberg
---
drivers/target/target_core_transport.c | 6 +-
1 file changed, 5
Sense data handling is also done in the target stack.
Hence, move sense handling routines to scsi_common so
the target will be able to use them as well.
Signed-off-by: Sagi Grimberg
---
drivers/scsi/scsi_common.c | 98 +
drivers/scsi/scsi_error.c
Instead of open coding the sense buffer construction, use
scsi scsi_build_sense_buffer() and scsi_set_sense_information()
helpers which moved to scsi_common.
Signed-off-by: Sagi Grimberg
---
drivers/target/target_core_spc.c | 31 +--
drivers/target
From: Bart Van Assche
Inline this function in its call site since it performs a trivial
task and since it is only called once.
Signed-off-by: Bart Van Assche
Signed-off-by: Sagi Grimberg
Reviewed-by: Hannes Reinecke
Reviewed-by: Christoph Hellwig
---
drivers/target/target_core_transport.c
On 7/6/2015 6:28 PM, Bart Van Assche wrote:
On 07/06/2015 06:15 AM, Sagi Grimberg wrote:
diff --git a/drivers/target/target_core_transport.c
b/drivers/target/target_core_transport.c
index 0181f8b..79bb8d1 100644
--- a/drivers/target/target_core_transport.c
+++ b/drivers/target
On 7/7/2015 1:51 AM, Nicholas A. Bellinger wrote:
On Mon, 2015-07-06 at 16:15 +0300, Sagi Grimberg wrote:
This patch set modifies the target sense data handling.
First, cleanup transport_send_check_condition_and_sense()
by splitting the sense translation to a separate function.
Second, convert
Hi,
I recently came across a lockdep complaint on a possible deadlock that
I don't understand yet (log output below).
Has anyone seen this? Any insights on why exactly is lockdep complaining?
The reproducer is very easy (reproduced both with tcp and iser):
- login to a target with one or more d
On 7/7/2015 12:01 PM, Nicholas A. Bellinger wrote:
Hey Sagi,
This addresses a regression with traditional iscsi-target that I noticed
recently, but has not been tested with iser-target yet.
Would you mind taking a quick spin with iser-target to verify, and try
to intentionally fail iscsit_start
On 7/8/2015 10:41 AM, Hannes Reinecke wrote:
If a disk reports an ALUA 'transitioning' state we should not
try to spin up the device.
Signed-off-by: Hannes Reinecke
---
drivers/scsi/sd.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/scsi/sd.c b/drivers/scsi/sd.c
index 7c0bdaa
On 7/8/2015 1:17 PM, Christoph Hellwig wrote:
--- a/include/target/target_core_base.h
+++ b/include/target/target_core_base.h
@@ -7,6 +7,7 @@
#include
#include
#include
+#include
#include
#include
Please only add the include in the files that need it. (And many of the
existing
On 7/8/2015 1:15 PM, Christoph Hellwig wrote:
+ if (r == (__force int)TCM_CHECK_CONDITION_UNIT_ATTENTION) {
You probably want to compare reason here to avoid the cast.
I do want it...
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to m
On 7/8/2015 1:19 PM, Christoph Hellwig wrote:
On Mon, Jul 06, 2015 at 04:15:08PM +0300, Sagi Grimberg wrote:
PI errors should be reported in a descriptor format sense data.
Fix that by adding a desc_format flag to struct sense_info and pass
it to scsi_build_sense_buffer() to do the right thing
On 7/8/2015 1:59 PM, Hannes Reinecke wrote:
On 07/08/2015 12:49 PM, Christoph Hellwig wrote:
On Wed, Jul 08, 2015 at 01:36:04PM +0300, Sagi Grimberg wrote:
We don't have any other information today, but sector is not the only
information that is requires a descriptor format, so maybe it
On 7/8/2015 2:44 PM, Christoph Hellwig wrote:
On Wed, Jul 08, 2015 at 12:59:18PM +0200, Hannes Reinecke wrote:
Actually it's controlled by the D_SENSE bit in the Control mode page
(that's bit[2] of byte 2 in the control mode page).
Which is currently set to '0', ie we will be returning fixed sen
On 7/8/2015 2:14 PM, Sagi Grimberg wrote:
And it's actually not true that you'd need descriptor sense to
encode the sector information; it'll be stored in the 'information'
section (byte 3-6) for fixed format sense.
But when I return the sector info in a fixed size
et: Inline transport_get_sense_codes()
target: Split transport_send_check_condition_and_sense()
Sagi Grimberg (3):
scsi: Move sense handling routines to scsi_common
target: Use scsi helpers to build the sense data correctly
target: Return ABORTED_COMMAND sense key for PI errors
drivers/scsi/scs
From: Bart Van Assche
Inline this function in its call site since it performs a trivial
task and since it is only called once.
Signed-off-by: Bart Van Assche
Signed-off-by: Sagi Grimberg
Reviewed-by: Hannes Reinecke
Reviewed-by: Christoph Hellwig
---
drivers/target/target_core_transport.c
.
Signed-off-by: Bart Van Assche
Signed-off-by: Sagi Grimberg
Reviewed-by: Hannes Reinecke
Reviewed-by: Christoph Hellwig
---
drivers/target/target_core_transport.c | 383 +
1 file changed, 148 insertions(+), 235 deletions(-)
diff --git a/drivers/target
PI errors were reported with ILLEGAL_REQUEST sense key but
there was actually no problem with the request. Target
detected PI errors should be reported with aborted command
sense key.
Signed-off-by: Sagi Grimberg
---
drivers/target/target_core_transport.c | 6 +++---
1 file changed, 3
Sense data handling is also done in the target stack.
Hence, move sense handling routines to scsi_common so
the target will be able to use them as well.
Signed-off-by: Sagi Grimberg
Reviewed-by: Bart Van Assche
Reviewed-by: Christoph Hellwig
---
drivers/scsi/scsi_common.c | 98
Instead of open coding the sense buffer construction, use
scsi scsi_build_sense_buffer() and scsi_set_sense_information()
helpers which moved to scsi_common.
Signed-off-by: Sagi Grimberg
Reviewed-by: Christoph Hellwig
---
drivers/target/target_core_spc.c | 31
In fixed size sense format the information field is a four byte
field.
Signed-off-by: Sagi Grimberg
---
drivers/scsi/scsi_common.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/scsi/scsi_common.c b/drivers/scsi/scsi_common.c
index 41432c1..8cfb7ee 100644
--- a
On 7/8/2015 6:06 PM, Hannes Reinecke wrote:
We're adding extra fields here, so we need to make sure to not
overflow the buffer. You probably have to pass in the buffersize
to avoid an overflow ...
Yeah, I know, it's theoretical at the moment.
But there's nothing which prevents anyone to add othe
Fix this issue by checking the return value of iscsi_send_nopout.
If it fails set the next_timeout to one second later.
Signed-off-by: Ariel Nahum
Signed-off-by: Sagi Grimberg
---
Changes from v0:
- Fixed coding style comment
drivers/scsi/libiscsi.c | 17 +++--
1 file changed,
This looks correct to me,
Reviewed-by: Christoph Hellwig
But: this will truncate > 32bit sector numbers. Maybe we need to
enable descriptor format sense data for large enough LUs.
That should be easy enough now that the sense data is constructed using
scsi helpers.
Let me prepare a patch
report descriptor format sense
data to avoid 64bit sector info truncation (reported by hch).
Sagi Grimberg (3):
scsi: Fix wrong additional sense length in descriptor format
scsi: Protect against buffer possible overflow in
scsi_set_sense_information
target: Return descriptor format sense
In case encountered a PI error, use scsi_set_sense_information
instead of open coding information descriptor format.
Signed-off-by: Sagi Grimberg
---
drivers/scsi/libiscsi.c | 9 +++--
1 file changed, 3 insertions(+), 6 deletions(-)
diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi
Make sure that the input sense buffer has sufficient length
to fit the information descriptor (12 additional bytes).
Modify scsi_set_sense_information to receive the sense buffer
length and adjust it's callers scsi target and libata.
Reported-by: Hannes Reinecke
Signed-off-by: Sagi Grimber
Fixed size sense data information field is only 32 bits which
means the sector (64 bits) information will be truncated.
Move to descriptor format sense data to correctly report full
sector information.
Reported-by: Christoph Hellwig
Signed-off-by: Sagi Grimberg
---
drivers/target
The sense header additional sense length should be the accumulated
size of all the descriptors. Information descriptor size is 12 bytes.
When setting the additional sense length we should add 0xc instead of
0xa.
Signed-off-by: Sagi Grimberg
---
drivers/scsi/scsi_common.c | 2 +-
1 file changed
se data information
field is only 32 bits which means the sector information will be truncated.
Thus, if the LU spans 64bit sectors, use descriptor format sense data to
correctly report sector information.
Reported-by: Christoph Hellwig
Signed-off-by: Sagi Grimberg
---
drivers/target/target_core
In case encountered a PI error, use scsi_set_sense_information
instead of open coding information descriptor format.
Signed-off-by: Sagi Grimberg
---
drivers/scsi/libiscsi.c | 9 +++--
1 file changed, 3 insertions(+), 6 deletions(-)
diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi
-by: Sagi Grimberg
---
drivers/target/target_core_hba.c | 5 +
drivers/target/target_core_spc.c | 12 +---
drivers/target/target_core_transport.c | 3 ++-
include/target/target_core_backend.h | 2 ++
4 files changed, 18 insertions(+), 4 deletions(-)
diff --git a
The sense header additional sense length should be the accumulated
size of all the descriptors. Information descriptor size is 12 bytes.
When setting the additional sense length we should add 0xc instead of
0xa.
Signed-off-by: Sagi Grimberg
Reviewed-by: Hannes Reinecke
Reviewed-by: Martin K
format only for LUs that
span 64bit sectors.
- Added reviwed-by tags.
Sagi Grimberg (4):
scsi: Fix wrong additional sense length in descriptor format
scsi: Protect against buffer possible overflow in
scsi_set_sense_information
target: Return descriptor format sense data in case the LU spans
Make sure that the input sense buffer has sufficient length
to fit the information descriptor (12 additional bytes).
Modify scsi_set_sense_information to receive the sense buffer
length and adjust its callers scsi target and libata.
Reported-by: Hannes Reinecke
Signed-off-by: Sagi Grimberg
-by: Sagi Grimberg
---
drivers/target/target_core_hba.c | 5 +
drivers/target/target_core_spc.c | 12 +---
drivers/target/target_core_transport.c | 3 ++-
include/target/target_core_backend.h | 2 ++
4 files changed, 18 insertions(+), 4 deletions(-)
diff --git a
G list in the target code?
In any event, looks good,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Signed-off-by: Joern Engel
Signed-off-by: Spencer Baugh
Looks good,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
- put_unaligned_be64(info, &buf[3]);
+ /*
+* Fixed format sense reserves only 32 bits for the
+* 'information' field
+*/
+ put_unaligned_be32((u32)info, &buf[3]);
}
}
EXPORT_SYMBOL(scsi_set_sens
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Thanks for moving it,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Looks good,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Looks fine,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi Robert,
Hey Robert, Christoph,
please explain your use cases that isn't handled. The one and only
reason to set MSDBD to 1 is to make the code a lot simpler given that
there is no real use case for supporting more.
RDMA uses memory registrations to register large and possibly
discontigu
Looks good,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Hey Bart,
The solution I prefer is to modify the SCSI scanning code such that
the scan_mutex is only held while performing the actual LUN scanning
and while ensuring that no SCSI device has been created yet for a
certain LUN number but not while the Linux device and its sysfs
attributes are crea
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Again,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Looks useful,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
and again,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Hi all,
I'd like to attend LSF/MM and would like to discuss polling for block drivers.
Currently there is blk-iopoll but it is neither as widely used as NAPI in the
networking field and accoring to Sagi's findings in [1] performance with
polling is not on par with IRQ usage.
On LSF/MM I'd lik
I'd like to attend LSF/MM and would like to discuss polling for block
drivers.
Currently there is blk-iopoll but it is neither as widely used as NAPI in
the networking field and accoring to Sagi's findings in [1] performance
with polling is not on par with IRQ usage.
On LSF/MM I'd like to whet
A typical Ethernet network adapter delays the generation of an
interrupt
after it has received a packet. A typical block device or HBA does not
delay
the generation of an interrupt that reports an I/O completion.
>>>
>>> NVMe allows for configurable interrupt coalescing, as
Hi Folks,
I would like to propose a general discussion on Storage stack and device driver
testing.
I think its very useful and needed.
Purpose:-
-
The main objective of this discussion is to address the need for
a Unified Test Automation Framework which can be used by different
I agree with Jens that we'll need some analysis if we want the
discussion to be affective, and I can spend some time this if I
can find volunteers with high-end nvme devices (I only have access
to client nvme devices.
I have a P3700 but somehow burned the FW. Let me see if I can bring it back
**Note: when I ran multiple threads on more cpus the performance
degradation phenomenon disappeared, but I tested on a VM with
qemu emulation backed by null_blk so I figured I had some other
bottleneck somewhere (that's why I asked for some more testing).
That could be because of the vmexits a
Add a helper to calculate the actual data transfer size for special
payload requests.
Signed-off-by: Christoph Hellwig
---
include/linux/blkdev.h | 13 +
1 file changed, 13 insertions(+)
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index ff3d774..1ca8e8f 100644
--
Looks good,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
This looks good,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Looks good,
Reviewed-by: Sagi Grimberg
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
--
[1]
queue = b'nvme0q1'
usecs : count distribution
0 -> 1 : 7310 ||
2 -> 3 : 11 | |
4 -> 7 : 10 | |
8 -> 15 : 20 | |
16
Hey, so I made some initial analysis of whats going on with
irq-poll.
First, I sampled how much time it takes before we
get the interrupt in nvme_irq and the initial visit
to nvme_irqpoll_handler. I ran a single threaded fio
with QD=32 of 4K reads. This is two displays of a
histogram of the laten
1 - 100 of 549 matches
Mail list logo
