* [PATCH] null_blk: add zoned config support information
@ 2019-01-04 15:42 John Pittman
2019-01-04 15:59 ` Laurence Oberman
2019-01-04 16:46 ` Bart Van Assche
0 siblings, 2 replies; 8+ messages in thread
From: John Pittman @ 2019-01-04 15:42 UTC (permalink / raw)
To: axboe; +Cc: linux-block, John Pittman
If the kernel is built without CONFIG_BLK_DEV_ZONED, a modprobe
of the null_blk driver with zoned=1 fails with 'Invalid argument'.
This can be confusing to users, prompting a search as to why the
parameter is invalid. To assist in that search, add a bit more
information to the failure, additionally adding to the documentation
that CONFIG_BLK_DEV_ZONED is needed for zoned=1.
Signed-off-by: John Pittman <jpittman@redhat.com>
---
Documentation/block/null_blk.txt | 3 ++-
drivers/block/null_blk.h | 1 +
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/Documentation/block/null_blk.txt b/Documentation/block/null_blk.txt
index ea2dafe49ae8..4cad1024fff7 100644
--- a/Documentation/block/null_blk.txt
+++ b/Documentation/block/null_blk.txt
@@ -88,7 +88,8 @@ shared_tags=[0/1]: Default: 0
zoned=[0/1]: Default: 0
0: Block device is exposed as a random-access block device.
- 1: Block device is exposed as a host-managed zoned block device.
+ 1: Block device is exposed as a host-managed zoned block device. Requires
+ CONFIG_BLK_DEV_ZONED.
zone_size=[MB]: Default: 256
Per zone size when exposed as a zoned block device. Must be a power of two.
diff --git a/drivers/block/null_blk.h b/drivers/block/null_blk.h
index b3df2793e7cd..cab4808f14bd 100644
--- a/drivers/block/null_blk.h
+++ b/drivers/block/null_blk.h
@@ -97,6 +97,7 @@ void null_zone_reset(struct nullb_cmd *cmd, sector_t sector);
#else
static inline int null_zone_init(struct nullb_device *dev)
{
+ pr_info("CONFIG_BLK_DEV_ZONED not enabled\n");
return -EINVAL;
}
static inline void null_zone_exit(struct nullb_device *dev) {}
--
2.17.2
^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [PATCH] null_blk: add zoned config support information
2019-01-04 15:42 [PATCH] null_blk: add zoned config support information John Pittman
@ 2019-01-04 15:59 ` Laurence Oberman
2019-01-04 16:46 ` Bart Van Assche
1 sibling, 0 replies; 8+ messages in thread
From: Laurence Oberman @ 2019-01-04 15:59 UTC (permalink / raw)
To: John Pittman, axboe; +Cc: linux-block
On Fri, 2019-01-04 at 10:42 -0500, John Pittman wrote:
> If the kernel is built without CONFIG_BLK_DEV_ZONED, a modprobe
> of the null_blk driver with zoned=1 fails with 'Invalid argument'.
> This can be confusing to users, prompting a search as to why the
> parameter is invalid. To assist in that search, add a bit more
> information to the failure, additionally adding to the documentation
> that CONFIG_BLK_DEV_ZONED is needed for zoned=1.
>
> Signed-off-by: John Pittman <jpittman@redhat.com>
> ---
> Documentation/block/null_blk.txt | 3 ++-
> drivers/block/null_blk.h | 1 +
> 2 files changed, 3 insertions(+), 1 deletion(-)
>
> diff --git a/Documentation/block/null_blk.txt
> b/Documentation/block/null_blk.txt
> index ea2dafe49ae8..4cad1024fff7 100644
> --- a/Documentation/block/null_blk.txt
> +++ b/Documentation/block/null_blk.txt
> @@ -88,7 +88,8 @@ shared_tags=[0/1]: Default: 0
>
> zoned=[0/1]: Default: 0
> 0: Block device is exposed as a random-access block device.
> - 1: Block device is exposed as a host-managed zoned block device.
> + 1: Block device is exposed as a host-managed zoned block device.
> Requires
> + CONFIG_BLK_DEV_ZONED.
>
> zone_size=[MB]: Default: 256
> Per zone size when exposed as a zoned block device. Must be a
> power of two.
> diff --git a/drivers/block/null_blk.h b/drivers/block/null_blk.h
> index b3df2793e7cd..cab4808f14bd 100644
> --- a/drivers/block/null_blk.h
> +++ b/drivers/block/null_blk.h
> @@ -97,6 +97,7 @@ void null_zone_reset(struct nullb_cmd *cmd,
> sector_t sector);
> #else
> static inline int null_zone_init(struct nullb_device *dev)
> {
> + pr_info("CONFIG_BLK_DEV_ZONED not enabled\n");
> return -EINVAL;
> }
> static inline void null_zone_exit(struct nullb_device *dev) {}
Looks good to me, useful change in my opinion.
Reviewed-by: Laurence Oberman <loberman@redhat.com>
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] null_blk: add zoned config support information
2019-01-04 15:42 [PATCH] null_blk: add zoned config support information John Pittman
2019-01-04 15:59 ` Laurence Oberman
@ 2019-01-04 16:46 ` Bart Van Assche
2019-01-04 16:58 ` Laurence Oberman
2019-01-04 20:37 ` John Pittman
1 sibling, 2 replies; 8+ messages in thread
From: Bart Van Assche @ 2019-01-04 16:46 UTC (permalink / raw)
To: John Pittman, axboe; +Cc: linux-block
On Fri, 2019-01-04 at 10:42 -0500, John Pittman wrote:
> static inline int null_zone_init(struct nullb_device *dev)
> {
> + pr_info("CONFIG_BLK_DEV_ZONED not enabled\n");
> return -EINVAL;
> }
Have you considered to use pr_err() instead of pr_info()?
Thanks,
Bart.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] null_blk: add zoned config support information
2019-01-04 16:46 ` Bart Van Assche
@ 2019-01-04 16:58 ` Laurence Oberman
2019-01-04 20:37 ` John Pittman
1 sibling, 0 replies; 8+ messages in thread
From: Laurence Oberman @ 2019-01-04 16:58 UTC (permalink / raw)
To: Bart Van Assche, John Pittman, axboe; +Cc: linux-block
On Fri, 2019-01-04 at 08:46 -0800, Bart Van Assche wrote:
> On Fri, 2019-01-04 at 10:42 -0500, John Pittman wrote:
> > static inline int null_zone_init(struct nullb_device *dev)
> > {
> > + pr_info("CONFIG_BLK_DEV_ZONED not enabled\n");
> > return -EINVAL;
> > }
>
> Have you considered to use pr_err() instead of pr_info()?
>
> Thanks,
>
> Bart.
Bart, good point, Thank you
Laurence
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] null_blk: add zoned config support information
2019-01-04 16:46 ` Bart Van Assche
2019-01-04 16:58 ` Laurence Oberman
@ 2019-01-04 20:37 ` John Pittman
2019-01-04 20:49 ` Bart Van Assche
2019-01-06 19:23 ` Jens Axboe
1 sibling, 2 replies; 8+ messages in thread
From: John Pittman @ 2019-01-04 20:37 UTC (permalink / raw)
To: Bart Van Assche; +Cc: Jens Axboe, linux-block
Thanks Bart; I made the changes and sent them in as a v2, I'm sure you
already saw. I have a quick, unrelated question if you have a moment.
In testing the null_blk driver, I found that trim commands sent by fio
were rejected due to lack of support. Tracking down Shaohua's commit
306eb6b4a ("nullb: support discard"), he mentions that "discard makes
sense for memory backed disk". Just to see what would happen, I
edited the source to make discard a configurable parameter at
modprobe, and after the edit & build, the trim commands submitted
fine. Does this sort of change make sense? I mean the ability to do
discard to null_blk without it being memory backed; solely for
testing/benchmarking purposes. I haven't found any good instructions
on creating a memory backed or discard enabled null_blk device from
the command line, so I assume a higher level driver would have to hook
in and enable these features manually. Thanks for your time and any
information.
John
On Fri, Jan 4, 2019 at 11:47 AM Bart Van Assche <bvanassche@acm.org> wrote:
>
> On Fri, 2019-01-04 at 10:42 -0500, John Pittman wrote:
> > static inline int null_zone_init(struct nullb_device *dev)
> > {
> > + pr_info("CONFIG_BLK_DEV_ZONED not enabled\n");
> > return -EINVAL;
> > }
>
> Have you considered to use pr_err() instead of pr_info()?
>
> Thanks,
>
> Bart.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] null_blk: add zoned config support information
2019-01-04 20:37 ` John Pittman
@ 2019-01-04 20:49 ` Bart Van Assche
2019-01-04 21:07 ` John Pittman
2019-01-06 19:23 ` Jens Axboe
1 sibling, 1 reply; 8+ messages in thread
From: Bart Van Assche @ 2019-01-04 20:49 UTC (permalink / raw)
To: John Pittman; +Cc: Jens Axboe, linux-block
On Fri, 2019-01-04 at 15:37 -0500, John Pittman wrote:
> Thanks Bart; I made the changes and sent them in as a v2, I'm sure you
> already saw. I have a quick, unrelated question if you have a moment.
> In testing the null_blk driver, I found that trim commands sent by fio
> were rejected due to lack of support. Tracking down Shaohua's commit
> 306eb6b4a ("nullb: support discard"), he mentions that "discard makes
> sense for memory backed disk". Just to see what would happen, I
> edited the source to make discard a configurable parameter at
> modprobe, and after the edit & build, the trim commands submitted
> fine. Does this sort of change make sense? I mean the ability to do
> discard to null_blk without it being memory backed; solely for
> testing/benchmarking purposes. I haven't found any good instructions
> on creating a memory backed or discard enabled null_blk device from
> the command line, so I assume a higher level driver would have to hook
> in and enable these features manually. Thanks for your time and any
> information.
Hi John,
Jens as the block layer maintainer has the last word about this. Personally
I would welcome that functionality. Before discard functionality was removed
from the brd driver I used the brd driver to test the discard functionality
in storage target stacks. If discard functionality would be added to the
null_blk driver then that would make it possible to use that driver for
testing the discard functionality in e.g. LIO. See also commit f09a06a193d9
("brd: remove discard support").
I'm not sure that we need a modprobe parameter to enable or disable trim
functionality in the null_blk driver. I'm fine with always enabling trim
functionality in that driver.
Bart.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] null_blk: add zoned config support information
2019-01-04 20:49 ` Bart Van Assche
@ 2019-01-04 21:07 ` John Pittman
0 siblings, 0 replies; 8+ messages in thread
From: John Pittman @ 2019-01-04 21:07 UTC (permalink / raw)
To: Bart Van Assche; +Cc: Jens Axboe, linux-block
Thanks Bart, I appreciate it.
On Fri, Jan 4, 2019 at 3:49 PM Bart Van Assche <bvanassche@acm.org> wrote:
>
> On Fri, 2019-01-04 at 15:37 -0500, John Pittman wrote:
> > Thanks Bart; I made the changes and sent them in as a v2, I'm sure you
> > already saw. I have a quick, unrelated question if you have a moment.
> > In testing the null_blk driver, I found that trim commands sent by fio
> > were rejected due to lack of support. Tracking down Shaohua's commit
> > 306eb6b4a ("nullb: support discard"), he mentions that "discard makes
> > sense for memory backed disk". Just to see what would happen, I
> > edited the source to make discard a configurable parameter at
> > modprobe, and after the edit & build, the trim commands submitted
> > fine. Does this sort of change make sense? I mean the ability to do
> > discard to null_blk without it being memory backed; solely for
> > testing/benchmarking purposes. I haven't found any good instructions
> > on creating a memory backed or discard enabled null_blk device from
> > the command line, so I assume a higher level driver would have to hook
> > in and enable these features manually. Thanks for your time and any
> > information.
>
> Hi John,
>
> Jens as the block layer maintainer has the last word about this. Personally
> I would welcome that functionality. Before discard functionality was removed
> from the brd driver I used the brd driver to test the discard functionality
> in storage target stacks. If discard functionality would be added to the
> null_blk driver then that would make it possible to use that driver for
> testing the discard functionality in e.g. LIO. See also commit f09a06a193d9
> ("brd: remove discard support").
>
> I'm not sure that we need a modprobe parameter to enable or disable trim
> functionality in the null_blk driver. I'm fine with always enabling trim
> functionality in that driver.
>
> Bart.
^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH] null_blk: add zoned config support information
2019-01-04 20:37 ` John Pittman
2019-01-04 20:49 ` Bart Van Assche
@ 2019-01-06 19:23 ` Jens Axboe
1 sibling, 0 replies; 8+ messages in thread
From: Jens Axboe @ 2019-01-06 19:23 UTC (permalink / raw)
To: John Pittman, Bart Van Assche; +Cc: linux-block
On 1/4/19 1:37 PM, John Pittman wrote:
> Thanks Bart; I made the changes and sent them in as a v2, I'm sure you
> already saw. I have a quick, unrelated question if you have a moment.
> In testing the null_blk driver, I found that trim commands sent by fio
> were rejected due to lack of support. Tracking down Shaohua's commit
> 306eb6b4a ("nullb: support discard"), he mentions that "discard makes
> sense for memory backed disk". Just to see what would happen, I
> edited the source to make discard a configurable parameter at
> modprobe, and after the edit & build, the trim commands submitted
> fine. Does this sort of change make sense? I mean the ability to do
> discard to null_blk without it being memory backed; solely for
> testing/benchmarking purposes. I haven't found any good instructions
> on creating a memory backed or discard enabled null_blk device from
> the command line, so I assume a higher level driver would have to hook
> in and enable these features manually. Thanks for your time and any
> information.
I think that would be fine, no option needed. It's really no different
than reads/writes being completed without having a memory backing.
--
Jens Axboe
^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2019-01-06 19:23 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-01-04 15:42 [PATCH] null_blk: add zoned config support information John Pittman
2019-01-04 15:59 ` Laurence Oberman
2019-01-04 16:46 ` Bart Van Assche
2019-01-04 16:58 ` Laurence Oberman
2019-01-04 20:37 ` John Pittman
2019-01-04 20:49 ` Bart Van Assche
2019-01-04 21:07 ` John Pittman
2019-01-06 19:23 ` Jens Axboe
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).