[dm-devel] [PATCH v2 0/3] brd discard patches

[dm-devel] [PATCH v2 0/3] brd discard patches

[Mikulas Patocka]

This is a new version of the brd discard patches.
--
dm-devel mailing list
dm-devel@redhat.com
https://listman.redhat.com/mailman/listinfo/dm-devel

Re: [dm-devel] [PATCH v2 0/3] brd discard patches

[Jens Axboe]

On 7/21/23 7:48?AM, Mikulas Patocka wrote:
> This is a new version of the brd discard patches.

Can you please:

1) Ensure that your postings thread properly, it's all separate emails
   and the patches don't nest under the cover letter parent.

2) Include a changelog. What changed since v1?

-- 
Jens Axboe

--
dm-devel mailing list
dm-devel@redhat.com
https://listman.redhat.com/mailman/listinfo/dm-devel

Re: [dm-devel] [PATCH v2 0/3] brd discard patches

[Mikulas Patocka]

On Fri, 21 Jul 2023, Jens Axboe wrote:

> On 7/21/23 7:48?AM, Mikulas Patocka wrote:
> > This is a new version of the brd discard patches.
> 
> Can you please:
> 
> 1) Ensure that your postings thread properly, it's all separate emails
>    and the patches don't nest under the cover letter parent.

I use alpine. I was testing it - and it turns out that when I delete the 
last character in the "Subject" field, it deletes the field "In-Reply-To" 
from the header. This must be some new bug/feature - it didn't do it in 
the past.

I'll try to be more careful to not make the "Subject" field empty when 
sending patches.

> 2) Include a changelog. What changed since v1?
> 
> -- 
> Jens Axboe

ChangeLog:

* Batch discarded pages into the "free_page_batch" structure and free all 
  of them with just one "call_rcu" call. In case of allocation failure, 
  fall back to per-page "call_rcu" calls. (suggested by Christoph Hellwig)

* Make the module parameter "/sys/module/brd/parameters/discard"  
  changeable at runtime. Changing it will iterate over all ramdisk devices 
  and call brd_set_discard_limits on them to enable or disable discard.
  (suggested by Christoph Hellwig)

* Use "switch (bio_op(bio))" in brd_submit_bio, so that the code looks 
  better. (suggested by Chaitanya Kulkarni)

* do "bio->bi_status = BLK_STS_NOTSUPP" in brd_submit_bio if unknown type 
  of bio is received.

Mikulas
--
dm-devel mailing list
dm-devel@redhat.com
https://listman.redhat.com/mailman/listinfo/dm-devel