[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20231219224821.GC38652@quark.localdomain>
Date: Tue, 19 Dec 2023 15:48:21 -0700
From: Eric Biggers <ebiggers@...nel.org>
To: Hongyu Jin <hongyu.jin.cn@...il.com>
Cc: agk@...hat.com, snitzer@...nel.org, mpatocka@...hat.com,
axboe@...nel.dk, zhiguo.niu@...soc.com, ke.wang@...soc.com,
yibin.ding@...soc.com, hongyu.jin@...soc.com,
linux-kernel@...r.kernel.org, dm-devel@...ts.linux.dev,
linux-block@...r.kernel.org
Subject: Re: [PATCH v5 RESEND 4/5] dm verity: Fix I/O priority lost when read
FEC and hash
On Mon, Dec 18, 2023 at 09:27:45AM +0800, Hongyu Jin wrote:
> From: Hongyu Jin <hongyu.jin@...soc.com>
>
> To fix this problem, when read FEC and hash from disk, I/O priority are
> inconsistent with data block and blocked by other I/O with low I/O
> priority.
>
> Make I/O for FEC and hash has same I/O priority with original data I/O.
"To fix this problem" is supposed to be in the second paragraph, not the first,
right?
> @@ -728,6 +730,7 @@ static void verity_submit_prefetch(struct dm_verity *v, struct dm_verity_io *io)
> sector_t block = io->block;
> unsigned int n_blocks = io->n_blocks;
> struct dm_verity_prefetch_work *pw;
> + struct bio *bio = dm_bio_from_per_bio_data(io, v->ti->per_io_data_size);
>
> if (v->validated_blocks) {
> while (n_blocks && test_bit(block, v->validated_blocks)) {
The caller has the bio pointer already, so maybe just add it as a parameter to
verity_submit_prefetch()?
- Eric
Powered by blists - more mailing lists