linux-kernel - Re: [PATCH v5 4/4] io_uring: remove ring quiesce for io_uring

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Message-ID: <130345c9-e8fb-2f13-59fd-368334b49235@gmail.com>
Date:   Fri, 4 Feb 2022 00:28:51 +0000
From:   Pavel Begunkov <asml.silence@...il.com>
To:     Usama Arif <usama.arif@...edance.com>, io-uring@...r.kernel.org,
        axboe@...nel.dk, linux-kernel@...r.kernel.org
Cc:     fam.zheng@...edance.com
Subject: Re: [PATCH v5 4/4] io_uring: remove ring quiesce for
 io_uring_register

On 2/3/22 23:47, Pavel Begunkov wrote:
> On 2/3/22 23:34, Usama Arif wrote:
>> Ring quiesce is currently only used for 2 opcodes
>> IORING_REGISTER_ENABLE_RINGS and IORING_REGISTER_RESTRICTIONS.
>> IORING_SETUP_R_DISABLED prevents submitting requests and
>> so there will be no requests until IORING_REGISTER_ENABLE_RINGS
>> is called. And IORING_REGISTER_RESTRICTIONS works only before
>> IORING_REGISTER_ENABLE_RINGS is called. Hence ring quiesce is
>> not needed for these opcodes and therefore io_uring_register.
> 
> I think I'd prefer to retain quiesce code than reverting this
> patch later.

btw, if it gets to reverting it'll be easier if this patch
is split in 2. The first puts these 2 opcodes into
io_register_op_must_quiesce(), we definitely want to keep
that. And the other doing the rest of cleanup

-- 
Pavel Begunkov