netdev - Re: [PATCH v2 2/5] io_uring/zcrx: add refcount to struct io_zcrx

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 for Android: free password hash cracker in your pocket

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

Message-ID: <74ce4fb9-3654-4a1d-9b8b-abee8aba9ca9@davidwei.uk>
Date: Sat, 25 Oct 2025 21:10:41 -0700
From: David Wei <dw@...idwei.uk>
To: Jens Axboe <axboe@...nel.dk>, io-uring@...r.kernel.org,
 netdev@...r.kernel.org
Cc: Pavel Begunkov <asml.silence@...il.com>
Subject: Re: [PATCH v2 2/5] io_uring/zcrx: add refcount to struct io_zcrx_ifq

On 2025-10-25 16:37, Jens Axboe wrote:
> On 10/25/25 1:15 PM, David Wei wrote:
>> diff --git a/io_uring/zcrx.c b/io_uring/zcrx.c
>> index a816f5902091..22d759307c16 100644
>> --- a/io_uring/zcrx.c
>> +++ b/io_uring/zcrx.c
>> @@ -730,6 +731,8 @@ void io_shutdown_zcrx_ifqs(struct io_ring_ctx *ctx)
>>   	lockdep_assert_held(&ctx->uring_lock);
>>   
>>   	xa_for_each(&ctx->zcrx_ctxs, index, ifq) {
>> +		if (refcount_read(&ifq->refs) > 1)
>> +			continue;
> 
> This is a bit odd, it's not an idiomatic way to use reference counts.
> Why isn't this a refcount_dec_and_test()? Given that both the later grab
> when sharing is enabled and the shutdown here are under the ->uring_lock
> this may not matter, but it'd be a lot more obviously correct if it
> looked ala:
> 
> 		if (refcount_dec_and_test(&ifq->refs)) {
>    			io_zcrx_scrub(ifq);
>    			io_close_queue(ifq);
> 		}
> 
> instead?
> 

Yeah, good idea. Your comments prompted me to try to find a better
solution that gets rid of ifq->proxy. Turns out xarray has 3 bits per
entry that can be 'marked'. With this I can get a cleaner solution. Will
respin tomorrow.