linux-kernel - Re: mempool_alloc_bulk and various mempool improvements

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

Message-ID: <454da918-bb53-410e-8046-49ffded8e5df@suse.cz>
Date: Wed, 12 Nov 2025 16:57:43 +0100
From: Vlastimil Babka <vbabka@...e.cz>
To: Christoph Hellwig <hch@....de>
Cc: Andrew Morton <akpm@...ux-foundation.org>,
 Christoph Lameter <cl@...two.org>, David Rientjes <rientjes@...gle.com>,
 Roman Gushchin <roman.gushchin@...ux.dev>, Harry Yoo <harry.yoo@...cle.com>,
 Eric Biggers <ebiggers@...nel.org>, linux-mm@...ck.org,
 linux-kernel@...r.kernel.org
Subject: Re: mempool_alloc_bulk and various mempool improvements

On 11/12/25 16:50, Christoph Hellwig wrote:
> On Wed, Nov 12, 2025 at 01:22:01PM +0100, Vlastimil Babka wrote:
>> On 11/11/25 14:52, Christoph Hellwig wrote:
>> > Hi all,
>> > 
>> > this series adds a bulk version of mempool_alloc that makes allocating
>> > multiple objects deadlock safe.
>> > 
>> > The initial users is the blk-crypto-fallback code:
>> > 
>> >   https://lore.kernel.org/linux-block/20251031093517.1603379-1-hch@lst.de/
>> > 
>> > with which v1 was posted, but I also have a few other users in mind.
>> 
>> How do we handle this then once it's settled?
> 
> Good question.  I think there is enough mm material now for a merge
> through an mm tree.  If we get it sorted out for this merge window just
> merging it through mm even when there is no user yet would be easiest as

OK should work hopefully...

> the blk-crypto work is probably not going to make this merge window with
> additional review from Eric and I'll most like have another user for the
> next merge window as well (block bounce buffering for !STABLE_WRITES
> dio).  Otherwise we'll have to operate with shared branches.

... and we can avoid the need for shared branches that way.