[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <396f9358-1e73-5448-96d1-98e36123093f@kernel.dk>
Date: Fri, 12 Nov 2021 11:07:52 -0700
From: Jens Axboe <axboe@...nel.dk>
To: Christoph Hellwig <hch@....de>, Pavel Machek <pavel@....cz>
Cc: Zubin Mithra <zsm@...omium.org>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
linux-kernel@...r.kernel.org, stable@...r.kernel.org,
Omar Sandoval <osandov@...com>, Ming Lei <ming.lei@...hat.com>
Subject: Re: [PATCH 4.19 01/16] block: introduce multi-page bvec helpers
On 11/11/21 10:48 PM, Christoph Hellwig wrote:
> On Thu, Nov 11, 2021 at 07:53:08PM +0100, Pavel Machek wrote:
>>> There is some more context on this at:
>>> https://lore.kernel.org/linux-block/YXweJ00CVsDLCI7b@google.com/T/#u
>>> and
>>> https://lore.kernel.org/stable/YYVZBuDaWBKT3vOS@google.com/T/#u
>>
>> Thank you!
>
> Honestly, this looks broken to me. multipage-bvec was a big invasive
> series with a lot of latter fixups. While taking this patch on it's
> own should be save by itself, but also useless. So if it is needed
> to make a KASAN warning go away we need to dig deeper and back something
> else out that should not have been backported to rely on multipage
> bvec becasue without the other patches they don't exist and can't
> acually work.
Yeah it makes no sense to me, to be honest. We're most certainly not
backporting multipage bvecs to stable, and if this particular patch
fixes something, it's by happy accident and we should find out why.
--
Jens Axboe
Powered by blists - more mailing lists