netdev - Re: please revert [PATCH bpf-next v5 03/15] xsk: create and free buffer pool independently from umem

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: <88fbfe56-8847-25a5-9301-a573788ca91a@intel.com>
Date:   Mon, 5 Oct 2020 10:47:08 +0200
From:   Björn Töpel <bjorn.topel@...el.com>
To:     Christoph Hellwig <hch@...radead.org>, davem@...emloft.net,
        Magnus Karlsson <magnus.karlsson@...el.com>
Cc:     ast@...nel.org, daniel@...earbox.net, netdev@...r.kernel.org,
        jonathan.lemon@...il.com, maximmi@...lanox.com,
        bpf@...r.kernel.org, jeffrey.t.kirsher@...el.com,
        anthony.l.nguyen@...el.com, maciej.fijalkowski@...el.com,
        maciejromanfijalkowski@...il.com, cristian.dumitrescu@...el.com
Subject: Re: please revert [PATCH bpf-next v5 03/15] xsk: create and free
 buffer pool independently from umem

On 2020-10-05 10:43, Christoph Hellwig wrote:
> On Mon, Oct 05, 2020 at 09:35:35AM +0100, Christoph Hellwig wrote:
>> Hi Dave,
>>
>> please can you rever this?  This is the second xsk patch this year
>> that pokes into dma-mapping internals for absolutely not reason.
>>
>> And we discussed this in detail the last time around: drivers have
>> absolutely no business poking into dma-direct.h and dma-noncoherent.h.
>> In fact because people do this crap I have a patch to remove these
>> headers that I'm about to queue up, so if this doesn't get reverted
>> the linux-next build will break soon.
> 
> Looks like it doesn't actually use any functionality and just
> pointlessly includes internal headers.  So just removing dma-direct.h,
> dma-noncoherent.h and swiotlb.h should do the job as well.
> 
> But guys, don't do this.
> 

Yeah, this commit just moves the work you helped out with [1], and by 
accident/sloppiness the internal dma headers were included in the move.

Sorry about that, and we'll try not to repeat this. Apologies for the 
extra work on your side.

I'll spin up a patch to fix this asap.


Thanks,
Björn


[1] https://lore.kernel.org/bpf/20200629130359.2690853-5-hch@lst.de/