linux-kernel - Re: [PATCH v1] ALSA: memalloc: Fix indefinite hang in non-iommu case

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: <87le7gog4k.wl-tiwai@suse.de>
Date: Mon, 19 Feb 2024 12:19:55 +0100
From: Takashi Iwai <tiwai@...e.de>
To: Sven van Ashbrook <svenva@...omium.org>
Cc: Takashi Iwai <tiwai@...e.de>,
	Kai Vehmanen <kai.vehmanen@...ux.intel.com>,
	Hillf Danton <hdanton@...a.com>,
	Karthikeyan Ramasubramanian <kramasub@...omium.org>,
	LKML <linux-kernel@...r.kernel.org>,
	Brian Geffon <bgeffon@...gle.com>,
	linux-sound@...r.kernel.org,
	"Arava, Jairaj" <jairaj.arava@...el.com>,
	Curtis Malainey <cujomalainey@...omium.org>
Subject: Re: [PATCH v1] ALSA: memalloc: Fix indefinite hang in non-iommu case

On Fri, 16 Feb 2024 17:22:45 +0100,
Sven van Ashbrook wrote:
> 
> On Fri, Feb 16, 2024 at 9:43 AM Takashi Iwai <tiwai@...e.de> wrote:
> >
> > OK, then how about the one like below?
> >
> > This changes:
> > - Back to __GFP_NORETRY as default
> > - Use __GFP_RETRY_MAYFAIL for SNDRV_DMA_TYPE_NONCONTIG with IOMMU;
> >   this should cover the commit a61c7d88d38c
> > - Also use __GFP_RETRY_MAYFAIL for the SG-fallback allocations of the
> >   minimal order, just like IOMMU allocator does.
> >
> > This should be less destructive, while still allowing more aggressive
> > allocations for SG buffers.
> 
> This one looks like it would keep the SOF firmware allocation issue at bay,
> in both iommu and non-iommu cases.
> 
> If there is no further discussion in this thread, we'll stress test this
> on iommu and non-iommu Chromebooks.

The test with my latest patch would be appreciated in anyway.


thanks,

Takashi