linux-kernel - Re: [PATCH] ALSA: hda - allow 40 bit DMA mask for NVidia devices

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: <s5hwph54uju.wl-tiwai@suse.de>
Date:   Tue, 18 Oct 2016 16:10:29 +0200
From:   Takashi Iwai <tiwai@...e.de>
To:     Mike Travis <travis@....com>
Cc:     Ard Biesheuvel <ard.biesheuvel@...aro.org>,
        <alsa-devel@...a-project.org>, <perex@...ex.cz>,
        <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] ALSA: hda - allow 40 bit DMA mask for NVidia devices

On Tue, 18 Oct 2016 14:46:29 +0200,
Mike Travis wrote:
> 
> 
> Have you tested this change on a system which has physical
> addressing > 40 bits to insure that addresses > 40 bits are
> not simply truncated to an incorrect (mod 40) address?
> By not restricting the addresses to 32 bits you are not
> restricting the device to the DMA32 zone.
> 
> This is what caused the original panic on UV systems.

But shouldn't it be handled properly by dma_set_mask() and
dma_set_coherent_mask()?  


Takashi

> Btw, the newest nVidia GPUs have > 40 bits of DMA address bits.
> 
> On 10/18/2016 3:25 AM, Takashi Iwai wrote:
> > On Mon, 17 Oct 2016 18:23:59 +0200,
> > Ard Biesheuvel wrote:
> >>
> >> Commit 49d9e77e72cf ("ALSA: hda - Fix system panic when DMA > 40 bits
> >> for Nvidia audio controllers") simply disabled any DMA exceeding 32
> >> bits for NVidia devices, even though they are capable of performing
> >> DMA up to 40 bits. On some architectures (such as arm64), system memory
> >> is not guaranteed to be 32-bit addressable by PCI devices, and so this
> >> change prevents NVidia devices from working on platforms such as AMD
> >> Seattle.
> >>
> >> Since the original commit already mentioned that up to 40 bits of DMA
> >> is supported, and given that the code has been updated in the meantime
> >> to support a 40 bit DMA mask on other devices, revert commit 49d9e77e72cf
> >> and explicitly set the DMA mask to 40 bits for NVidia devices.
> >>
> >> Signed-off-by: Ard Biesheuvel <ard.biesheuvel@...aro.org>
> > 
> > Thanks, applied.  Since this seems causing a real problem, I added Cc
> > to stable.
> > 
> > 
> > Takashi
> > 
> > 
> >> ---
> >>  sound/pci/hda/hda_intel.c | 7 +++++--
> >>  1 file changed, 5 insertions(+), 2 deletions(-)
> >>
> >> diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c
> >> index c3469f756ec2..c64d986009a9 100644
> >> --- a/sound/pci/hda/hda_intel.c
> >> +++ b/sound/pci/hda/hda_intel.c
> >> @@ -341,8 +341,7 @@ enum {
> >>  
> >>  /* quirks for Nvidia */
> >>  #define AZX_DCAPS_PRESET_NVIDIA \
> >> -	(AZX_DCAPS_NO_MSI | /*AZX_DCAPS_ALIGN_BUFSIZE |*/ \
> >> -	 AZX_DCAPS_NO_64BIT | AZX_DCAPS_CORBRP_SELF_CLEAR |\
> >> +	(AZX_DCAPS_NO_MSI | AZX_DCAPS_CORBRP_SELF_CLEAR |\
> >>  	 AZX_DCAPS_SNOOP_TYPE(NVIDIA))
> >>  
> >>  #define AZX_DCAPS_PRESET_CTHDA \
> >> @@ -1716,6 +1715,10 @@ static int azx_first_init(struct azx *chip)
> >>  		}
> >>  	}
> >>  
> >> +	/* NVidia hardware normally only supports up to 40 bits of DMA */
> >> +	if (chip->pci->vendor == PCI_VENDOR_ID_NVIDIA)
> >> +		dma_bits = 40;
> >> +
> >>  	/* disable 64bit DMA address on some devices */
> >>  	if (chip->driver_caps & AZX_DCAPS_NO_64BIT) {
> >>  		dev_dbg(card->dev, "Disabling 64bit DMA\n");
> >> -- 
> >> 2.7.4
> >>
> >>
>