linux-kernel - Re: [PATCH 08/16] dma-debug: add core checking functions

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 for Android: free password hash cracker in your pocket

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

Date:	Fri, 13 Feb 2009 11:48:02 -0800
From:	Jesse Brandeburg <jesse.brandeburg@...il.com>
To:	Joerg Roedel <joro@...tes.org>
Cc:	David Woodhouse <dwmw2@...radead.org>,
	Joerg Roedel <joerg.roedel@....com>,
	fujita.tomonori@....ntt.co.jp, iommu@...ts.linux-foundation.org,
	mingo@...hat.com, linux-kernel@...r.kernel.org,
	davem@...emloft.net, Jesse Brandeburg <jesse.brandeburg@...el.com>
Subject: Re: [PATCH 08/16] dma-debug: add core checking functions

On Thu, Feb 12, 2009 at 6:49 AM, Joerg Roedel <joro@...tes.org> wrote:
> On Thu, Feb 05, 2009 at 09:07:59PM +0000, David Woodhouse wrote:
>> On Mon, 2009-01-19 at 15:46 +0100, Joerg Roedel wrote:
>> >
>> > +       if (ref->type != entry->type) {
>> > +               err_printk(ref->dev, "DMA-API: device driver frees "
>> > +                          "DMA memory with wrong function "
>> > +                          "[device address=0x%016llx] [size=%llu bytes] "
>> > +                          "[mapped as %s] [unmapped as %s]\n",
>> > +                          ref->dev_addr, ref->size,
>> > +                          type2name[entry->type], type2name[ref->type]);
>>
>> e1000e 0000:00:19.0: DMA-API: device driver frees DMA memory with wrong
>> function [device address=0x00000000ffedc002] [size=90 bytes] [mapped as
>> single] [unmapped as page]
>>
>>
>> <davem> That debug check is excessive, substituting those two things
>> works for every DMA API implementation I am aware of
>
> True. Buts it is an API violation, isn't it?

I have a fix for this, for both e1000 and e1000e, the patches are not
(deeply) tested and are RFC right now.
Both e1000e patches must be applied for the actual bug fix to apply cleanly
01 - e1000e_remove_old_workarounds
02 - e1000e_fix_dma_unmap

Jesse

View attachment "e1000e_remove_old_workarounds.patch" of type "text/x-patch" (1500 bytes)

View attachment "e1000_fix_dma_unmap.patch" of type "text/x-patch" (3288 bytes)

View attachment "e1000e_fix_dma_unmap.patch" of type "text/x-patch" (3383 bytes)