On Sun, Feb 13, 2011 at 07:22, Dave Young <hidave.darkstar@xxxxxxxxx> wrote:
> Finally I bisected it, results:
> 47970b1b2aa64464bc0a9543e86361a622ae7c03 is first bad commit
> commit 47970b1b2aa64464bc0a9543e86361a622ae7c03
> Author: Chris Wright <chrisw@xxxxxxxxxxxx>
> Date: Â Thu Feb 10 15:58:56 2011 -0800
>
> Â Âpci: use security_capable() when checking capablities during config space read
>
> Â ÂEric Paris noted that commit de139a3 ("pci: check caps from sysfs file
> Â Âopen to read device dependent config space") caused the capability check
> Â Âto bypass security modules and potentially auditing. ÂRectify this by
> Â Âcalling security_capable() when checking the open file's capabilities
> Â Âfor config space reads.
>
> Â ÂReported-by: Eric Paris <eparis@xxxxxxxxxx>
> Â ÂSigned-off-by: Chris Wright <chrisw@xxxxxxxxxxxx>
> Â ÂSigned-off-by: James Morris <jmorris@xxxxxxxxx>
>
Actually, even reading the PCI capabilities fails with lspci
reporting "Capabilities: <access denied>" if run as root.
"libpciaccess" should have handled this situation, but still
it looks like a regression and it breaks existing systems.
_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/dri-devel
