On Sun, 2010-10-31 at 01:40 +0200, Daniel GlÃckner wrote: > On Mon, Oct 25, 2010 at 11:34:40AM -0400, Laurent Birtz wrote: > > I've observed that the poison was corrupted at some random pages, but > > always in the first four bytes. The value of those bytes is 0x23232323. > > This byte sequence often appears in kernel oops reports on my setup. > I couldn't verify your findings by mallocing an equivalent of the > physical memory, filling everything with 0xdeadbeef, and then > periodically checking it didn't change while watching tv at 768x576 > in YUY2. > Here are other apparent instances with bttv: http://linux.derkeiler.com/Mailing-Lists/Kernel/2004-08/2233.html http://www.mail-archive.com/linux-kernel@xxxxxxxxxxxxxxx/msg247084.html http://kerneltrap.org/mailarchive/linux-kernel/2008/11/19/4214764 http://lkml.org/lkml/2006/6/2/16 http://lkml.indiana.edu/hypermail/linux/kernel/0407.0/1124.html http://fixunix.com/kernel/323946-bug-unable-handle-kernel-paging-request-virtual-address.html BTW, I did notice that drivers/acpi/acpica/utglobal.c:acpi_ut_get_node_name() has a return("####"); statement in it explicitly. 0x23232323 is "####" in ASCII. It could possibly be the case that anything that calls that function could be a source of those "####" bytes. Although the initial patch submission comments make that seem unlikely. Regards, Andy > Which video mode and resolution did you use in your tests? > Did you try loading bttv with triton1=1 and vsfx=1? > > Daniel > -- > To unsubscribe from this list: send the line "unsubscribe linux-media" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html -- To unsubscribe from this list: send the line "unsubscribe linux-media" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
