Re: [Bug 13286] dma_timer_expiry: DMA status (0x21)

[Bartlomiej Zolnierkiewicz <bzolnier@gmail.com>]

On Wednesday 13 May 2009 20:01:31 bugzilla-daemon@bugzilla.kernel.org wrote:
> http://bugzilla.kernel.org/show_bug.cgi?id=13286
> 
> 
> 
> 
> 
> --- Comment #4 from fibre_optic@go2.pl  2009-05-13 18:01:31 ---
> "Does 2.6.27.21 use DMA?"
> 
> Here is part of dmesg output from 2.6.27.21:
> ...
> platform rtc_cmos: registered platform RTC device (no PNP device found)
> squashfs: version 3.4 (2008/08/26) Phillip Lougher
> msgmni has been set to 498
> io scheduler noop registered (default)
> pci 0000:00:01.0: Boot video device
> brd: module loaded
> Uniform Multi-Platform E-IDE driver
> ide_generic: please use "probe_mask=0x3f" module parameter for probing all
> legacy ISA IDE ports
> Probing IDE interface ide0...
> hda: SanDisk SDCFJ-32, CFA DISK drive

This kernel is using ide_generic host driver which doesn't support DMA
instead of piix one (it explains why 2.6.27.21 works and 2.6.29.3 doesn't).

> "Unfortunately there is no software method to check it so
> you need to look at PCB and verify if there are any routes
> connected to pins 21 and 29 (DMARQ and -DMACK lines)."
> 
> Take a look at attached pictures. Do they help to figure this out?

No DMA lines... this is hardware problem -- you can workaround it with
"ide_core.nodma=0.0" boot parameter (see Documentation/ide/ide.txt for
more details on ide parameters).

I'm closing the bug.