Skip to contentSkip to conversation

IDE corruption during heavy bt878-induced interrupt load [LKM]

DISCUSSION joe briggs 1 participant 1 message Apr 22, 2003 1 message not in the archive

lore
Share & tools
lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

joe briggs

author
#1 lore raw

I forgot to mention 2.4.19 on Debian Woody, GigaByte GA-7VAXP/Athlon MX2500/512 MB (and several other Intel motherboards).

I am glad that you mentioned ext3 because while I curse ReiserFS, I really don't think that it is part of the problem. Definately PCI-dma related, but does onboard IDE (i.e., my system disk) use DMA in the same way that a PCI adapter such as Promise does?

Quoted 31 lines · you
Quoted 12 lines

I create multi-channel digital surveillance systems using cards with 4 or more multiplexed bt878 framegrabbers; each one capturing 5 or more frames per second on each of its two input channels (total 4 * 2 * 5 = 40 fps). Typically I run using either a Promise or Adaptec HPT370 IDE-RAID controller with 2 WD-120GB/8MB-cache drives striped in RAID-0, with another IDE as hda for my system drive. What happens is that every few seconds I get a "BTTV: RISC ERROR - resetting" from the frame grabber driver. After a few days of this I have corruption on my Reiser file system; which usually I am able to clean up with mkreiserfs --fix-fixable or --rebuild-tree. The corruption is both on my RAID and my system drive. Missing doing this maintenance action can really ruin my day.

Quoted 3 lines

All comments/suggestions/etc. appreciated.

Joe

Hi,

I'm seeing the same type of corruption (kernel 2.4.21pre7-ac1+acpi) when mencoding from BT878 to ext3 on ide raid1 (VIA 8233A on-board ide) with single 25 FPS. The kernel even oopsed on umount of the array, showing signs of memory corruption.
I haven't come far with investigation, since I cannot affort to ruin that (production) system, however it seems clear to me that the v4l driver is the culprit. We also thought of dma contention, so my next test will be to switch the drives to PIO, to see if the dma traffic reduction helps.

If you find some solution, please let me know

Best Regards
Andy

Signature · 6 lines
Joe Briggs
Briggs Media Systems
105 Burnsen Ave.
Manchester NH 01304 USA
TEL/FAX 603-232-3115 MOBILE 603-493-2386
www.briggsmedia.com