Bug 36017 - large diskactivity -> kernel freezes
Summary: large diskactivity -> kernel freezes
Alias: None
Product: Red Hat Raw Hide
Classification: Retired
Component: kernel   
(Show other bugs)
Version: 1.0
Hardware: alpha
OS: Linux
Target Milestone: ---
Assignee: Phil Copeland
QA Contact: Brock Organ
Depends On:
TreeView+ depends on / blocked
Reported: 2001-04-16 10:30 UTC by Dries Verachtert
Modified: 2007-04-18 16:32 UTC (History)
0 users

Fixed In Version:
Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of:
Last Closed: 2001-05-29 20:25:04 UTC
Type: ---
Regression: ---
Mount Type: ---
Documentation: ---
Verified Versions:
Category: ---
oVirt Team: ---
RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: ---

Attachments (Terms of Use)

Description Dries Verachtert 2001-04-16 10:30:22 UTC
Used system: Personal Workstation 500a MiataGL
  scsi card: QLogic ISP1020 SCSI on PCI bus 01 device 20 irq 27 I/O base
  2 harddisks of type RZ2CC-KB (C) DEC Rev: DC2B

When I'm copying a lot of files, the system sometimes hangs. I don't get an
oops or anything. Only thing i still can do is a hard reset.

It only crashes with linux 2.4.x (i'm now using kernel-2.4.2-0.1.49 from
rawhide, whole system is rawhide) The 2.2.x kernels don't have the problem.

Easy reproducable:
dd if=/dev/sda of=/dev/null &
dd if=/dev/sdb of=/dev/null &
and wait 5 minutes

Comment 1 Dries Verachtert 2001-05-24 21:25:13 UTC
the latest patch of Andrea Arcangeli which can be found at 
seems to fix this problem. 

Wesley Daemen is the one who told me this good news. He already applied the 
patch and has started those dd commands to create a lot of harddisk activity 
and his miata doesn't crash.

Comment 2 Dries Verachtert 2001-05-25 13:38:07 UTC
2.4.4-ac17 also seems to have the fix. The fix is contained in the patch of 
Andrea but the original author seems to be  Ivan Kokshaysky.

Comment 3 Dries Verachtert 2001-05-27 15:29:05 UTC
2.4.5 also fixes the problem.

Comment 4 Phil Copeland 2001-05-29 21:15:32 UTC
I believe this is the tbia patch in pci_iommu.c

it's a bit dirty in that as far as I remember the bits to differnetiate between
tbia of 1 and 8 depending on what chipset  you're running (tsunami/clipper are 8
and I think all the rest are 1. I could very well be wrong)

I'm hacking in changes to the 2.4.3 kernel that will be released for 7.1 alpha


Comment 5 Dries Verachtert 2001-05-29 22:03:48 UTC
Let me know where i can find the rpms if you want me to test a patched version :-)

Comment 6 Phil Copeland 2001-05-29 22:40:00 UTC
Hopefully rawhide should have kernel-2.4.3-7
That *should* work but probably not for tsunami/clipper motherboards


Note You need to log in before you can comment on or make changes to this bug.