repeated kernel crash, posibly after last update (dma_timer_expiry)
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I have been experiencing over the last three weeks several machine crashes, more or less since I installed the last YOU update, kernel-default-2.6.11.4-21.9.i586.rpm (SuSE 9.3). Yesterday evening I had three crashes. Around 10-20 minutes after actively using the computer, it suddenly became very sluggish, the mouse jumped, and finally froze completely, within 15" of the first symptoms: not even num-lock key/led worked. After rebooting, nothing showed in any log. The last time I was fast enough to jump to console-10, and I saw this line appearing (hand copied): kernel: hdb dma_timer_expiry: dma status= 0x64 I could still switch consoles, but that was all I could do, it did not respond to commands. Memtest ran fine; so did "smart" long test. No apparent hardware error. Thinking that it could be the last kernel upgrade, I have reverted to 2.6.11.4-21.8, and so far, it is working... "unfortunately" at the same time I have reseated the drive connector, so I'm not absolutely sure which maneuver did the trick - if it did, as it is still to early to know. But knowing that some of the SuSE kernel developers read this list, perhaps they know if there was some thing modified in the kernel that could have affected, seeing the error message above. - -- Cheers, Carlos Robinson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFDPSogtTMYHG2NR9URAmO3AJ4+qXkTpRlgrlGKURRNa/6zwBgsOQCbBfC6 X4VdVA67AqUvCsQE7RbJt5U= =FhRd -----END PGP SIGNATURE-----
Carlos E. R. said:
Yesterday evening I had three crashes. Around 10-20 minutes after actively using the computer, it suddenly became very sluggish, the mouse jumped, and finally froze completely, within 15" of the first symptoms: not even num-lock key/led worked. After rebooting, nothing showed in any log.
The last time I was fast enough to jump to console-10, and I saw this line appearing (hand copied):
kernel: hdb dma_timer_expiry: dma status= 0x64 [...] Thinking that it could be the last kernel upgrade, I have reverted to 2.6.11.4-21.8, and so far, it is working...
I don't think this problem affects only the latest Suse kernel. It seems to be a bug common to all "new" IDE driver versions. I have the same behaviour on my home PC under Debian Sarge (kernel 2.6.8...) and Suse 9.0 (kernel 2.4.18...), but only if I switch the harddisks to MWDMA. With UltraDMA, the error frequency is higher, but after some time the driver disables DMA completly and resets the IDE bus thus avoiding the crash. Until now, I had no problems with disabled DMA (but that is not an option for a multimedia PC with video editing/playback). Some more observations: I have 3 IDE harddisks, but only the newer ones (Samsung 120GB, ATA7 and Seagate 160GB, ATA6) are effected by these DMA timeeout errors. The older drive (IBM 30GB, ATA5) runs without problem. Changing all IDE cables made no differences. I tried nearly all possible drive connections (IDE bus 0/1, master/slave), but the errors always effect the same drives (with no noticeable change in the error rate). -- Michel Messerschmidt, lists@michel-messerschmidt.de
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Michel Messerschmidt wrote:
Carlos E. R. said:
Yesterday evening I had three crashes. Around 10-20 minutes after actively using the computer, it suddenly became very sluggish, the mouse jumped, and finally froze completely, within 15" of the first symptoms: not even num-lock key/led worked. After rebooting, nothing showed in any log.
The last time I was fast enough to jump to console-10, and I saw this line appearing (hand copied):
kernel: hdb dma_timer_expiry: dma status= 0x64 [...] Thinking that it could be the last kernel upgrade, I have reverted to 2.6.11.4-21.8, and so far, it is working...
I don't think this problem affects only the latest Suse kernel. It seems to be a bug common to all "new" IDE driver versions.
I have the same behaviour on my home PC under Debian Sarge (kernel 2.6.8...) and Suse 9.0 (kernel 2.4.18...), but only if I switch the harddisks to MWDMA. With UltraDMA, the error frequency is higher, but after some time the driver disables DMA completly and resets the IDE bus thus avoiding the crash. Until now, I had no problems with disabled DMA (but that is not an option for a multimedia PC with video editing/playback).
Some more observations: I have 3 IDE harddisks, but only the newer ones (Samsung 120GB, ATA7 and Seagate 160GB, ATA6) are effected by these DMA timeeout errors. The older drive (IBM 30GB, ATA5) runs without problem. Changing all IDE cables made no differences. I tried nearly all possible drive connections (IDE bus 0/1, master/slave), but the errors always effect the same drives (with no noticeable change in the error rate).
The same here: SUSE kernel 2.6.13-9-default, disk IBM IC35, controller VIA vt8233a Jürgen -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.2 (GNU/Linux) Comment: Using GnuPG with Thunderbird - http://enigmail.mozdev.org iD8DBQFDQ3ettMrl3JEeRvwRAqT5AJ4//agiphAj3+XvwN3pdGbTUtEFbgCfZQp2 EvPv1LqMOoLGDt2IZe2sMe0= =IfFS -----END PGP SIGNATURE-----
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 The Wednesday 2005-10-05 at 08:50 +0200, Jürgen Mell wrote:
The same here: SUSE kernel 2.6.13-9-default, disk IBM IC35, controller VIA vt8233a
Mine are two Seagate disks on the IDE0 cable: Model=ST360020A, FwRev=3.34, 60.0 GB Model=ST340810A, FwRev=3.34, 40.0 GB The one I catched the error message was the second, where most of Linux resides; but if I had waited for more crashes, I could perhaps see more messages. I have a new 160 drive, ultra ATA, not yet formatted. I hope not to have too many problems with it. My chipset is (lspci): 0000:00:00.0 Host bridge: Intel Corporation 82845 845 (Brookdale) Chipset Host Bridge (rev 03) 0000:00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev 12) 0000:00:1f.1 IDE interface: Intel Corporation 82801BA IDE U100 (rev 12) so it is not only the VIA chipset that gives problems, nor big disks. But with the kernel: Linux version 2.6.11.4-21.8-default (geeko@buildhost) (gcc version 3.3.5 20050117 (prerelease) (SUSE Linux)) #1 Tue Jul 19 12:42:37 UTC 2005 I have not seen more crashes. I intend thus not to update the kernel till I hear comments from the SuSE people - I hope they are reading this and taking notes, because as my connection is not permanent, I don't like to use feedback web form, I prefer email. But it is then obvious that it is a generalized problem. - -- Cheers, Carlos Robinson -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.0 (GNU/Linux) Comment: Made with pgp4pine 1.76 iD8DBQFDRBv1tTMYHG2NR9URAl6qAJsGAWv3uVfWmJXNImKfLl6WuqKOhwCeKdBf Xh3fJjGPmTG8psU0isFvMg8= =osHg -----END PGP SIGNATURE-----
participants (3)
-
Carlos E. R. -
Jürgen Mell -
Michel Messerschmidt