Re: [SLE] dma_timer_expiry weirdness with SUSE LINUX 9.2.
Hmm.... This is very weird. Just tried the disk in my other computer,
besides testing it on 3 of 4 IDE controllers in the original machine.
Same thing happens. Ergo: Not cable or motherboard problem. When did
they start with these half-height sliding pins and what's the deal
with them anyway? Is my problem even related to that stupid pin? It's
the only visual difference I have to my other IDE disks.
There are issues with both 40 IDE and 80 IDE cables. There are no SATA
connectors on the disk (and I have not SATA controllers), so....
Basically, the disk *can* function, but only without DMA enabled,
meaning 5-6 MB/s. Not very fun. I suppose I have to have a word with
the shop that sold it to me...
/Martin.
PS: I have run tests of the two disks with kernel 2.4.25. Although
they *say* they have DMA enabled they either cannot be timed or end up
with the above 5-6 MB/s for read tests anyway.
As Madonna once said: I'm going bananas.
PPS: As for Xine, you can get that from packman.links2linux.org if you
wish. Should be fine and uncrippled.
On Fri, 05 Nov 2004 21:00:22 -0600, Donn Washburn
Martin Moeller wrote:
On Fri, 05 Nov 2004 19:34:41 -0600, Donn Washburn
wrote: ? are you using the 80 IDE cable? Try hdparm --help and see if it will tell you anything. Also what file system are you using. I have heard reiser may give you problems. But after about 5 years of reiser as the main file sytem type - no problem.
also check " /usr/sbin/smartctl "
Unfortunately the problem show up even before the disk is partitioned so filesystems have little to do with it. Yes, I'm using the 80 cable and most of my partitions are reiserfs, but the problem manifests itself way before we get to filesystem level.
The 'funny' thing is that linux 2.4.x doesn't have this issue, although I have not yet run tests in that environment, I think.
Maybe of no consequence, but one of the connectors on the disk is 'half-height' and can be moved back and both about half a millimetre. No clue as to why....
/Martin.
Well you are correct that the OS has nothing to do with it. If it is a WD drive get a DOS based copy of thier diagnostic software.
As for the short pins - that is a real problem. Good luck
I also tonight got SuSE 92 and kernel 2.6.9 working. Best of all Nvidia came out with a new file today and it is working also.
Compiling the kernel with the correct parts all came about because I am trying to ger xine to compile correctly and it refuses with the convoluted kernel setup SuSE 9.2 has.
-- 73 de Donn Washburn __" http://www.hal-pc.org/~n5xwb " Ham Callsign N5XWB / / __ __ __ __ __ __ __ 307 Savoy St. / /__ / / / \/ / / /_/ / \ \/ / Sugar Land, TX 77478 /_____/ /_/ /_/\__/ /_____/ /_/\_\ LL# 1.281.242.3256 Dump Microsoft Software - Stop virus email Email: n5xwb@hal-pc.org " http://counter.li.org " #279316
Well, it was my own fault, actually.
Somehow I have put one of the connector cables on the wrong way
around, thus pressing that one pin down. A colleague of mine managed
to get it most of the way up again and now the drive works as it's
supposed to. Reads are now up to 57 MB/s. MUCH better.
Now to see if I can salvage my older data from the other disk....
/Martin.
On Sun, 7 Nov 2004 00:33:06 +0100, Martin Moeller
Hmm.... This is very weird. Just tried the disk in my other computer, besides testing it on 3 of 4 IDE controllers in the original machine. Same thing happens. Ergo: Not cable or motherboard problem. When did they start with these half-height sliding pins and what's the deal with them anyway? Is my problem even related to that stupid pin? It's the only visual difference I have to my other IDE disks.
There are issues with both 40 IDE and 80 IDE cables. There are no SATA connectors on the disk (and I have not SATA controllers), so....
Basically, the disk *can* function, but only without DMA enabled, meaning 5-6 MB/s. Not very fun. I suppose I have to have a word with the shop that sold it to me...
/Martin.
PS: I have run tests of the two disks with kernel 2.4.25. Although they *say* they have DMA enabled they either cannot be timed or end up with the above 5-6 MB/s for read tests anyway.
As Madonna once said: I'm going bananas.
PPS: As for Xine, you can get that from packman.links2linux.org if you wish. Should be fine and uncrippled.
participants (1)
-
Martin Moeller