2.4.4 kernel freeze for unknown reason
Hi everyone,
Mark, I got your point about the dma/udma stuffs. My hdparm setting is UDMA w/ MultiSector 16..
I had recompiled my kernel and disabled the FB option but my linux box still hanged (another completely freeze) yesterday... Oh well..
I have been tracking this thread for a few days and it seem the source of this problem is related to swap space. Vincent, would you mind to send me the patch for swap space problem if Alan had sent it to you? So I can test it on my machine and report the result later.
Mark, please suggest a setting for the hdparm so I can test it on my machine. Thanks alot for your time.
Best Regards,
Jacky Liu
Quoted 10 lines
Date: Thu, 10 May 2001 23:20:17 -0400 (EDT)
From: Mark Hahn <[email protected]>
To: Jacky Liu <[email protected]>
SUBJECTQuoted 1 line
You are right for the other assumption, I am running the harddisk in UDMA w/32bits mode. Are you suggesting me to turn off both functions? But if I turn them off, the performance will decrease alot.
the only thing that matters is dma/udma or not. 32b mode is irrelevant to the actual dma/udma transfer, as is -m settings. even -u has no measurable effect.
Quoted 5 lines
Quoted 1 line
Is there any way I can get any crash information? e.g. any function I can turn on for logging or something?
if your hang is as I'm imagining, there's nothing you can do, since it's purely hardware. you could try compiling in magic sysrq, but I suspect the hardware is hung.
------------------------------------------------------------
--== Sent via Deja.com ==--
http://www.deja.com/
Quoted 5 lines · Jacky Liu
Hi everyone,
Mark, I got your point about the dma/udma stuffs. My hdparm setting is UDMA w/ MultiSector 16..
Quoted 8 lines
I had recompiled my kernel and disabled the FB option but my linux box still hanged (another completely freeze) yesterday... Oh well..
I have been tracking this thread for a few days and it seem the source of this problem is related to swap space. Vincent, would you mind to send me the patch for swap space problem if Alan had sent it to you? So I can test it on my machine and report the result later.
I havn't received the patch but, on Byron Stanoszek's suggestion, I have compiled 2.4.4 with gcc-2.95.3 to run on it a few days to see if it is any better. So far, it appears at least as stable as with egcs-1.1.2 and most of the swap was freed up this morning for the first time since 2.4.0. However, it is pretty full tonight. I should know tomorrow if it really made a difference. This is usually the state in which I find it locked up the next morning. It has been up a little over 2 days now. Byron actually suggested I use the ac7 patch but I first wanted to see if the compiler had anything to do with it without changing anything else.
Quoted 6 lines
Mark, please suggest a setting for the hdparm so I can test it on my machine. Thanks alot for your time.
Best Regards,
Jacky Liu