Roberto Sanchez wrote: > > ... > > BTW, what kernel are you compiling (sorry, late to the thread)
Debian's 2.4.22-2, with Debian's 2.4.22-2 k7-smp config. file. > and what IDE chipset do you have? AMD 768 (amd7441) (dual-Athlon motherboard). Of course, thanks to the still-extant, long-time IDE DMA corruption bug in the Linux kernel, and thanks to whomever (Debian config.?) irresponsibly turned DMA on by default in the kernel (before my hdparm init. script can turn it back off), my regular root partition is COMPLETELY SCREWED! (Luckily I had an old root partition to boot from.) (I had a hang. I rebooted. The system started recovering the ext3 root filesystem from the journal. Because the default configuration now enables DMA, and because recovery runs before my rc.boot or init.d script that turns DMA off, the recovery ran with DMA on, which of course means that the "recovery" scrambled my filesystem way beyond repair.) Does anyone know: - Why the IDE DMA corruption bug has been around unfixed for so long? - Why the kernel doesn't retry the timed-out operation? - Is the bug believed to be fixed (and in which kernel version)? - Why on earth Debian's default configuration would turn DMA on? (Why wouldn't the default be to leave it off since it's so dangerous?) Thanks, Daniel -- Daniel Barclay [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
