https://bugzilla.novell.com/show_bug.cgi?id=338319 Summary: Kernel Locks Up With HPT366 Disks on BP6 SMP During udev Hardware Detection Product: openSUSE 10.3 Version: Final Platform: i686 OS/Version: openSUSE 10.3 Status: NEW Severity: Major Priority: P5 - None Component: Kernel AssignedTo: kernel-maintainers@forge.provo.novell.com ReportedBy: rob.opensuse.linux@googlemail.com QAContact: qa@suse.de Found By: Customer Kernel locks up if devices are attached to HPT366 UDMA-66 controller on Abit BP6 motherboard, on every OS-10.3 boot. The machine runs very stabely with SuSE-8.2 and a Linux 2.4.20-64GB-SMP kernel from k_smp-2.4.20-102 rpm. My hunch is that it's a 2.6 HPT366 driver problem caused by the PATA reworking of the IDE code from 2.4. Reboot to configure running system after initial install failed, after installing with nodma had appeared to succeed. In past I'd tweak the IDE settings after a nodma install to force the disk to UDMA3 mode, and set the CD-RW to mdma2, which would then run pretty stably. Now the system hangs every time if HPT366 is active, when udev has started. ( This also occured with a recent Gentoo-sources kernel as well so I don't think it's SuSE patches responsible, but the new drivers reworking the IDE code in mainstream stock Linus codebase. ) Usually the last message is : rtc_cmos : probe of 00:04 failed error -16 udevd_event [1558] : mode_symlink device node /dev/rtc0 already exists ... But on 1 occasion I saw a line or two about CD-RW, making me suspect the IDE subsystem. Another time after enabling ACPI (a no-no in past) as this machine ran SuSE 6.2, 7.0 .. 7.3 and 8.0..8.2; I saw a line following /dev/rtc0 about ACPI. The symtoms are similar to the old BP6 lock-up problem in linux-2.2 which I think Andre Hedrik (who had a BP6 at one point) cured with IDE re-working for 2.4. Perhaps there'll be a clue in the 2.4 source or changelog? The I/O LED stays on. No keyboard response. Machine totally frozen until reset is pushed. The 2.4 series kernels improved stability of these boards to acceptable, which under 2.2 had trouble locking up regularly. Seems like the reworking of the IDE layers, has caused a regression. I did manage to run a Linus stock kernel 2.6.0 (after upgrading 1 or 2 tools), so I think it's been introduced in a later version. I've classified this as 'Major' because it causes installation to fail, and furthermore it's hard to use an old machine with just 2 IDE motherboard connectors so the work round makes the machine pretty useless, whilst with multiple disks and CD-RW it's a good testbed system. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug.