Messages in this thread | | | Subject | Re: Total freeze with 2.1.12[7,8] | Date | Sun, 15 Nov 1998 10:36:11 -0400 | From | Horst von Brand <> |
| |
"Michael H. Warfield" <mhw@wittsend.com> said: > Matthew Hawkins enscribed thusly:
[...]
> > Just as a general observance, all post-2.1.125 kernels appear to be > > living up to the development kernel inference of instability. 117/9 > > and 125 were fine.
This is i586/100 (UP mobo, SMP disabled). IDE system, lots of modules (CDROM, floppy, Zip; filesystems; SLIP et al; IPv6). RedHat-5.0 + most RedHat-5.1 rpms and updates + lots of installs from source: modutils-2.1.121 (+ preferred patch from RedHat srpm + fix for infinite loop (not advancing on argv) in modinfo), net-tools-1.47, sysvinit-2.26, gpm-1.15.7, rc files hacked to not start kerneld.
I've run lots of post-2.1.125 kernels (and pre-patches, Alan Cox's patches, Andrea Arcangeli's patches, you name it). No real problems at all. egcs-1.1, binutils-2.9.1.0.15 throughout, various glibc-2.0.9x or 10x. CFLAGS for kernel essentially -O2 -fomit-frame-pointer -march=pentium
Currently I'm running 2.1.129pre1 compiled with egcs-1.1.1pre2. No freezes or other strange symptoms. Just that lately (after 2.1.120 or so) the ctrl key seems not to work sometimes (might be the keyboard here), and the text screen seems to have a higher than normal tendency to go into noecho mode when exiting less(1). Can't say for sure, my usage of the system has changed quite a bit.
> > I'm also willing to bet that egcs is the problem - its pentium > > optimisations are probably generating some unwanted/unexpected opcodes. > > Wish I knew enough AT&T syntax to figure it out.
I've been using egcs (snapshots up to 1.1 and lately 1.1.1pre2) here for a _long_ time (february 1998?), only rather minor trouble, lately (after linux-2.1.100 or so at least) none at all. I don't think egcs is at fault here. Haven't used egcs snapshots after 1.1, they seem to hate glibc-2.1 snapshots.
The other reports I've seen here about this kind of problem are running kerneld on 2.1.1xx kernels, or are running a dual CPU system with different CPUs, or are running UP on a SMP mobo. I dimly remember somebody explaining different CPUs _don't_ work; and reports on phantom i386 CPUs on certain mobos for dual CPUs with only one installed.
It just isn't true that this is a generic 2.1.x, x >= 125 UP problem; if it was, the yelling in l-k would have me stone deaf. -- Horst von Brand vonbrand@sleipnir.valparaiso.cl Casilla 9G, Viña del Mar, Chile +56 32 672616
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |