On Mon, 14 Feb 2000, Riley Williams wrote:
Sorry for spoiling a party, but it does NOT even compile (up to
2.2.15pre7)...
=== Cut ===
make -C hfmodem modules
make[3]: Entering directory `/tmp/build-kernel/usr/src/linux-2.2.15pre7/drivers/char/hfmodem'
gcc -D__KERNEL__ -I/tmp/build-kernel/usr/src/linux-2.2.15pre7/include -Wall -Wstrict-prototypes -O2 -fomit-frame-pointer -fno-strict-aliasing -D__SMP__ -pipe -march=pentiumpro -mcpu=pentiumpro -malign-loops=2 -malign-jumps=2 -malign-functions=2 -DCPU=686 -DMODULE -DMODVERSIONS -include /tmp/build-kernel/usr/src/linux-2.2.15pre7/include/linux/modversions.h -c -o refclock.o refclock.c
refclock.c: In function `hfmodem_refclock_current':
refclock.c:136: Invalid `asm' statement:
refclock.c:136: fixed or forbidden register 0 (ax) was spilled for class AREG.
make[3]: *** [refclock.o] Error 1
make[3]: Leaving directory `/tmp/build-kernel/usr/src/linux-2.2.15pre7/drivers/char/hfmodem'
make[2]: *** [_modsubdir_hfmodem] Error 2
make[2]: Leaving directory `/tmp/build-kernel/usr/src/linux-2.2.15pre7/drivers/char'
make[1]: *** [_modsubdir_char] Error 2
make[1]: Leaving directory `/tmp/build-kernel/usr/src/linux-2.2.15pre7/drivers'
make: *** [_mod_drivers] Error 2
=== Cut ===
I DO know that hfmodem is not the most used part of the kernel. I DO know
that it'd probably compile with kgcc-2.7.2.3 (although breaking in some
other part), but it's to illustrate that NO vanilla 2.2.xx kernel is
compileable outta the box...
People !!! It's already 15th release candidate (less for all that pre's and
ac's) and it does NOT even compile... "Use kgcc-2.7.2.3" is NOT to be taken
as a valid excuse, it's simply broken... There have been patches addressing
wrong assembly in 2.2.xx in l-k but nobody cared to merge them.
Anyone else for a "stable" kernel ? Someone to call the configuration with
hfmodem exotic ? As I do recall, someone here did say that he does compile
all 2.2.xx kernels with all imaginable config combination with 100% success
rate...
I do not even try to mention the parts which are not SMP-safe...
> Hi Dominik.
>
> >> I've built EVERY one of the 2.2 series kernels several times for
> >> several different configurations, and I've yet to find one that
> >> was unstable, announced or otherwise.
>
> > And because you didnīt encounter instability it does not exist?
>
> Instability ALWAYS exists, given the right hardware combination.
>
> > I have three systems here (Dual P54C-200, 128 MB RAM DPT-3334UW,
> > AHA-2940UW, AHA-2940 and 3c905TX) which froze with every kernel
> > between 2.2.5 and 2.2.13 within hours. No OOPS, no nothing. Just
> > dead.
>
> The most obvious point is that they are "Dual P54C-200" systems, which
> indicates that they are SMP systems ("Dual") running with non-MMX
> Pentium processors ("P54C") clocked at 200 MHz ("-200").
>
> ALAN: Am I right in thinking that there's a known bug in the P54C
> processor regarding its use in SMP configurations, and one needs the
> P55C (Px/MMX) processor as a minimum for reliable SMP operation?
>
> > And since i didnīt like to check that 150GB Hardware-RAID5 every
> > boot, i went back to 2.0.36 UP and enjoyed a rock solid system.
> > (Just compiling 2.2.x as UP was not enough...)
>
> That sounds like the bug referred to above is the cause.
>
> > Now finally 2.2.14 appears to be stable on this hardware: one
> > system is up for over four days now... As soon as the uptime of
> > this system reaches 30 days i will consider switching the system
> > with the RAID over to 2.2.14...
>
> It's good to hear, but I would still be wary of your choice of
> processor in an SMP system.
>
> Best wishes from Riley.
>
>
> -
> 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/
>
===========================================================================
Sergey Kubushin aka the Tamer < > The impossible we do immediately.
e-mail: ksi@ksi-linux.com SK320-RIPE < > Miracles require 24-hour notice.
===========================================================================
-
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/
This archive was generated by hypermail 2b29 : Tue Feb 15 2000 - 21:00:27 EST