Re: Bug#146883: Kernel compiled for an Alpha SMP is not bootable.
Hi,
Yes, that patch addresses the problem.
The 3.0.4 compiler generates codes that works.
The bottom line is that it is a weird combination of `dead code' in the
kernel and code generation bug.
Thanks to all of you!
On Wed, 2002-05-15 at 21:34, Christopher C. Chimelis wrote:
>
> On 13 May 2002, Frederic Roussel wrote:
>
> > Package: gcc-2.95
> > Version: 1:2.95.4-9
> > Severity: normal
> >
> > -- System Information
> > Debian Release: 3.0
> > Kernel Version: Linux big2 2.4.19-pre8-30 #1 SMP Mon May 13 18:01:01 PDT
> > 2002 alpha unknown
> >
> > Versions of the packages gcc-2.95 depends on:
> > ii binutils 2.12.90.0.1-5 The GNU assembler, linker and binary
> > utiliti
> > ii cpp-2.95 2.95.4-9 The GNU C preprocessor.
> > ii gcc 2.95.4-14 The GNU C compiler.
> > ii libc6.1 2.2.5-6 GNU C Library: Shared libraries and
> > Timezone
> >
> > If I compile a kernel 2.4.x (I tried with x >= 17) on a Compaq DS20, the
> > newly compiled kernel oopses in the fontion:
> > smp_tune_scheduling defined in file arch/alpha/kernel/smp.c
> >
> > The `switch' statement in that function is where the Oops happens, at
> > least I believe so: as an experiment I've hard-coded the `on_chip_cache'
> > variable and commented out the whole switch, the kernel booted up fine.
> >
> > I compiled with gcc-3.0.4 and the problem disappeared.
>
> Try the patch below to see if it fixes things. There is a code generation
> problem in 2.95.4, but it's mostly irrelevent since the kernel code that's
> being miscompiled doesn't even need to be used in the 2.4 kernels. I
> tried submitting this patch to upstream kernel folks, but it got ignored
> several times and I didn't have the time to follow up further.
>
> C
>
> Index: arch/alpha/kernel/smp.c
> ===================================================================
> RCS file: /home/cvs/linux-24/arch/alpha/kernel/smp.c,v
> retrieving revision 1.1.1.4
> retrieving revision 1.4
> diff -u -r1.1.1.4 -r1.4
> --- arch/alpha/kernel/smp.c 11 Oct 2001 15:41:35 -0000 1.1.1.4
> +++ arch/alpha/kernel/smp.c 11 Oct 2001 15:51:36 -0000 1.4
> @@ -80,7 +80,6 @@
> int smp_num_probed; /* Internal processor count */
> int smp_num_cpus = 1; /* Number that came online. */
> int smp_threads_ready; /* True once the per process idle is forked. */
> -cycles_t cacheflush_time;
>
> int __cpu_number_map[NR_CPUS];
> int __cpu_logical_map[NR_CPUS];
> @@ -212,63 +211,6 @@
>
>
> /*
> - * Rough estimation for SMP scheduling, this is the number of cycles it
> - * takes for a fully memory-limited process to flush the SMP-local cache.
> - *
> - * We are not told how much cache there is, so we have to guess.
> - */
> -static void __init
> -smp_tune_scheduling (int cpuid)
> -{
> - struct percpu_struct *cpu;
> - unsigned long on_chip_cache;
> - unsigned long freq;
> -
> - cpu = (struct percpu_struct*)((char*)hwrpb + hwrpb->processor_offset
> - + cpuid * hwrpb->processor_size);
> - switch (cpu->type)
> - {
> - case EV45_CPU:
> - on_chip_cache = 16 + 16;
> - break;
> -
> - case EV5_CPU:
> - case EV56_CPU:
> - on_chip_cache = 8 + 8 + 96;
> - break;
> -
> - case PCA56_CPU:
> - on_chip_cache = 16 + 8;
> - break;
> -
> - case EV6_CPU:
> - case EV67_CPU:
> - on_chip_cache = 64 + 64;
> - break;
> -
> - default:
> - on_chip_cache = 8 + 8;
> - break;
> - }
> -
> - freq = hwrpb->cycle_freq ? : est_cycle_freq;
> -
> -#if 0
> - /* Magic estimation stolen from x86 port. */
> - cacheflush_time = freq / 1024L * on_chip_cache / 5000L;
> -
> - printk("Using heuristic of %d cycles.\n",
> - cacheflush_time);
> -#else
> - /* Magic value to force potential preemption of other CPUs. */
> - cacheflush_time = INT_MAX;
> -
> - printk("Using heuristic of %d cycles.\n",
> - cacheflush_time);
> -#endif
> -}
> -
> -/*
> * Send a message to a secondary's console. "START" is one such
> * interesting message. ;-)
> */
> @@ -607,7 +549,6 @@
> current->processor = boot_cpuid;
>
> smp_store_cpu_info(boot_cpuid);
> - smp_tune_scheduling(boot_cpuid);
> smp_setup_percpu_timer(boot_cpuid);
>
> init_idle();
--
Frederic.R.Roussel -o) (o-
/\\ Join the penguin force (o_ (o_ //\
_\_v The Linux G3N3R47!0N (/)_ (/)_ v_/_
--
To UNSUBSCRIBE, email to debian-gcc-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: