2.6.36-rc1: Doesn't boot on an AMD-based machine
Hi,
The subject says it all, the last message printed is:
CPU0: AMD Athlon(tm) X2 Dual Core Processor L310 stepping 02
so I guess the AMD C1E detection is broken.
I'm going to identify the offending commit.
Thanks,
Rafael
Quoted 9 lines · Rafael J. Wysocki
Hi,
The subject says it all, the last message printed is:
CPU0: AMD Athlon(tm) X2 Dual Core Processor L310 stepping 02
so I guess the AMD C1E detection is broken.
I'm going to identify the offending commit.
There you go:
commit 9d8888c2a214aece2494a49e699a097c2ba9498b
Author: Hans Rosenfeld <[email protected]> Date: Wed Jul 28 19:09:31 2010 +0200
x86, cpu: Clean up AMD erratum 400 workaround
Remove check_c1e_idle() and use the new AMD errata checking framework instead.
LKML-Reference: <[email protected]>
It doesn't revert cleanly so appended is the revert patch I used for the verification if this commit is really responsible for the boot failure.
Thanks,
Rafael
---
From: Rafael J. Wysocki <[email protected]>
Revert commit 9d8888c2a214aece2494a49e699a097c2ba9498b (x86, cpu: Clean up AMD erratum 400 workaround) that causes my Acer Ferrari One, based on AMD Athlon(tm) X2 Dual Core Processor L310 stepping 02, to hang solid during boot (while configuring the CPU).
--- arch/x86/include/asm/processor.h | 1 - arch/x86/kernel/cpu/amd.c | 5 ----- arch/x86/kernel/process.c | 39 +++++++++++++++++++++++++++++++++++++-- 3 files changed, 37 insertions(+), 8 deletions(-)
arch/x86/include/asm/processor.h +0 -1
@@ -1031,7 +1031,6 @@ unsigned long calc_aperfmperf_ratio(stru*/ #ifdef CONFIG_CPU_SUP_AMD extern const int amd_erratum_383[]; -extern const int amd_erratum_400[];extern bool cpu_has_amd_erratum (const int *);#define AMD_LEGACY_ERRATUM (...) { -1, __VA_ARGS__, 0 }
arch/x86/kernel/cpu/amd.c +0 -5
@@ -628,11 +628,6 @@ cpu_dev_register(amd_cpu_dev);* AMD_MODEL_RANGE(0x10, 0x9, 0x0, 0x9, 0x0)); */ -const int amd_erratum_400[] = -AMD_OSVW_ERRATUM (1 ,AMD_MODEL_RANGE (0xf ,0x41 ,0x2 ,0xff ,0xf ), -AMD_MODEL_RANGE (0x10 ,0x2 ,0x1 ,0xff ,0xf )); -EXPORT_SYMBOL_GPL (amd_erratum_400); -const int amd_erratum_383[] =AMD_OSVW_ERRATUM (3 ,AMD_MODEL_RANGE (0x10 ,0 ,0 ,0xff ,0xf ));EXPORT_SYMBOL_GPL (amd_erratum_383);
arch/x86/kernel/process.c +37 -2
@@ -529,6 +529,42 @@ static int __cpuinit mwait_usable(constbool c1e_detected;EXPORT_SYMBOL (c1e_detected); +/* +* Check for AMD CPUs, where APIC timer interrupt does not wake up CPU from C1e. +* For more information see +* - Erratum #400 for NPT family 0xf and family 0x10 CPUs +* - Erratum #365 for family 0x11 (not affected because C1e not in use) +*/ +static int __cpuinit check_c1e_idle (const struct cpuinfo_x86 *c) +{ +u64 val; +if (c->x86_vendor !=X86_VENDOR_AMD ) + goto no_c1e_idle; + +/* Family 0x0f models < rev F do not have C1E */ +if (c->x86 ==0x0F && c->x86_model >=0x40 ) +return 1 ; + +if (c->x86 ==0x10 ) { +/* +* check OSVW bit for CPUs that are not affected +* by erratum #400 +*/ +if (cpu_has (c,X86_FEATURE_OSVW )) { +rdmsrl (MSR_AMD64_OSVW_ID_LENGTH , val); +if (val >=2 ) { +rdmsrl (MSR_AMD64_OSVW_STATUS , val); +if (!(val &BIT (1 ))) + goto no_c1e_idle; + } + } +return 1 ; + } + +no_c1e_idle: +return 0 ; +} +static cpumask_var_t c1e_mask;void c1e_remove_cpu (int cpu)
@@ -605,8 +641,7 @@ void __cpuinit select_idle_routine(const*/ printk (KERN_INFO "using mwait in idle threads.\n" ); pm_idle = mwait_idle; - }else if (cpu_has_amd_erratum (amd_erratum_400)) { -/* E400: APIC timer interrupt does not wake up CPU from C1e */ + }else if (check_c1e_idle (c)) {printk (KERN_INFO "using C1E aware idle routine\n" ); pm_idle = c1e_idle; }else
Quoted 3 lines · Rafael J. Wysocki
It doesn't revert cleanly so appended is the revert patch I used for the verification if this commit is really responsible for the boot failure.
Hans, Peter? What's the status of this? Should I just take the revert from Rafael? Ping?
Linus
H. Peter Anvin
reviewerQuoted 8 lines · Linus Torvalds
Quoted 3 lines · Rafael J. Wysocki
It doesn't revert cleanly so appended is the revert patch I used for the verification if this commit is really responsible for the boot failure.
Hans, Peter? What's the status of this? Should I just take the revert from Rafael? Ping?
I have a fix from Hans queued up in x86/urgent and it will be in the next push of the fixes queue, which I am planning to send you today.
http://git.kernel.org/tip/07a7795ca2e6e66d00b184efb46bd0e23d90d3fe
-hpa
Signature · 2 lines
H. Peter Anvin, Intel Open Source Technology Center I work for Intel. I don't speak on their behalf.
Quoted 5 lines · H. Peter Anvin
I have a fix from Hans queued up in x86/urgent and it will be in the next push of the fixes queue, which I am planning to send you today.
http://git.kernel.org/tip/07a7795ca2e6e66d00b184efb46bd0e23d90d3fe
Rafael, can you confirm that that fixes it?
Linus
Quoted 8 lines · Linus Torvalds
Quoted 5 lines · H. Peter Anvin
I have a fix from Hans queued up in x86/urgent and it will be in the next push of the fixes queue, which I am planning to send you today.
http://git.kernel.org/tip/07a7795ca2e6e66d00b184efb46bd0e23d90d3fe
Rafael, can you confirm that that fixes it?
Yes, it does (sorry for being late).
Rafael