From: Wanpeng Li <hidden> Date: 2016-06-22 01:28:49
From: Wanpeng Li <redacted>
------------[ cut here ]------------
WARNING: CPU: 0 PID: 0 at arch/x86/mm/extable.c:50 ex_handler_rdmsr_unsafe+0x6a/0x70
unchecked MSR access error: RDMSR from 0xce
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.7.0-rc3+ #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
0000000000000000 ffffffff81c03ce0 ffffffff813b3eae ffffffff81c03d30
0000000000000000 ffffffff81c03d20 ffffffff81067181 0000003200000001
ffffffff81c03df8 ffffffff8179676c 0000000000000000 ffffffff81fcd2c0
Call Trace:
dump_stack+0x67/0x99
__warn+0xd1/0xf0
warn_slowpath_fmt+0x4f/0x60
ex_handler_rdmsr_unsafe+0x6a/0x70
fixup_exception+0x39/0x50
do_general_protection+0x93/0x1b0
general_protection+0x22/0x30
? cpu_khz_from_msr+0xd8/0x1c0
native_calibrate_cpu+0x30/0x5b0
tsc_init+0x2b/0x297
x86_late_time_init+0xf/0x11
start_kernel+0x398/0x451
? set_init_arg+0x55/0x55
x86_64_start_reservations+0x2f/0x31
x86_64_start_kernel+0xea/0xed
After commit (fc141535ad8 : "x86 tsc_msr: Extend to include Intel Core
Architecture"), rdmsr MSR_PLATFORM_INFO is used to get maximum non-turbo
ratio for recent Intel Core Architecture which results in kvm guest rdmsr
unsafe warning.
As Radim pointed out before:
| MSR_PLATFORM_INFO: Intel changes it from family to family and there is
| no obvious overlap or default. If we picked 0 (any other fixed value),
| then the guest would have to know that 0 doesn't mean that
| MSR_PLATFORM_INFO returned 0, but that KVM doesn't emulate this MSR and
| the value cannot be used. This is very similar to handling a #GP in the
| guest, but also has a disadvantage, because KVM cannot say that
| MSR_PLATFORM_INFO is 0. Simple emulation is not possible.
This patch fix it by using rdmsr_safe to read MSR_PLATFORM_INFO in kvm
guest in order that #GP can be fixed up, then tsc will be calibrated by
PIT, HPET etc.
Reported-by: kernel test robot <redacted>
Cc: Len Brown <redacted>
Cc: "Rafael J. Wysocki" <redacted>
Cc: Zhang Rui <rui.zhang@intel.com>
Cc: Chen Yu <redacted>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Radim Krčmář <redacted>
Cc: jacob.jun.pan@intel.com
Cc: Thomas Gleixner <redacted>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Wanpeng Li <redacted>
---
arch/x86/kernel/tsc_msr.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
@@ -70,7 +70,7 @@ static int match_cpu(u8 family, u8 model)*/unsignedlongcpu_khz_from_msr(void){-u32lo,hi,ratio,freq_id,freq;+u32lo,hi,freq_id,freq,ratio=0;unsignedlongres;intcpu_index;
@@ -123,8 +123,8 @@ unsigned long cpu_khz_from_msr(void)}get_ratio:-rdmsr(MSR_PLATFORM_INFO,lo,hi);-ratio=(lo>>8)&0xff;+if(!rdmsr_safe(MSR_PLATFORM_INFO,&lo,&hi))+ratio=(lo>>8)&0xff;done:/* TSC frequency = maximum resolved freq * maximum resolved bus ratio */
I prefer this patch can be applied separately instead of fold into the
bad commit since it shows the issue when access MSR_PLATFORM_INFO in
kvm guest and other guys who want to access MSR_PLATFORM_INFO later
can find the changelog and make better decisions.
Regards,
Wanpeng Li
On Thu, Jun 23, 2016 at 3:09 PM, Thomas Gleixner [off-list ref] wrote:
On Wed, 22 Jun 2016, Wanpeng Li wrote:
quoted
After commit (fc141535ad8 : "x86 tsc_msr: Extend to include Intel Core
Where did you find that commit? It's neither in Linus tree nor in tip.
It is in Len's tree, we are planing to resend the patchset with Wanpeng's fix
merged with a credit to him in commit msg, thanks for Wanpeng's effort.
thanks all.
From: Wanpeng Li <hidden> Date: 2016-06-23 07:41:16
Hi Yu，
2016-06-23 15:37 GMT+08:00 Yu Chen [off-list ref]:
On Thu, Jun 23, 2016 at 3:09 PM, Thomas Gleixner [off-list ref] wrote:
quoted
On Wed, 22 Jun 2016, Wanpeng Li wrote:
quoted
After commit (fc141535ad8 : "x86 tsc_msr: Extend to include Intel Core
Where did you find that commit? It's neither in Linus tree nor in tip.
It is in Len's tree, we are planing to resend the patchset with Wanpeng's fix
merged with a credit to him in commit msg, thanks for Wanpeng's effort.
thanks all.
I prefer this patch can be applied separately instead of fold into the
bad commit since it shows the issue when access MSR_PLATFORM_INFO in
kvm guest and other guys who want to access MSR_PLATFORM_INFO later
can find the changelog and make better decisions.
Regards,
Wanpeng Li
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
From: Wanpeng Li <hidden> Date: 2016-06-23 08:00:13
2016-06-23 15:41 GMT+08:00 Wanpeng Li [off-list ref]:
Hi Yu，
2016-06-23 15:37 GMT+08:00 Yu Chen [off-list ref]:
quoted
On Thu, Jun 23, 2016 at 3:09 PM, Thomas Gleixner [off-list ref] wrote:
quoted
On Wed, 22 Jun 2016, Wanpeng Li wrote:
quoted
After commit (fc141535ad8 : "x86 tsc_msr: Extend to include Intel Core
Where did you find that commit? It's neither in Linus tree nor in tip.
It is in Len's tree, we are planing to resend the patchset with Wanpeng's fix
merged with a credit to him in commit msg, thanks for Wanpeng's effort.
thanks all.
I prefer this patch can be applied separately instead of fold into the
bad commit since it shows the issue when access MSR_PLATFORM_INFO in
kvm guest and other guys who want to access MSR_PLATFORM_INFO later
can find the changelog and make better decisions.
Thomas, does it make sense to keep separate?
Regards,
Wanpeng Li
From: Wanpeng Li <hidden> Date: 2016-07-11 02:38:37
Hi Ingo, Thomas,
2016-06-22 9:28 GMT+08:00 Wanpeng Li [off-list ref]:
From: Wanpeng Li <redacted>
------------[ cut here ]------------
WARNING: CPU: 0 PID: 0 at arch/x86/mm/extable.c:50 ex_handler_rdmsr_unsafe+0x6a/0x70
unchecked MSR access error: RDMSR from 0xce
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.7.0-rc3+ #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
0000000000000000 ffffffff81c03ce0 ffffffff813b3eae ffffffff81c03d30
0000000000000000 ffffffff81c03d20 ffffffff81067181 0000003200000001
ffffffff81c03df8 ffffffff8179676c 0000000000000000 ffffffff81fcd2c0
Call Trace:
dump_stack+0x67/0x99
__warn+0xd1/0xf0
warn_slowpath_fmt+0x4f/0x60
ex_handler_rdmsr_unsafe+0x6a/0x70
fixup_exception+0x39/0x50
do_general_protection+0x93/0x1b0
general_protection+0x22/0x30
? cpu_khz_from_msr+0xd8/0x1c0
native_calibrate_cpu+0x30/0x5b0
tsc_init+0x2b/0x297
x86_late_time_init+0xf/0x11
start_kernel+0x398/0x451
? set_init_arg+0x55/0x55
x86_64_start_reservations+0x2f/0x31
x86_64_start_kernel+0xea/0xed
After commit (fc141535ad8 : "x86 tsc_msr: Extend to include Intel Core
Architecture"), rdmsr MSR_PLATFORM_INFO is used to get maximum non-turbo
I just saw commit (fc273eeef314c : "x86 tsc_msr: Extend to include
Intel Core Architecture") was merged in tip tree, then this patch is
needed to fix that commit. The bug is reported by LKP several weeks
ago, and kvm maintainer Paolo has already replied "This looks good" to
this patch.
Regards,
Wanpeng Li
quoted hunk
ratio for recent Intel Core Architecture which results in kvm guest rdmsr
unsafe warning.
As Radim pointed out before:
| MSR_PLATFORM_INFO: Intel changes it from family to family and there is
| no obvious overlap or default. If we picked 0 (any other fixed value),
| then the guest would have to know that 0 doesn't mean that
| MSR_PLATFORM_INFO returned 0, but that KVM doesn't emulate this MSR and
| the value cannot be used. This is very similar to handling a #GP in the
| guest, but also has a disadvantage, because KVM cannot say that
| MSR_PLATFORM_INFO is 0. Simple emulation is not possible.
This patch fix it by using rdmsr_safe to read MSR_PLATFORM_INFO in kvm
guest in order that #GP can be fixed up, then tsc will be calibrated by
PIT, HPET etc.
Reported-by: kernel test robot <redacted>
Cc: Len Brown <redacted>
Cc: "Rafael J. Wysocki" <redacted>
Cc: Zhang Rui <rui.zhang@intel.com>
Cc: Chen Yu <yu.c.chen@intel.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: Radim Krčmář <redacted>
Cc: jacob.jun.pan@intel.com
Cc: Thomas Gleixner <redacted>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: "H. Peter Anvin" <hpa@zytor.com>
Signed-off-by: Wanpeng Li <redacted>
---
arch/x86/kernel/tsc_msr.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
@@ -70,7 +70,7 @@ static int match_cpu(u8 family, u8 model)*/unsignedlongcpu_khz_from_msr(void){-u32lo,hi,ratio,freq_id,freq;+u32lo,hi,freq_id,freq,ratio=0;unsignedlongres;intcpu_index;
@@ -123,8 +123,8 @@ unsigned long cpu_khz_from_msr(void)}get_ratio:-rdmsr(MSR_PLATFORM_INFO,lo,hi);-ratio=(lo>>8)&0xff;+if(!rdmsr_safe(MSR_PLATFORM_INFO,&lo,&hi))+ratio=(lo>>8)&0xff;done:/* TSC frequency = maximum resolved freq * maximum resolved bus ratio */--
Hi Ingo, Thomas,
2016-06-22 9:28 GMT+08:00 Wanpeng Li [off-list ref]:
quoted
From: Wanpeng Li <redacted>
------------[ cut here ]------------
WARNING: CPU: 0 PID: 0 at arch/x86/mm/extable.c:50 ex_handler_rdmsr_unsafe+0x6a/0x70
unchecked MSR access error: RDMSR from 0xce
Modules linked in:
CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.7.0-rc3+ #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
0000000000000000 ffffffff81c03ce0 ffffffff813b3eae ffffffff81c03d30
0000000000000000 ffffffff81c03d20 ffffffff81067181 0000003200000001
ffffffff81c03df8 ffffffff8179676c 0000000000000000 ffffffff81fcd2c0
Call Trace:
dump_stack+0x67/0x99
__warn+0xd1/0xf0
warn_slowpath_fmt+0x4f/0x60
ex_handler_rdmsr_unsafe+0x6a/0x70
fixup_exception+0x39/0x50
do_general_protection+0x93/0x1b0
general_protection+0x22/0x30
? cpu_khz_from_msr+0xd8/0x1c0
native_calibrate_cpu+0x30/0x5b0
tsc_init+0x2b/0x297
x86_late_time_init+0xf/0x11
start_kernel+0x398/0x451
? set_init_arg+0x55/0x55
x86_64_start_reservations+0x2f/0x31
x86_64_start_kernel+0xea/0xed
After commit (fc141535ad8 : "x86 tsc_msr: Extend to include Intel Core
Architecture"), rdmsr MSR_PLATFORM_INFO is used to get maximum non-turbo
I just saw commit (fc273eeef314c : "x86 tsc_msr: Extend to include
Intel Core Architecture") was merged in tip tree, then this patch is
needed to fix that commit. The bug is reported by LKP several weeks
ago, and kvm maintainer Paolo has already replied "This looks good" to
this patch.
From: tip-bot for Wanpeng Li <hidden> Date: 2016-07-11 07:41:14
Commit-ID: 37c528ee1af7f24eb31f4195b8b7d4f23e6c716d
Gitweb: http://git.kernel.org/tip/37c528ee1af7f24eb31f4195b8b7d4f23e6c716d
Author: Wanpeng Li [off-list ref]
AuthorDate: Wed, 22 Jun 2016 09:28:28 +0800
Committer: Ingo Molnar [off-list ref]
CommitDate: Mon, 11 Jul 2016 09:20:36 +0200
x86/tsc_msr: Fix rdmsr(MSR_PLATFORM_INFO) unsafe warning in KVM guest
After this commit:
fc273eeef314 ("x86/tsc_msr: Extend to include Intel Core Architecture")
The following unsafe MSR reading warning triggers:
WARNING: CPU: 0 PID: 0 at arch/x86/mm/extable.c:50 ex_handler_rdmsr_unsafe+0x6a/0x70
unchecked MSR access error: RDMSR from 0xce
Call Trace:
dump_stack+0x67/0x99
__warn+0xd1/0xf0
warn_slowpath_fmt+0x4f/0x60
ex_handler_rdmsr_unsafe+0x6a/0x70
fixup_exception+0x39/0x50
do_general_protection+0x93/0x1b0
general_protection+0x22/0x30
? cpu_khz_from_msr+0xd8/0x1c0
native_calibrate_cpu+0x30/0x5b0
tsc_init+0x2b/0x297
x86_late_time_init+0xf/0x11
start_kernel+0x398/0x451
? set_init_arg+0x55/0x55
x86_64_start_reservations+0x2f/0x31
x86_64_start_kernel+0xea/0xed
As Radim pointed out before:
| MSR_PLATFORM_INFO: Intel changes it from family to family and there is
| no obvious overlap or default. If we picked 0 (any other fixed value),
| then the guest would have to know that 0 doesn't mean that
| MSR_PLATFORM_INFO returned 0, but that KVM doesn't emulate this MSR and
| the value cannot be used. This is very similar to handling a #GP in the
| guest, but also has a disadvantage, because KVM cannot say that
| MSR_PLATFORM_INFO is 0. Simple emulation is not possible.
Fix it by using rdmsr_safe(MSR_PLATFORM_INFO) in KVM guest to
not trigger a #GP, then tsc will be calibrated by a fallback
method: PIT, HPET etc.
Reported-by: kernel test robot <redacted>
Signed-off-by: Wanpeng Li <redacted>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Cc: Chen Yu <redacted>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Len Brown <redacted>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Linux PM list <redacted>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Radim Krčmář <redacted>
Cc: Rafael J. Wysocki <redacted>
Cc: Thomas Gleixner <redacted>
Cc: Zhang Rui <rui.zhang@intel.com>
Cc: jacob.jun.pan@intel.com
Cc: kvm@vger.kernel.org
Cc: linux-acpi@vger.kernel.org
Cc: lkp@01.org
Link: http://lkml.kernel.org/r/1466558908-3524-1-git-send-email-wanpeng.li@hotmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
---
arch/x86/kernel/tsc_msr.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
@@ -70,7 +70,7 @@ static int match_cpu(u8 family, u8 model)*/unsignedlongcpu_khz_from_msr(void){-u32lo,hi,ratio,freq_id,freq;+u32lo,hi,freq_id,freq,ratio=0;unsignedlongres;intcpu_index;
@@ -123,8 +123,8 @@ unsigned long cpu_khz_from_msr(void)}get_ratio:-rdmsr(MSR_PLATFORM_INFO,lo,hi);-ratio=(lo>>8)&0xff;+if(!rdmsr_safe(MSR_PLATFORM_INFO,&lo,&hi))+ratio=(lo>>8)&0xff;done:/* TSC frequency = maximum resolved freq * maximum resolved bus ratio */