x86/tsc: Cure fallout from the TSC ADJUST series
cover x86/tsc: Cure fallout from the TSC ADJUST series
The following patches address the fallout of the TSC ADJUST changes in Linus tree.
- Forward jump of dmesg timestamps
- Make it work when TSC is marked reliable
Thanks,
tglx
PATCH 1/2 x86/tsc: Avoid the large time jump when sanitizing TSC ADJUST
Olof reported that on a machine which has a BIOS wreckaged TSC the timestamps in dmesg are making a large jump because the TSC value is jumping forward after resetting the TSC ADJUST register to a sane value.
This can be avoided by calling the TSC ADJUST saniziting function before initializing the per cpu sched clock machinery. That takes the offset into account and avoid the time jump.
What cannot be avoided is that the 'Firmware Bug' warnings on the secondary CPUs are printed with the large time offsets because it would be too much effort and ugly hackery to print those warnings into a buffer and emit them after the adjustemt on the starting CPUs. It's a firmware bug and should be fixed in firmware. The weird timestamps are collateral damage and just illustrate the sillyness of the BIOS folks:
[ 0.397445] smp: Bringing up secondary CPUs ... [ 0.402100] x86: Booting SMP configuration: [ 0.406343] .... node #0, CPUs: #1 [1265776479.930667] [Firmware Bug]: TSC ADJUST differs: Reference CPU0: -2978888639075328 CPU1: -2978888639183101 [1265776479.944664] TSC ADJUST synchronize: Reference CPU0: 0 CPU1: -2978888639183101 [ 0.508119] #2 [1265776480.032346] [Firmware Bug]: TSC ADJUST differs: Reference CPU0: -2978888639075328 CPU2: -2978888639183677 [1265776480.044192] TSC ADJUST synchronize: Reference CPU0: 0 CPU2: -2978888639183677 [ 0.607643] #3 [1265776480.131874] [Firmware Bug]: TSC ADJUST differs: Reference CPU0: -2978888639075328 CPU3: -2978888639184530 [1265776480.143720] TSC ADJUST synchronize: Reference CPU0: 0 CPU3: -2978888639184530 [ 0.707108] smp: Brought up 1 node, 4 CPUs [ 0.711271] smpboot: Total of 4 processors activated (21698.88 BogoMIPS)
--- arch/x86/kernel/tsc.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
arch/x86/kernel/tsc.c +3 -2
@@ -1356,6 +1356,9 @@ void __init tsc_init(void) (unsigned long )cpu_khz /1000 , (unsigned long)cpu_khz % 1000); +/* Sanitize TSC ADJUST before cyc2ns gets initialized */ +tsc_store_and_check_tsc_adjust (true); +/* * Secondary CPUs do not run through tsc_init(), so set up * all the scale factors for all CPUs, assuming the same
@@ -1386,8 +1389,6 @@ void __init tsc_init(void)if (unsynchronized_tsc ())mark_tsc_unstable ("TSCs unsynchronized" ); -else -tsc_store_and_check_tsc_adjust (true);check_system_tsc_reliable ();
tip-bot for Thomas Gleixner
merge bot[tip:x86/urgent] x86/tsc: Avoid the large time jump when sanitizing TSC ADJUST
x86/tsc: Avoid the large time jump when sanitizing TSC ADJUST
Commit-ID: f2e04214ef7f7e49d1e06109ad1b2718155dab25 Gitweb: http://git.kernel.org/tip/f2e04214ef7f7e49d1e06109ad1b2718155dab25 Author: Thomas Gleixner <[email protected]> AuthorDate: Thu, 9 Feb 2017 16:08:41 +0100 Committer: Thomas Gleixner <[email protected]> CommitDate: Fri, 10 Feb 2017 09:47:16 +0100
x86/tsc: Avoid the large time jump when sanitizing TSC ADJUST
Olof reported that on a machine which has a BIOS wreckaged TSC the timestamps in dmesg are making a large jump because the TSC value is jumping forward after resetting the TSC ADJUST register to a sane value.
This can be avoided by calling the TSC ADJUST saniziting function before initializing the per cpu sched clock machinery. That takes the offset into account and avoid the time jump.
What cannot be avoided is that the 'Firmware Bug' warnings on the secondary CPUs are printed with the large time offsets because it would be too much effort and ugly hackery to print those warnings into a buffer and emit them after the adjustemt on the starting CPUs. It's a firmware bug and should be fixed in firmware. The weird timestamps are collateral damage and just illustrate the sillyness of the BIOS folks:
[ 0.397445] smp: Bringing up secondary CPUs ... [ 0.402100] x86: Booting SMP configuration: [ 0.406343] .... node #0, CPUs: #1 [1265776479.930667] [Firmware Bug]: TSC ADJUST differs: Reference CPU0: -2978888639075328 CPU1: -2978888639183101 [1265776479.944664] TSC ADJUST synchronize: Reference CPU0: 0 CPU1: -2978888639183101 [ 0.508119] #2 [1265776480.032346] [Firmware Bug]: TSC ADJUST differs: Reference CPU0: -2978888639075328 CPU2: -2978888639183677 [1265776480.044192] TSC ADJUST synchronize: Reference CPU0: 0 CPU2: -2978888639183677 [ 0.607643] #3 [1265776480.131874] [Firmware Bug]: TSC ADJUST differs: Reference CPU0: -2978888639075328 CPU3: -2978888639184530 [1265776480.143720] TSC ADJUST synchronize: Reference CPU0: 0 CPU3: -2978888639184530 [ 0.707108] smp: Brought up 1 node, 4 CPUs [ 0.711271] smpboot: Total of 4 processors activated (21698.88 BogoMIPS)
--- arch/x86/kernel/tsc.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
arch/x86/kernel/tsc.c +3 -2
@@ -1356,6 +1356,9 @@ void __init tsc_init(void) (unsigned long )cpu_khz /1000 , (unsigned long)cpu_khz % 1000); +/* Sanitize TSC ADJUST before cyc2ns gets initialized */ +tsc_store_and_check_tsc_adjust (true); +/* * Secondary CPUs do not run through tsc_init(), so set up * all the scale factors for all CPUs, assuming the same
@@ -1386,8 +1389,6 @@ void __init tsc_init(void)if (unsynchronized_tsc ())mark_tsc_unstable ("TSCs unsynchronized" ); -else -tsc_store_and_check_tsc_adjust (true);check_system_tsc_reliable ();
PATCH 2/2 x86/tsc: Make the TSC ADJUST sanitizing work for tsc_reliable
In reply to Thomas GleixnerWhen the TSC is marked reliable then the synchronization check is skipped, but that also skips the TSC ADJUST sanitizing code. So on a machine with a wreckaged BIOS the TSC deviation between CPUs might go unnoticed.
Let the TSC adjust sanitizing code run unconditionally and just skip the expensive synchronization checks when TSC is marked reliable.
--- arch/x86/kernel/tsc_sync.c | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-)
arch/x86/kernel/tsc_sync.c +7 -9
@@ -286,13 +286,6 @@ void check_tsc_sync_source(int cpu)if (unsynchronized_tsc ())return ; -if (tsc_clocksource_reliable) { -if (cpu == (nr_cpu_ids-1 ) || system_state !=SYSTEM_BOOTING ) -pr_info ( -"Skipped synchronization checks as TSC is reliable.\n" ); -return ; - } -/* * Set the maximum number of test runs to * 1 if the CPU does not provide the TSC_ADJUST MSR
@@ -380,14 +373,19 @@ void check_tsc_sync_target(void)int cpus =2 ;/* Also aborts if there is no TSC. */ -if (unsynchronized_tsc () || tsc_clocksource_reliable) +if (unsynchronized_tsc ())return ;/* * Store, verify and sanitize the TSC adjust register. If * successful skip the test. +* +* The test is also skipped when the TSC is marked reliable. This +* is true for SoCs which have no fallback clocksource. On these +* SoCs the TSC is frequency synchronized, but still the TSC ADJUST +* register might have been wreckaged by the BIOS.. */ -if (tsc_store_and_check_tsc_adjust (false)) { +if (tsc_store_and_check_tsc_adjust (false) || tsc_clocksource_reliable) {atomic_inc (&skip_test);return ; }
tip-bot for Thomas Gleixner
merge bot[tip:x86/urgent] x86/tsc: Make the TSC ADJUST sanitizing work for tsc_reliable
x86/tsc: Make the TSC ADJUST sanitizing work for tsc_reliable
Commit-ID: 5f2e71e71410ecb858cfec184ba092adaca61626 Gitweb: http://git.kernel.org/tip/5f2e71e71410ecb858cfec184ba092adaca61626 Author: Thomas Gleixner <[email protected]> AuthorDate: Thu, 9 Feb 2017 16:08:42 +0100 Committer: Thomas Gleixner <[email protected]> CommitDate: Fri, 10 Feb 2017 09:47:17 +0100
x86/tsc: Make the TSC ADJUST sanitizing work for tsc_reliable
When the TSC is marked reliable then the synchronization check is skipped, but that also skips the TSC ADJUST sanitizing code. So on a machine with a wreckaged BIOS the TSC deviation between CPUs might go unnoticed.
Let the TSC adjust sanitizing code run unconditionally and just skip the expensive synchronization checks when TSC is marked reliable.
--- arch/x86/kernel/tsc_sync.c | 16 +++++++--------- 1 file changed, 7 insertions(+), 9 deletions(-)
arch/x86/kernel/tsc_sync.c +7 -9
@@ -286,13 +286,6 @@ void check_tsc_sync_source(int cpu)if (unsynchronized_tsc ())return ; -if (tsc_clocksource_reliable) { -if (cpu == (nr_cpu_ids-1 ) || system_state !=SYSTEM_BOOTING ) -pr_info ( -"Skipped synchronization checks as TSC is reliable.\n" ); -return ; - } -/* * Set the maximum number of test runs to * 1 if the CPU does not provide the TSC_ADJUST MSR
@@ -380,14 +373,19 @@ void check_tsc_sync_target(void)int cpus =2 ;/* Also aborts if there is no TSC. */ -if (unsynchronized_tsc () || tsc_clocksource_reliable) +if (unsynchronized_tsc ())return ;/* * Store, verify and sanitize the TSC adjust register. If * successful skip the test. +* +* The test is also skipped when the TSC is marked reliable. This +* is true for SoCs which have no fallback clocksource. On these +* SoCs the TSC is frequency synchronized, but still the TSC ADJUST +* register might have been wreckaged by the BIOS.. */ -if (tsc_store_and_check_tsc_adjust (false)) { +if (tsc_store_and_check_tsc_adjust (false) || tsc_clocksource_reliable) {atomic_inc (&skip_test);return ; }