xen /proc/mtrr implementation
Quoted 7 lines · Ingo Molnar
Quoted 5 lines
Quoted 3 lines
Quoted 1 line
Oops, the third "proper technical solutions" is missing.
Yeah, the third one is to not touch MTRRs after bootup and use PAT.
Works only in case the CPU has PAT support.
Which specific CPU without PAT support do you worry about?
Are these ones (below) good CPU to worry about? That's slightly off-topic as they are KVM guests, but still, the guest kernel seems to disable PAT (unless I'm mistaken).
KVM-86 + 2.6.27.19 guest:
# dmesg Initializing cgroup subsys cpuset Linux version 2.6.27.19-server-1mnb ([email protected]) (gcc version 4.3.2 (GCC) ) #1 SMP Thu Mar 5 00:09:04 EST 2009 PAT WC disabled due to known CPU erratum. (...)
# cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 2 model name : QEMU Virtual CPU version 0.10.50 stepping : 3 cpu MHz : 2133.285 cache size : 2048 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall lm up pni bogomips : 4266.57 clflush size : 64 power management:
KVM-85 + 2.6.29.1 guest:
# dmesg Initializing cgroup subsys cpuset Initializing cgroup subsys cpu Linux version 2.6.29.1-server-4mnb ([email protected]) (gcc version 4.3.2 (GCC) ) #1 SMP Mon Apr 20 20:10:24 EDT 2009 KERNEL supported cpus: Intel GenuineIntel AMD AuthenticAMD NSC Geode by NSC Cyrix CyrixInstead Centaur CentaurHauls Transmeta GenuineTMx86 Transmeta TransmetaCPU UMC UMC UMC UMC PAT WC disabled due to known CPU erratum. (...)
# cat /proc/cpuinfo processor : 0 vendor_id : GenuineIntel cpu family : 6 model : 2 model name : QEMU Virtual CPU version 0.10.0 stepping : 3 cpu MHz : 2128.004 cache size : 2048 KB fdiv_bug : no hlt_bug : no f00f_bug : no coma_bug : no fpu : yes fpu_exception : yes cpuid level : 2 wp : yes flags : fpu de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 syscall nx lm up pni hypervisor bogomips : 4256.00 clflush size : 64 power management:
Signature · 2 lines
Tomasz Chmielewski http://wpkg.org
Quoted 3 lines
Are these ones (below) good CPU to worry about? That's slightly off-topic as they are KVM guests, but still, the guest kernel seems to disable PAT (unless I'm mistaken).
To my knowledge the only CPUs with broken PAT are some very old Pentium Pros dating from back the time when PAT was new and noone used it. On those it should be disabled or the high bits not used.
The kernel is overall too conservative with its white list.
Today a lot of other common operating systems use PAT extensively, so in general it works on the hardware level.
-Andi
Signature · 1 line
[email protected] -- Speaking for myself only.
Quoted 4 lines
Quoted 1 line
Quoted 1 line
Andi Kleen <[email protected]> 25.05.09 11:15 >>>
To my knowledge the only CPUs with broken PAT are some very old Pentium Pros dating from back the time when PAT was new and noone used it. On those it should be disabled or the high bits not used.
??? See PentiumIII erratum E27.
Jan
Quoted 6 lines · Jan Beulich
Quoted 4 lines
Quoted 1 line
Quoted 1 line
Andi Kleen <[email protected]> 25.05.09 11:15 >>>
To my knowledge the only CPUs with broken PAT are some very old Pentium Pros dating from back the time when PAT was new and noone used it. On those it should be disabled or the high bits not used.
??? See PentiumIII erratum E27.
Thanks for the correction.
It's still all pretty old CPUs.
-Andi
Signature · 1 line
[email protected] -- Speaking for myself only.
Quoted 10 lines · Andi Kleen
To my knowledge the only CPUs with broken PAT are some very old Pentium Pros dating from back the time when PAT was new and noone used it. On those it should be disabled or the high bits not used.
The kernel is overall too conservative with its white list.
Today a lot of other common operating systems use PAT extensively, so in general it works on the hardware level.
Well, we don't use the high bits exactly for this reason.
-hpa
Signature · 2 lines
H. Peter Anvin, Intel Open Source Technology Center I work for Intel. I don't speak on their behalf.