Messages in this thread | | | Date | Mon, 17 Jun 2019 01:09:09 -0700 | From | Fenghua Yu <> | Subject | Re: [PATCH 1/3] x86/resctrl: Get max rmid and occupancy scale directly from CPUID instead of cpuinfo_x86 |
| |
On Mon, Jun 17, 2019 at 09:52:14AM +0200, Borislav Petkov wrote: > On Sun, Jun 16, 2019 at 08:18:09PM -0700, Fenghua Yu wrote: > > diff --git a/arch/x86/kernel/cpu/common.c b/arch/x86/kernel/cpu/common.c > > index 2c57fffebf9b..f080be35da41 100644 > > --- a/arch/x86/kernel/cpu/common.c > > +++ b/arch/x86/kernel/cpu/common.c > > @@ -801,6 +801,31 @@ static void init_speculation_control(struct cpuinfo_x86 *c) > > } > > } > > > > +static void get_cqm_info(struct cpuinfo_x86 *c) > > +{ > > + if (cpu_has(c, X86_FEATURE_CQM_LLC)) { > > + u32 eax, ebx, ecx, edx; > > + > > + /* QoS sub-leaf, EAX=0Fh, ECX=0 */ > > + cpuid_count(0x0000000F, 0, &eax, &ebx, &ecx, &edx); > > + /* will be overridden if occupancy monitoring exists */ > > + c->x86_cache_max_rmid = ebx; > > + > > + if (cpu_has(c, X86_FEATURE_CQM_OCCUP_LLC) || > > + cpu_has(c, X86_FEATURE_CQM_MBM_TOTAL) || > > + cpu_has(c, X86_FEATURE_CQM_MBM_LOCAL)) { > > + /* QoS sub-leaf, EAX=0Fh, ECX=1 */ > > + cpuid_count(0x0000000F, 1, &eax, &ebx, &ecx, &edx); > > + > > + c->x86_cache_max_rmid = ecx; > > + c->x86_cache_occ_scale = ebx; > > + } > > + } else { > > + c->x86_cache_max_rmid = -1; > > + c->x86_cache_occ_scale = -1; > > + } > > +} > > + > > void get_cpu_cap(struct cpuinfo_x86 *c) > > { > > u32 eax, ebx, ecx, edx; > > @@ -832,33 +857,6 @@ void get_cpu_cap(struct cpuinfo_x86 *c) > > c->x86_capability[CPUID_D_1_EAX] = eax; > > } > > > > - /* Additional Intel-defined flags: level 0x0000000F */ > > - if (c->cpuid_level >= 0x0000000F) { > > - > > - /* QoS sub-leaf, EAX=0Fh, ECX=0 */ > > - cpuid_count(0x0000000F, 0, &eax, &ebx, &ecx, &edx); > > - c->x86_capability[CPUID_F_0_EDX] = edx; > > - > > - if (cpu_has(c, X86_FEATURE_CQM_LLC)) { > > - /* will be overridden if occupancy monitoring exists */ > > - c->x86_cache_max_rmid = ebx; > > - > > - /* QoS sub-leaf, EAX=0Fh, ECX=1 */ > > - cpuid_count(0x0000000F, 1, &eax, &ebx, &ecx, &edx); > > - c->x86_capability[CPUID_F_1_EDX] = edx; > > - > > - if ((cpu_has(c, X86_FEATURE_CQM_OCCUP_LLC)) || > > - ((cpu_has(c, X86_FEATURE_CQM_MBM_TOTAL)) || > > - (cpu_has(c, X86_FEATURE_CQM_MBM_LOCAL)))) { > > - c->x86_cache_max_rmid = ecx; > > - c->x86_cache_occ_scale = ebx; > > - } > > - } else { > > - c->x86_cache_max_rmid = -1; > > - c->x86_cache_occ_scale = -1; > > - } > > - } > > Why are you doing this carving out into a separate function since you're > keeping the cpuinfo_x86 members?
I just keep the code a bit uniform around the calling area where a few functions are called. So get_cqm_info() makes the code a bit more readable.
init_scattered_cpuid_features(c); init_speculation_control(c); + get_cqm_info(c);
/* * Clear/Set all flags overridden by options, after probe. * This needs to happen each time we re-probe, which may happen * several times during CPU initialization. */ apply_forced_caps(c); }
Maybe not? If the function is not good, I can directly put the code here?
Thanks.
-Fenghua
| |