| 1 |
On 05/13/2018 10:39 PM, tuxic@××××××.de wrote: |
| 2 |
|
| 3 |
> Hi, |
| 4 |
> |
| 5 |
> from a previous thread I learned which micorcode file needs |
| 6 |
> to be loaded for the "fight against Spectre and Melddown"... |
| 7 |
> |
| 8 |
> Now two question came up for me: |
| 9 |
> 1.) WIth this particular micorcode loaded: Do I need to activate |
| 10 |
> the Spectre/Meldtown fix in the kernel itsself? |
| 11 |
> 2.) Which microcode file gets loaded? |
| 12 |
> |
| 13 |
> cat /proc/cpuinfo (from one of the six cores): |
| 14 |
> processor : 4 |
| 15 |
> vendor_id : AuthenticAMD |
| 16 |
> cpu family : 16 |
| 17 |
> model : 10 |
| 18 |
> model name : AMD Phenom(tm) II X6 1090T Processor |
| 19 |
> stepping : 0 |
| 20 |
> microcode : 0x10000bf |
| 21 |
> cpu MHz : 1755.530 |
| 22 |
> cache size : 512 KB |
| 23 |
> physical id : 0 |
| 24 |
> siblings : 6 |
| 25 |
> core id : 4 |
| 26 |
> cpu cores : 6 |
| 27 |
> apicid : 4 |
| 28 |
> initial apicid : 4 |
| 29 |
> fpu : yes |
| 30 |
> fpu_exception : yes |
| 31 |
> cpuid level : 6 |
| 32 |
> wp : yes |
| 33 |
> flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm 3dnowext 3dnow constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf pni monitor cx16 popcnt lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt cpb hw_pstate vmmcall npt lbrv svm_lock nrip_save pausefilter |
| 34 |
> bugs : tlb_mmatch apic_c1e fxsave_leak sysret_ss_attrs null_seg amd_e400 spectre_v1 spectre_v2 |
| 35 |
> bogomips : 7224.41 |
| 36 |
> TLB size : 1024 4K pages |
| 37 |
> clflush size : 64 |
| 38 |
> cache_alignment : 64 |
| 39 |
> address sizes : 48 bits physical, 48 bits virtual |
| 40 |
> power management: ts ttp tm stc 100mhzsteps hwpstate cpb |
| 41 |
> |
| 42 |
> |
| 43 |
> from dmesg (microcode-related lines): |
| 44 |
> [ 1.517265] microcode: CPU0: patch_level=0x010000bf |
| 45 |
> [ 1.517367] microcode: CPU1: patch_level=0x010000bf |
| 46 |
> [ 1.517450] microcode: CPU2: patch_level=0x010000bf |
| 47 |
> [ 1.517538] microcode: CPU3: patch_level=0x010000bf |
| 48 |
> [ 1.517628] microcode: CPU4: patch_level=0x010000bf |
| 49 |
> [ 1.517716] microcode: CPU5: patch_level=0x010000bf |
| 50 |
> [ 1.517824] microcode: Microcode Update Driver: v2.2. |
| 51 |
> |
| 52 |
> from linux-firmware (related lines): |
| 53 |
> /lib/firmware/amd-ucode/microcode_amd.bin |
| 54 |
> /lib/firmware/amd-ucode/microcode_amd_fam15h.bin |
| 55 |
> /lib/firmware/amd-ucode/microcode_amd.bin.asc |
| 56 |
> /lib/firmware/amd-ucode/microcode_amd_fam16h.bin *** |
| 57 |
> /lib/firmware/amd-ucode/microcode_amd_fam17h.bin |
| 58 |
> /lib/firmware/amd-ucode/microcode_amd_fam16h.bin.asc *** |
| 59 |
> /lib/firmware/amd-ucode/microcode_amd_fam15h.bin.asc |
| 60 |
> |
| 61 |
> (***) I think, these microcodes do I need... |
| 62 |
> |
| 63 |
> Thanks for any help in advance! |
| 64 |
> Cheers |
| 65 |
> Meino |
| 66 |
> |
| 67 |
> |
| 68 |
> |
| 69 |
> |
| 70 |
> |
| 71 |
. |
| 72 |
|
| 73 |
No Spectre/Meltdown fixes in these files, apparently. |
| 74 |
|
| 75 |
Simple experiment(s) to confirm the ucode file to use : ( repetitive, but simple ) |
| 76 |
|
| 77 |
Build a kernel with NO ucode, boot, note the patch_level. |
| 78 |
Build a kernel with microcode_amd.bin ucode ONLY, boot, note the patch_level. |
| 79 |
Build a kernel with microcode_amd_fam15h ucode ONLY, boot, note the patch_level. |
| 80 |
Build a kernel with microcode_amd_fam16h ucode ONLY, boot, note the patch_level. |
| 81 |
Build a kernel with microcode_amd_fam17h ucode ONLY, boot, note the patch_level. |
| 82 |
|
| 83 |
Results : |
| 84 |
When I did this on a Phenom II 980 x4, the only time the patch level changed was with "microcode_amd.bin" ucode. |
| 85 |
When I do this on a FX-9590, the only time the patch level changes is when "microcode_amd_fam15h.bin" ucode is used. |
| 86 |
|
| 87 |
Corbin |
Archives