I just noticed this message in the logs running 4.4.11 pmd_set_huge: Cannot satisfy [mem 0xf8000000-0xf8200000] with a huge-page mapping due to MTRR override. I did not see this in 4.1.21, nor in any of the 4.1x LTS branch... Something new seems to of been introduced into 4.4x LTS? Can anyone explain what is going on here, bug, harmless message, etc.? If anyone needs to see some logs, please let me know what you need and I'll upload...
I would greatly appreciate it if someone can take a few moments to reply to this, and give some insight here? Also if anyone knows about this in the Kernel DEV Team and if any kernels 4.5x upwards have been resolved of this? thank you
I found a solution; enable_mtrr_cleanup=1
(In reply to KernelBug from comment #2) > I found a solution; > > enable_mtrr_cleanup=1 I'd like to be able to read about some background of this solution, because I see the same problem with kernel 4.4.79-18.26-default from openSUSE 42.2: Sep 10 16:34:51 i7 kernel: BIOS-e820: [mem 0x00000000f8000000-0x00000000fbffffff] reserved [...] Sep 10 16:34:51 i7 kernel: MTRR fixed ranges enabled: [...] Sep 10 16:34:51 i7 kernel: E8000-FFFFF write-protect [...] Sep 10 16:34:51 i7 kernel: PCI: MMCONFIG at [mem 0xf8000000-0xfbffffff] reserved in E820 Sep 10 16:34:51 i7 kernel: pmd_set_huge: Cannot satisfy [mem 0xf8000000-0xf8200000] with a huge-page mapping due to MTRR override.
Problem happens on HP Spectre with 4.16.0 unless enable_mtrr_cleanup=1 is used.
Created attachment 298957 [details] dmseg log of Fedora 34 which has 5.13 kernel saying, "pmd_set_huge: Cannot satisfy [mem 0xe0000000-0xe0200000] with a huge-page mapping due to MTRR override." 314 - 319 0.017603] acpiphp: ACPI Hot Plug PCI Controller Driver version: 0.5 [ 0.018076] PCI: MMCONFIG for domain 0000 [bus 00-ff] at [mem 0xe0000000-0xefffffff] (base 0xe0000000) [ 0.018081] PCI: MMCONFIG at [mem 0xe0000000-0xefffffff] reserved in E820 [ 0.018088] pmd_set_huge: Cannot satisfy [mem 0xe0000000-0xe0200000] with a huge-page mapping due to MTRR override. [ 0.018449] PCI: Using configuration type 1 for base access [ 0.018709] ENERGY_PERF_BIAS: Set to 'normal', was 'performance' Is there some workaround toward this issue!
https://web.archive.org/web/20190904223631/http://my-fuzzy-logic.de/blog/index.php?/archives/41-Solving-linux-MTRR-problems.html