Forget to mention one small detail: On 15.01.23 15:21, Linux kernel regression tracking (Thorsten Leemhuis) wrote: > [CCing the regression list, as it should be in the loop for regressions: > https://docs.kernel.org/admin-guide/reporting-regressions.html] > > Hi, this is your Linux kernel regression tracker. > > On 15.01.23 10:21, Peifeng Qiu wrote: >> >> I'm using a dual Xeon system with Intel e810 25G network card and make use >> of SRIOV feature heavily. I have a script to setup the NIC the first step is >> echo $VFS > /sys/class/net/$DEVNAME/device/sriov_numvfs >> >> After switching from v6.1 to v6.2-rc1 "sriov_numvfs" is no longer present. If I >> switch back to v6.1 it's back. Command line parameters are the same so it's >> most likely kernel changes. I did git bisect and found the culprit to be >> 07eab0901ed(efi/x86: Remove EfiMemoryMappedIO from E820 map) > > This is not my area of expertise, but there is another report about an > issue that was bisected to that particular commit: > > https://lore.kernel.org/lkml/ac2693d8-8ba3-72e0-5b66-b3ae008d539d@xxxxxxxxxxxxxxx/ > > A fix for that one was posted here: > > https://lore.kernel.org/lkml/20230110180243.1590045-1-helgaas@xxxxxxxxxx/ > > You might want to look into the report and if it looks like a possible > duplicate give the proposed fix a try. FWIW, that patch was recently mainlined and thus will be in 6.2-rc4 that Linus will likely release in a few hours. Ciao, Thorsten > But as I said, this is not my area of expertise, you hence might want to > wait for the expert to share their view of things. > > Anyway, to be sure the issue doesn't fall through the cracks unnoticed, > I'm adding it to regzbot, the Linux kernel regression tracking bot: > > #regzbot ^introduced 07eab0901ed > #regzbot title pci: sriov_numvfs missing after removal of > EfiMemoryMappedIO from E820 map (possible duplicate) > #regzbot ignore-activity > > Ciao, Thorsten (wearing his 'the Linux kernel's regression tracker' hat) > -- > Everything you wanna know about Linux kernel regression tracking: > https://linux-regtracking.leemhuis.info/about/#tldr > If I did something stupid, please tell me, as explained on that page. > >> I tested v6.2-rc3 and it's the same. I reverted this commit on top of v6.2-rc3 >> then sriov_numvfs is back again. Comparing the dmesg output, I found that >> with commit 07eab0901ed these lines are present: >> [ 0.000000] efi: Remove mem94: MMIO range=[0x80000000-0x8fffffff] >> (256MB) from e820 map >> [ 0.000000] e820: remove [mem 0x80000000-0x8fffffff] reserved >> [ 0.000000] efi: Remove mem95: MMIO range=[0xfd000000-0xfe7fffff] >> (24MB) from e820 map >> [ 0.000000] e820: remove [mem 0xfd000000-0xfe7fffff] reserved >> [ 0.000000] efi: Not removing mem96: MMIO >> range=[0xfed20000-0xfed44fff] (148KB) from e820 map >> [ 0.000000] efi: Remove mem97: MMIO range=[0xff000000-0xffffffff] >> (16MB) from e820 map >> [ 0.000000] e820: remove [mem 0xff000000-0xffffffff] reserved >> [ 0.000000] efi: Remove mem99: MMIO >> range=[0x1ffc00000000-0x1fffffffffff] (16384MB) from e820 map >> [ 0.000000] e820: remove [mem 0x1ffc00000000-0x1fffffffffff] reserved >> >> I think that's what the commit actually does. But the following are missing: >> [ 2.516119] pci 0000:ca:00.0: reg 0x184: [mem >> 0x208ffd000000-0x208ffd01ffff 64bit pref] >> [ 2.516121] pci 0000:ca:00.0: VF(n) BAR0 space: [mem >> 0x208ffd000000-0x208ffdffffff 64bit pref] (contains BAR0 for 128 VFs) >> [ 2.516134] pci 0000:ca:00.0: reg 0x190: [mem >> 0x208ffe220000-0x208ffe223fff 64bit pref] >> [ 2.516136] pci 0000:ca:00.0: VF(n) BAR3 space: [mem >> 0x208ffe220000-0x208ffe41ffff 64bit pref] (contains BAR3 for 128 VFs) >> >> Not sure whether this is a driver issue specific to Intel e810(module ice) or >> a more general one. Any thoughts on this issue? >> >> Best regards, >> Peifeng Qiu