Quote:
Originally Posted by binkyd
#1
Code:
0.004724] No NUMA configuration found
|
On some multi-core systems and most multiprocessor (multi-socket) systems, access to different memory regions is is tied to certain CPUs. For instance, if memory range A is only directly accessible by CPU 1, CPU 2 will need to use the memory lines of CPU 1 to access that region of memory.
This happens pretty much transparently, but there's a performance penalty when CPUs access memory tied to other CPUs. Hence, it is useful for the OS to know exactly how memory is partitioned when allocating tasks to CPU cores.
The "No NUMA configuration found" message may mean that the system in question simply isn't using NUMA, or that the Linux kernel isn't able to determine how it works. In the latter case, the system will not perform optimally.
Quote:
Originally Posted by binkyd
#2
Code:
0.346063] ACPI BIOS Error (bug): Could not resolve symbol [\_SB.PCI0.GPP8.VGA], AE_NOT_FOUND (20200326/dswload2-162)
[ 0.346063] fbcon: Taking over console
[ 0.346063] ACPI Error: AE_NOT_FOUND, During name lookup/catalog (20200326/psobject-220)
[ 0.346063] ACPI: Skipping parse of AML opcode: OpcodeName unavailable (0x0010)
[ 0.346063] ACPI BIOS Error (bug): Could not resolve symbol [\_SB.PCI0.GPP8.HDAU], AE_NOT_FOUND (20200326/dswload2-162)
|
Your system most likely has a buggy BIOS. This is surprisingly common.
The ACPI regions of the BIOS contain information written in a language called
AML, and any OS wanting to access ACPI functions needs to parse the information found in the various ACPI tables.
The error message says that the kernel AML parser is unable to make sense of some of the symbols in the ACPI table(s). As a result, any function described (probably incorrectly) by such a table entry will be inaccessible.
In some cases, the Linux kernel developers create workarounds to handle certain BIOS bugs, but evidently no workaround exists for your system. Unless there's an actual bug in the Linux AML parser, the only way to really fix this is a BIOS upgrade.
Quote:
Originally Posted by binkyd
#3
Code:
0.400982] pci 0000:02:00.1: can't claim BAR 6 [mem 0xfff80000-0xffffffff pref]: no compatible bridge window
[ 0.400984] pci 0000:09:00.0: can't claim BAR 6 [mem 0xfffe0000-0xffffffff pref]: no compatible bridge window
|
This means that there's a PCI/PCIe device that wants to use a certain memory range, but as far as the kernel knows, the region in question is outside what the PCI(e) bridge can access, so it just wouldn't work.
The BARs in question could be referencing expansion ROMs that are of no use to the OS, in which case the error message is just cosmetic. On the other hand, it could be that the BARs should have been moved to another area by the driver for the device in question, in which case this could indicate a drive bug. And yet again, it could be an ACPI bug affecting a motherboard resource of some kind.
It's impossible to be more exact without knowing exactly what PCI devices "0000:02:00.1" and "0000:09:00.0" are. You'll have to post the output of
lspci -v.
Quote:
Originally Posted by binkyd
#4
Code:
[0.709541] usb usb2: We don't know the algorithms for LPM for this host, disabling LPM.
|
The reference to "this host" means "this particular make/model USB host controller". In other words, it's a driver issue that may or may not have been addresses in a later kernel release.
"LPM" is an acronym for
Link Power Management, and message #5 probably appears as a consequence of this issue.
Quote:
Originally Posted by binkyd
#6
Code:
4.738241] tpm_crb MSFT0101:00: can't request region for resource [mem 0xdfb6f000-0xdfb72fff]
[ 4.738286] tpm_crb: probe of MSFT0101:00 failed with error -16
[ 4.743693] wmi_bus wmi_bus-PNP0C14:00: WQBC data block query control method not found
|
The kernel isn't able to initialize the Trusted Platform Module on your motherboard. Could be a driver bug, could be an(other) ACPI issue.
Quote:
Originally Posted by binkyd
#7
Code:
dell_smbios: No SMBIOS backends available (wmi: -19, smm: -19)
|
The kernel decoder for the System Management BIOS information on your system is unable to make sense of it.
I see this is a Dell system. That would explain all the BIOS bugs.
