Hi,
On Sun, Oct 18, 2026 at 11:21:50PM +0200, Linux-Fan wrote:
> Hello,
>
> there is newly an upstream solution to the issue:
>
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=4766082b791b515953d87f7a24c2ea857e6c10b0
>
> https://lore.kernel.org/all/[email protected]/
>
> The two patches from linux/master apply as-is on the current Debian stable
> Linux kernel 6.12.113 and I am now running such a variant locally to bypass
> the issue.
For completeness, the single commits applied in mainline are:
e8df1d194c39 ("PCI: Don't treat ~0 as ready in pci_dev_wait() with RRS SV")
0f8b5d90d179 ("PCI: Ignore Immediate Readiness on Samsung 990 PRO and 990 EVO
Plus NVMe")
which is in v7.3.
Both landed as well v7.2.13 upstream.
e8df1d194c39 ("PCI: Don't treat ~0 as ready in pci_dev_wait() with RRS
SV") is queued as well for 6.12.y but the 0f8b5d90d179 ("PCI: Ignore
Immediate Readiness on Samsung 990 PRO and 990 EVO Plus NVMe") is not,
where the latter is said to fix a regression in v7.2-rc1.
So things should be good with the queued version.
Regards,
Salvatore
