Re: [PATCH v3 1/2] PCI: Ignore Immediate Readiness on Samsung 990 PRO NVMe

[Alexander Gruhlke]

On Sun, Oct 11, 2026 at 07:14:44PM +0000, sashiko-bot@xxxxxxxxxx wrote:
> Can this result in a C bitfield data race where either the imm_ready = 0
> update or the sysfs attribute update is silently lost, potentially breaking
> the VFIO passthrough (the bug this patch intended to fix) or corrupting PCI
> power management state?

imm_ready is set in pci_init_capabilities(), after the header fixups, so
a final fixup is the earliest place a quirk can clear it.  At boot, final
fixups run before userspace starts, and other final quirks such as
quirk_no_msi() and quirk_broken_intx_masking() set pci_dev bitfields the
same way.  So I'd leave it as is.

Thanks,
Alexander