Re: [PATCH v2 0/2] PCI: Update quirk timing

From: Bjorn Helgaas
Date: Mon Mar 19 2018 - 14:51:30 EST


On Wed, Mar 14, 2018 at 08:30:26AM -0500, Bjorn Helgaas wrote:
> This is a minor update of a patch I posted earlier [1] when Paul
> reported an unreasonably long time spent in PCI quirks during boot
> [2]. This doesn't fix the slow quirks, but might help find them.
>
> We previously printed the time taken by each quirk if "initcall_debug"
> was specified. With these patches, we'll print it even without
> "initcall_debug" if the quirk takes longer than 10ms.
>
> I did notice one quirk (quirk_apple_poweroff_thunderbolt()) in the
> suspend path that sleeps for 300ms, so this will probably make suspend
> noisier on Apple systems.
>
> [1] https://lkml.kernel.org/r/20171228212723.GB211339@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
> [2] https://lkml.kernel.org/r/44cada166e42007d27b4c3e3aa0744d7@xxxxxxxxxxxxx
>
> ---
>
> Bjorn Helgaas (2):
> PCI: Report quirk timings with pci_info() instead of pr_debug()
> PCI: Report quirks that take more than 10ms
>
>
> drivers/pci/quirks.c | 24 ++++++++----------------
> 1 file changed, 8 insertions(+), 16 deletions(-)

Applied to pci/misc for v4.17.