[AMD Official Use Only] Hi Lucas, We can some changes we can see in lspci from working to non-working case. Below are changes Link Speed = 8GT/s -> 2.5GT/s. DLActive+ -> DLActive- BWMgmt+ -> BWMgmt+ PresDet+ -> PresDet+ EqualizationComplete+ -> EqualizationComplete+ Also when we do reset via sysfs, we don't see this issue. I have created bug here https://bugzilla.kernel.org/show_bug.cgi?id=215590 Thanks Rahul -----Original Message----- From: Lukas Wunner <lukas@xxxxxxxxx> Sent: Friday, February 11, 2022 3:08 AM To: Grodzovsky, Andrey <Andrey.Grodzovsky@xxxxxxx> Cc: linux-pci@xxxxxxxxxxxxxxx; helgaas@xxxxxxxxxx; Antonovitch, Anatoli <Anatoli.Antonovitch@xxxxxxx>; Kumar1, Rahul <Rahul.Kumar1@xxxxxxx>; Deucher, Alexander <Alexander.Deucher@xxxxxxx> Subject: Re: Question about deadlock between AER and pceihp interrupts during resume from S3 with unplugged device On Thu, Feb 10, 2022 at 03:47:10PM -0500, Andrey Grodzovsky wrote: > So the patches indeed helped resolving the deadlock but when we try > again to hotplug back there is a link status failure > > pcieport 0000:00:01.1: pciehp: Slot(0): Card present pcieport > 0000:00:01.1: Data Link Layer Link Active not set in 1000 msec > pcieport 0000:00:01.1: pciehp: Failed to check link status > > and more detailed bellow, > we are trying to debug but again, you might have a quick insight Well, the link doesn't come up. Is the Link Disable bit in the Link Control Register set for some reason? Perhaps some ACPI method fiddled with it? Compare the output of lspci -vv before and after the system sleep transition, do you see anything suspicious? If you reset the slot via sysfs, does the link come back up? You may want to open a bug over at bugzilla.kernel.org and attach the full dmesg output which didn't reach the list, as well as lspci output. Did you apply only my deadlock fix or also Kai-Heng Feng's AER disablement patch? Thanks, Lukas
