Created attachment 232653 [details] Call dev_pm_ops::prepare() from linux_pci_suspend() The dev_pm_ops' prepare() method should be called prior to its suspend() method, according to the Linux documentation: https://www.kernel.org/doc/html/v5.4/driver-api/pm/types.html#c.dev_pm_ops Without doing so, at least on my X1 Carbon 6th gen., Xorg usually (but not always) crashes after resume. I also suspect that the lacking of calling prepare() contributes to the hang on resume, infrequent but not rare, albeit no evidence. I have a couple of variants to call prepare() method at the time: * call prepare() via power_suspend_early eventhandler: stucks inside switch_to_kernel_context_sync() once per several trials. * (patch attached) call prepare() directly from linux_pci_suspend(): survives at least 10 trials, no failures yet.
Keyword: patch or patch-ready – in lieu of summary line prefix: [patch] * bulk change for the keyword * summary lines may be edited manually (not in bulk). Keyword descriptions and search interface: <https://bugs.freebsd.org/bugzilla/describekeywords.cgi>