Hello, On Sat, Apr 18, 2020 at 03:18:35PM +0200, Uwe Kleine-König wrote: > On Sat, Apr 18, 2020 at 01:22:26PM +0200, Uwe Kleine-König wrote: > > On Fri, Apr 17, 2020 at 11:15:28AM +0200, Uwe Kleine-König wrote: > > > Control: found -1 5.5.13-2 > > > > > > On Thu, Jan 16, 2020 at 08:39:44AM +0100, Lennert Van Alboom wrote: > > > > Package: src:linux > > > > Version: 5.5~rc5-1~exp1 > > > > Severity: normal > > > > > > > > Neither shutdown nor suspend works in this kernel version. Systemd seems > > > > to indicate it is trying at least, but the system hangs either right > > > > before shutdown (clean situation, requiring a poweroff with the power > > > > button) or before sleep (hung system, can't do anything except also hard > > > > powering off and rebooting later). > > > > > > > > Syslog from a suspend action: > > > > > > > > Jan 16 06:43:53 Nesbitt NetworkManager[749]: <info> [1579153433.6342] manager: sleep: sleep requested (sleeping: no enabled: yes) > > > > Jan 16 06:43:53 Nesbitt NetworkManager[749]: <info> [1579153433.6343] device (p2p-dev-wlan0): state change: disconnected -> unmanaged (reason 'sleeping', sys-iface-state: 'managed') > > > > Jan 16 06:43:53 Nesbitt NetworkManager[749]: <info> [1579153433.6345] manager: NetworkManager state is now ASLEEP > > > > Jan 16 06:43:53 Nesbitt systemd[1]: Reached target Sleep. > > > > Jan 16 06:43:53 Nesbitt systemd[1]: Starting Suspend... > > > > Jan 16 06:43:53 Nesbitt kernel: [55203.294410] PM: suspend entry (s2idle) > > > > Jan 16 06:43:53 Nesbitt systemd-sleep[19110]: Suspending system... > > > > > > I have the same problem with 5.5.13-2 on a Lenovo Thinkpad T460p, with > > > 4.19 from buster there is no such problem. I didn't debug any further > > > yet. > > > > I just reconfirmed: linux-image-4.19.0-8-amd64 4.19.98-1 doesn't suffer > > from this problem, but linux-image-5.6.0-trunk-amd64-unsigned > > 5.6.4-1~exp1 has the same problem. > > I started bisecting (but didn't complete yet, the weather is too good to > sit inside :-). Currently I see > > bad: 5.2.17-1+b1 > good: 5.2.6-1 > I checked the remaining versions in between the two above and the bisect result is that 5.2.7-1 is the last good; and 5.2.9-1 is the first bad kernel image. The relevant change between these two is: * [x86] iommu: Enable INTEL_IOMMU_DEFAULT_ON (Closes: #934309) . At least 5.2.7-1 shows the problematic behaviour when I add "intel_iommu=on" on the kernel command line. I suspect I can boot newer kernels with "intel_iommu=off", but didn't try that yet. Related: https://bugzilla.kernel.org/show_bug.cgi?id=197029 @Lennert: I assume you can still reproduce the problem. Do you care to test with intel_iommu=off and report about the result? Just to make sure you and I see the same problems? Best regards Uwe
Attachment:
signature.asc
Description: PGP signature