I've been suffering from laptop (dell xps 13 w/ skylake) freezes after it shuts down the screen due to inactivity. I call that frezes because when it happens there is no way to wake the screen up. SSH is down. ctrl-alt-f1 and bling-typing is not working. suspend button is not responding. s2ram command is ignorant too. And since I do not have a screen (a laptop, remember?) I try all these things blindly.
So the laptop basically stops responding. Each time (3-5 occurences this year) post mortem analysis reveals that one of last syslog entries is:
syslog.1:9316:Sep 20 20:51:56 netikras-xps kernel: [477714.359255] [drm:intel_pipe_update_end [i915_bpo]] *ERROR* Atomic update failure on pipe A (start=330999 end=331000) time 147 us, min 1788, max 1799, scanline start 1785, end 1801
As you see from the log attached this error occurs rather often. Maybe it's a coincidence, maybe it's not. But I have surely noticed that this error is always one of the last syslog records ante mortem.
Ping me if anyone needs any more details, I'm subscribing to this thread
I'll spill some gasoline on this fire.
I've been suffering from laptop (dell xps 13 w/ skylake) freezes after it shuts down the screen due to inactivity. I call that frezes because when it happens there is no way to wake the screen up. SSH is down. ctrl-alt-f1 and bling-typing is not working. suspend button is not responding. s2ram command is ignorant too. And since I do not have a screen (a laptop, remember?) I try all these things blindly.
So the laptop basically stops responding. Each time (3-5 occurences this year) post mortem analysis reveals that one of last syslog entries is:
syslog.1:9316:Sep 20 20:51:56 netikras-xps kernel: [477714.359255] [drm:intel_ pipe_update_ end [i915_bpo]] *ERROR* Atomic update failure on pipe A (start=330999 end=331000) time 147 us, min 1788, max 1799, scanline start 1785, end 1801
As you see from the log attached this error occurs rather often. Maybe it's a coincidence, maybe it's not. But I have surely noticed that this error is always one of the last syslog records ante mortem.
Ping me if anyone needs any more details, I'm subscribing to this thread