Opened 3 years ago
Closed 3 years ago
#17365 closed bug (fixed)
Boot hangs on rocket in QEMU
Reported by: | bitigchi | Owned by: | waddlesplash |
---|---|---|---|
Priority: | critical | Milestone: | R1/beta4 |
Component: | Servers/launch_daemon | Version: | R1/Development |
Keywords: | Cc: | axeld | |
Blocked By: | Blocking: | ||
Platform: | All |
Description
I am getting the following error on latest nightly after the rocket icon:
qemu-x86_64-softmmu: warning: global PIIX4_PM.disable_s3=1 not used
It stopped working around two weeks ago, before that I am not able to give an estimate since I have been a bit busy.
Attachments (1)
Change History (14)
comment:1 by , 3 years ago
comment:2 by , 3 years ago
comment:3 by , 3 years ago
Tried with hrev55602. Still the same error, nothing evident in the new syslog as well.
comment:4 by , 3 years ago
Version: | R1/beta3 → R1/Development |
---|
comment:5 by , 3 years ago
qemu-x86_64-softmmu: warning: Spice: record:0 (0x7f868b86ea00): setsockopt failed, Operation not supported on socket
No obvious error on the syslog.
comment:6 by , 3 years ago
QEMU boots occasionally fail randomly for me, too, but only 1 in 10 times or so. I haven't yet investigated that.
comment:7 by , 3 years ago
Milestone: | Unscheduled → R1/beta4 |
---|---|
Priority: | normal → critical |
Seems to happen more than 1 in 10 times for me actually.
comment:8 by , 3 years ago
I’ve managed to boot using a fresh QEMU config, but this time Deskbar fails to load every other reboot. I have to launch Team Monitor to be able to relaunch. Shall I open a new ticket for that?
comment:10 by , 3 years ago
Component: | - General → Servers/launch_daemon |
---|---|
Owner: | changed from | to
Summary: | Boot error on QEMU → Boot hangs on rocket in QEMU |
This seems to be much easier to reproduce with the Storage Kit compiled in debug mode...
app_server is started as per my logs, so it seems the real culprit is probably once again the launch_daemon failing to start the desktop.
comment:11 by , 3 years ago
Owner: | changed from | to
---|---|
Status: | new → in-progress |
The problem is specifically that the launch_daemon is starting the user session before the initial_volumes_mounted event is registered. This should not matter as the messages should be forwarded to the user launch_daemon anyway, but for some reason it seems they are not.
comment:12 by , 3 years ago
Cc: | added |
---|
Proposed fix: https://review.haiku-os.org/c/haiku/+/4968
CC'ing axeld just for good measure on this, too. :)
Please post a full syslog (obtained via
-serial stdout
on QEMU should work.)