On 09/28/2012 09:31 AM, Maurizio Lombardi wrote:
> some days ago I noticed that the HelenOS boot process hangs on my Sony
> VAIO (amd64) immediatly after launching devman, the kernel console is
> still responsive.
> 
> I decided to ran bzr bisect to find out when the bug has been introduced.
> Apparently, something went wrong when Vojtech merged the mainline
> changes into his own "logging" branch at revision 1537.1.51, but:
> 
> - the previous revision, 1537.1.50 works perfectly.
> - the revision coming from the mainline, 1624, works as well.
> - I tried to checkout revision 1537.1.50 and do the merge with
> revision 1624 by myself (just to be sure Vojtech did the merge
> correctly)  but it does not boot.
> 
> Looks like the bug is not reproducible with qemu, Jan Vesely was
> succesful to reproduce it on his Aspire One (ia32).
> Is there someone else who is able to reproduce it?

Yes, I can reproduce it as well on my laptop based on your recipe:

- amd64 build with ext4fs root fs

kconsole shows devman has a connecting phone to ns which is not
responding to its IPC_M_CONNECT_ME_TO call for some reason. In fact, ns
is not responding to the preceding 1025 call from the init task either.

This issue has also revealed another issue with killing a task which has
an unanswered IPC_M_CONNECT_ME_TO and a connecting phone. Despite all my
beliefs, such a task cannot be killed as of now. I will try to fix this.

Jakub

_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/cgi-bin/listinfo/helenos-devel

Reply via email to