suspend/resume kills usb0

Andy Green andy at openmoko.com
Sun Nov 2 23:13:20 CET 2008


-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Somebody in the thread at some point said:
| Am Sunday 02 November 2008 22:44:33 schrieb Andy Green:
|> Somebody in the thread at some point said:
|> | I'm logged in over usb0, in one session running a program that emits
|>
|> quite
|>
|> | some debug output via shell, in the other session running nothing. On
|>
|> resume,
|>
|> | the session with traffic quits.
|>
|> I guess it makes sense, initially and for some time the logical usb0
|> connection on the GTA02 is still down at resume.  Only when there is the
|> asynchronously completed USB negotiation with the host about enumeration
|> and so on done will the usb0 on the GTA02 exist again.
|>
|> So from the point of view of whatever is putting stuff out on stdout on
|> the GTA02, it probably gets some fatal "console has gone away" signal
|> sent to it if it tries to issue stuff during that window between resume
|> and usb0 getting recreated.  Whereas the silent session is unaware of
|> the whole period when it would have been death to issue something.
|
| Yes, that's sounds like an accurate analysis. Back to the original
question (I
| don't know too much about the console layer) -- is there a
straightforward
| way to fix that?

It sounds like the kind of thing that would respond to nohup or screen
type treatment.

- -Andy

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora - http://enigmail.mozdev.org

iEYEARECAAYFAkkOJfEACgkQOjLpvpq7dMq/RwCdHe9+j2C1vBGLl+d3tENj02iz
x00An3Au2Rkv8mQph/zgA4jP6SpyaNaC
=PGaI
-----END PGP SIGNATURE-----



More information about the openmoko-kernel mailing list