James Bottomley wrote: > On Mon, 2005-04-25 at 08:46 +0900, Tejun Heo wrote: > >> If you're talking about scmd->eh_timeout, it's our main timer for >>normal command timeouts. If you're suggesting renaming it to something >>more apparant, I agree. Maybe just scmd->timeout will do. > > > Sorry ... actually on the ball now; I was assuming you simply wanted not > to use the field for efficiency. > > So, actually having read the description, you think that reusing the > eh_timeout in the error handler command submission path could confuse > the normal done routine if the host still has the command pending and > completes it? Hi, James. Sorry about late reply. Been busy and currently on the run, so please excuse me for being brief. * A command is passed to lldd and starts execution * It times out. * eh runs * abort isn't implemented or fails * eh issues eh cmd (TUL, STU...) * The command miraculously & stupidly completes just now. * The lldd succeeds to delete timer and normal completion path runs. * We're fucked up now. If anything is wrong, please point out. Thanks. Gotta go. -- tejun - : send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
