[<<] [<] Page 1 of 1 [>] [>>] | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subject:
GsSelect timeout
From: Allan Hessenflow ####@####.#### Date: 12 Aug 2006 03:14:02 +0100 Message-Id: <20060811191322.A23764@kallisti.com> Hello, How is the timeout passed to GsSelect() intended to be interpreted? From looking at some of the calls to it, I'm guessing: -1: do not wait 0: wait forever all other n: wait n milliseconds. If that's correct, the RTEMS port isn't handling it correctly (the timeout is definitely not being handled correctly somewhere - I think that it is meant to be as above which would mean that the RTEMS port is where the problem is). I just want to be sure of the correct interpretation before fixing it. allan -- Allan N. Hessenflow ####@####.#### | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subject:
Re: [nanogui] GsSelect timeout
From: "Greg Haerr" ####@####.#### Date: 12 Aug 2006 04:38:27 +0100 Message-Id: <096a01c6bdc0$af838380$6401a8c0@winXP> : -1: do not wait : 0: wait forever : all other n: wait n milliseconds. That's correct : : If that's correct, the RTEMS port isn't handling it correctly (the : timeout is definitely not being handled correctly somewhere - I think : that it is meant to be as above which would mean that the RTEMS port : is where the problem is). I just want to be sure of the correct : interpretation before fixing it. A quick looks shows the timeout value is passed unchanged to uid_read_message, FYI. I'm a bit surprised that this isn't working correctly, perhaps its just the -1L (poll) that's broken? Regards, Greg | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Subject:
Re: [nanogui] GsSelect timeout
From: Allan Hessenflow ####@####.#### Date: 12 Aug 2006 05:25:03 +0100 Message-Id: <20060811212426.A25144@kallisti.com> > : -1: do not wait > : 0: wait forever > : all other n: wait n milliseconds. > > That's correct > > > : > : If that's correct, the RTEMS port isn't handling it correctly (the > : timeout is definitely not being handled correctly somewhere - I think > : that it is meant to be as above which would mean that the RTEMS port > : is where the problem is). I just want to be sure of the correct > : interpretation before fixing it. > > A quick looks shows the timeout value is passed unchanged > to uid_read_message, FYI. I'm a bit surprised that this > isn't working correctly, perhaps its just the -1L (poll) > that's broken? -1 and 0 are both broken; inside uid_read_message the timeout is being interpreted as no wait for 0 and wait for a very long time for -1. It's fine for all other values. allan -- Allan N. Hessenflow ####@####.#### | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
[<<] [<] Page 1 of 1 [>] [>>] |