Hi,
I'm on FreeBSD and have been having segfaults and SIGBUS sporadically when changing VTs, and trying to trace this down.
There are several of logs from GLIB on g_close; running in a debugger, I see that the file descriptor here is -1. I will probably file this with the BSD bugtracker, but was just wondering as I'm not familiar with the codebase, if this is indicative of a more serious problem that has origintated elsewhere, or if it's perhaps undesireable and an issue but the field could be considered "nullable" if somehting has gone wrong along the way.
Should a device's file descriptor always be non-negative, or is this an important assumption that shouldn't be broken? I see it'd checked one frame up here, and is reopened on the following line, but wanted to sanity-check before assuming I can ignore it and see what happens if I do that.
Thanks a lot!
Hi,
I'm on FreeBSD and have been having segfaults and SIGBUS sporadically when changing VTs, and trying to trace this down.
There are several of logs from GLIB on g_close; running in a debugger, I see that the file descriptor here is -1. I will probably file this with the BSD bugtracker, but was just wondering as I'm not familiar with the codebase, if this is indicative of a more serious problem that has origintated elsewhere, or if it's perhaps undesireable and an issue but the field could be considered "nullable" if somehting has gone wrong along the way.
Should a device's file descriptor always be non-negative, or is this an important assumption that shouldn't be broken? I see it'd checked one frame up here, and is reopened on the following line, but wanted to sanity-check before assuming I can ignore it and see what happens if I do that.
Thanks a lot!