From: Jeff Dike <jdike@ka...> - 2002-03-28 23:09:44
The major theme of this patch is the fix for the console flow control bug.
This involved various things such as:
a boot-time test for support for output SIGIO in the host pty driver
when it's not there, emulation of output SIGIOs is enabled
the drivers themselves implement flow control by buffering whatever
they can't get out immediately and having write_room return the amount of
space remaining in the buffer.
It seems to work fairly well, but let me know if there are problems.
I also killed a couple more nasty SIGWINCH errors, fixed the NULL tty
dereference that popped up here, and fixed the kmalloc bug that Matt