On Thu, Oct 14, 2010 at 12:19:49PM -0700, Adam Jackson wrote: > On Thu, 2010-10-14 at 12:12 -0700, Aaron Plattner wrote: > > On Thu, Oct 14, 2010 at 09:49:05AM -0700, Adam Jackson wrote: > > > Signed-off-by: Adam Jackson <[email protected]> > > > --- > > > hw/xfree86/os-support/shared/sigio.c | 11 ----------- > > > hw/xfree86/os-support/shared/sigiostubs.c | 5 ----- > > > hw/xfree86/os-support/xf86_OSproc.h | 1 - > > > 3 files changed, 0 insertions(+), 17 deletions(-) > > > > We use these asserts in the nvidia driver to make sure that SIGIO is > > properly blocked around code that touches the display engine. It would be > > a shame to lose this, but we could roll our own checks if you really think > > these need to be removed. > > Fair enough, though I'm curious what that's caught.
Due to quirks in our hardware, we need to prevent cursor changes during modesets. The asserts catch internal bugs where we don't do that. That's why I said we can roll our own if we have to, we just have to then identify exactly which pieces of code can be called from the signal handler. Speaking of that, it might be nice to have a runtime flag that indicates whether the "conservative" SIGIO handler is enabled. -- Aaron _______________________________________________ [email protected]: X.Org development Archives: http://lists.x.org/archives/xorg-devel Info: http://lists.x.org/mailman/listinfo/xorg-devel
