|
From: Josef W. <Jos...@gm...> - 2005-02-01 14:11:31
|
On Tuesday 01 February 2005 11:34, you wrote: > On Tue, 2005-02-01 at 10:17 +0000, Julian Seward wrote: > > On Monday 31 January 2005 12:13, Ross Burton wrote: > > > I'm trying to use callgrind with Xephyr, a re-implementation of XNest > > > using the kdrive X server (resulting in a xnest-like tool which supports > > > Render, RANDR, etc). However, when I start it I just get this error: > > > > > > valgrind --tool=callgrind Xephyr -ac :1 > > > [...] > > > XIO: fatal IO error 4 (Interrupted system call) on X server ":0.0" > > > after 99 requests (99 known processed) with 0 events remaining. > > I'm afraid I don't know what the problem is. > > > > Does it happen if you use --tool=none ? What about if you use > > --tool=memcheck ? > > With none, I get exactly the same behaviour. With memcheck I get two > jump depends on uninitalised values warnings before the crash. These > may be the problem but they are concerned with the actual drawing > routines, so probably not. > > This is an interesting mail we found yesterday: > > "valgrind may help here; be ware that valgrind only works with Xfbdev, > and then only if you disable the SIGIO-based mouse code." > http://lists.freedesktop.org/pipermail/xserver/2004-March/001086.html > > I can confirm that it crashes the moment I move the mouse, so is this a > limitation of the signal handling in valgrind? Sorry, no idea. Can you provide a log with --trace-signals=yes? Josef |