Xconfigurator defaults to X 3.3.6 for the following chipset, due to some
minor display corruption in X 4.0's "chips" driver. But if I wanted to, I
was always able to run X 4.0 with the following chipset. It worked fine,
except for some minor display corruption.
Starting with 4.1, X fails to start no matter what. I've tried multiple
resolutions, and depths, the end result is always the same - a segfault
after loading libramdac.a
3.3.6 in beta3 still works fine for this chipset. The problem is the 4.1
Created attachment 25532 [details]
XFree86.0.log of X 4.1.0 crashing after loading libramdac.a
Created attachment 25533 [details]
Created attachment 25534 [details]
A working XF86Config for 3.3.6
Created attachment 25535 [details]
Thanks for the detailed report mrsam, with all the attachments! I will
diff the two trees and see if there are any changes that might have
caused this regression. Thanks again for the complete report!
Well, the only thing I can get with gdb is this:
(II) Module ramdac: vendor="The XFree86 Project"
compiled for 4.1.0, module version = 0.1.0
(no debugging symbols found)...
Program received signal SIGSEGV, Segmentation fault.
0x084c4b5a in ?? () at eval.c:41
eval.c: No such file or directory.
#0 0x084c4b5a in ?? () at eval.c:41
#1 0x080c4463 in AddScreen () at eval.c:41
#2 0x0806e0c2 in InitOutput () at eval.c:41
#3 0x080c3a22 in main () at eval.c:41
#4 0x4007d18e in __libc_start_main (main=0x80c3750 <main>, argc=1,
ubp_av=0xbffffb04, init=0x806c2d8 <_init>, fini=0x817ebb0 <_fini>,
rtld_fini=0x4000cf08 <_dl_fini>, stack_end=0xbffffafc)
3.3.6 doesn't work any more in roswell2, because of an issue with xfs.
The xfs issue has been long since fixed.
libramdac.a still segfaults in enigma.
Created attachment 31698 [details]
libramdac.a appears to make some additional progress before segfaulting.
Lots of chips driver fixes went into XFree86 4.2.0. I recommend
trying out my RHL 7.2 build of 4.2.0 located at:
I do not have this hardware for testing, nor its documentation, however
the new release may fix it. Please update the bug report with the
This bug is fixed in 4.2 (other bugs from 4.0 are still there, though :-( ).