> On Fri, Feb 27, 2009 at 10:40 AM, Alexander Beregalov > <a.beregalov@xxxxxxxxx> wrote: > > The same problem: ld segfaults: > > do_page_fault() pid=15739 command='ld' type=15 address=0x00000c84 > > This is likely a bug in the static linker, or a miscompilation of the > static linker by your toolchain. If the problem is fully, reproducable, then I agree its likely a toolchain issue. However, as indicated multiple times on the list, we get random segfaults in various programs. In my experience, this occurs more frequently with kernels after 2.6.22 but I'm sure 2.6.22 is affected. I've tried many kernels after 2.6.22 but I haven't found one that I like. 2.6.28 doesn't crash on PA 8800 (SMP 64-bit), but I get too many random segfaults doing GCC builds. The extract/deposit/shift bug that Helge found recently looks important as it causes undefined behavior that appears to vary from processor to processor. A patch to check for this was recently applied to GAS. I was hoping to try Helge's recent changes but I don't think they are in Kyle's parisc tree. If the problem is reproducible, try running gdb to get a backtrace (e.g., gdb -c core /usr/bin/ld). I am fairly certain the binutils testsuite for ld 2.19.1 is clean. Dave -- J. David Anglin dave.anglin@xxxxxxxxxxxxxx National Research Council of Canada (613) 990-0752 (FAX: 952-6602) -- To unsubscribe from this list: send the line "unsubscribe linux-parisc" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
