This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: PR debug/17406: ICE dwarf2out_frame_debug_expr on MinGW


Aaron W. LaFramboise wrote:
Could somebody give me some quick information about the prognosis of
this bug?

We rely on volunteer labor to fix bugs, so there is no way for us to predict when it might be fixed. Also, the tree is rather unsetttled at the moment, as there are multiple bugs causing bootstrap problems on multiple targets.


In addition, as who is still very much a GCC beginner, do I have any
chance at all of being able to fix this myself?

This isn't clear from the info in the PR. If you want to try taking a look, it wouldn't hurt.


You can pretty print rtl by using some gdb macros, though should be automatically loaded from the .gdbinit file. So you can do
(gdb) print dest
(gdb) pr
to see what the RTL for dest is. You probably want expr, dest, and src here.


From the line number, it looks like it hit the gcc_unreachable line at the end, but you seem to have a MEM, and there is a MEM case, so it isn't clear why this happened.

Try putting a breakpoint in the dwarf2out_frame_debug_expr function and step through it to see what is happening.

You can generate a testcase by using --save-temps which will give you a .i file to use as input to cc1.

If the function is too hard to debug, you might try rebuilding dwarf2out.c with -O1 or -O0. The default -O2 sometimes makes debugging too difficult to be useful.
--
Jim Wilson, GNU Tools Support, http://www.SpecifixInc.com



Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]