Bug#804002: libxcb FTBFS on mips: cc1 segfaults building xkb.c

control: reassign -1 gcc-5
control: affects -1 libxcb
control: forwarded -1 https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68302

On 2015-11-06 07:52, Helmut Grohne wrote:
> On Wed, Nov 04, 2015 at 06:26:01AM +0100, Helmut Grohne wrote:
> > | libtool: compile:  gcc -DHAVE_CONFIG_H -I. -I../../src -D_FORTIFY_SOURCE=2 -Wall -Wpointer-arith -Wmissing-declarations -Wformat=2 -Wstrict-prototypes -Wmissing-prototypes -Wnested-externs -Wbad-function-cast -Wold-style-definition -Wdeclaration-after-statement -Wunused -Wuninitialized -Wshadow -Wmissing-noreturn -Wmissing-format-attribute -Wredundant-decls -Wlogical-op -Werror=implicit -Werror=nonnull -Werror=init-self -Werror=main -Werror=missing-braces -Werror=sequence-point -Werror=return-type -Werror=trigraphs -Werror=array-bounds -Werror=write-strings -Werror=address -Werror=int-to-pointer-cast -Werror=pointer-to-int-cast -g -O2 -fstack-protector-strong -Wformat -Werror=format-security -c xkb.c -o xkb.o
> > | gcc: internal compiler error: Segmentation fault (program cc1)
> > | Please submit a full bug report,
> > | with preprocessed source if appropriate.
> > | See <file:///usr/share/doc/gcc-5/README.Bugs> for instructions.
> > | Makefile:996: recipe for target 'xkb.lo' failed
> > 
> > The attached file xkb.pre.c is the result of running the same invocation
> > past the preprocessor. Running gcc -g -O2 xkb.pre.c also gives that
> > segfault. Dropping either -g or -O2 or using -O1 or -O3 makes it go
> > away. Thus I cc [email protected]. This is gcc-5 5.2.1-23.
> 
> I am attaching a reduced file crash.c that reproduces the crash with
> just 28 lines and no includes. It seems that the bug is in the code
> generator or optimizer, because we can only observe a crash for mips and
> subtle changes make the issue disappear. Judging from a quick glance at
> a traceback (w/o debug symbols), the issue at hand is a stack space
> exhaustion of two mutually recursive functions. That's all I know for
> now.

Thanks for the reduced testcase. The problem is reproducible with gcc
trunk from today on both mips and mipsel. One need to add
-march=mips32r2 to reproduce the issue on mipsel as it is not (yet)
the default there.

I have reported the bug upstream, and I am re-affecting this bug to
gcc-5 as the bug is actually there.

Aurelien