On Wed, 5 Jul 2023 at 10:25, Philippe Mathieu-Daudé <phi...@linaro.org> wrote:
>
> Cc'ing John.
>
> On 5/7/23 06:53, Richard Henderson wrote:
> > On 7/4/23 18:36, Peter Maydell wrote:
> >> +int main(int argc, char **argv)
> >> +{
> >> + const char *shm_name = "qemu-test-tcg-aarch64-icivau";
> >> + int fd;
> >> +
> >> + fd = shm_open(shm_name, O_CREAT | O_RDWR, S_IRUSR | S_IWUSR);
> >
> > Build failures:
> >
> > https://gitlab.com/qemu-project/qemu/-/jobs/4592433393#L3958
> > https://gitlab.com/qemu-project/qemu/-/jobs/4592433395#L4149
> > https://gitlab.com/qemu-project/qemu/-/jobs/4592433400#L3694
> >
> >
> > /usr/lib/gcc-cross/aarch64-linux-gnu/10/../../../../aarch64-linux-gnu/bin/ld:
> >
> > /usr/lib/gcc-cross/aarch64-linux-gnu/10/../../../../aarch64-linux-gnu/lib/../lib/librt.a(shm_open.o):
> > in function `shm_open':
> > (.text+0x3c): undefined reference to `__shm_directory'
> > /usr/lib/gcc-cross/aarch64-linux-gnu/10/../../../../aarch64-linux-gnu/bin/ld:
> > (.text+0xcc): undefined reference to `pthread_setcancelstate'
> > /usr/lib/gcc-cross/aarch64-linux-gnu/10/../../../../aarch64-linux-gnu/bin/ld:
> > (.text+0xfc): undefined reference to `pthread_setcancelstate'
> > /usr/lib/gcc-cross/aarch64-linux-gnu/10/../../../../aarch64-linux-gnu/bin/ld:
> >
> > /usr/lib/gcc-cross/aarch64-linux-gnu/10/../../../../aarch64-linux-gnu/lib/../lib/librt.a(shm_unlink.o):
> > in function `shm_unlink':
> > (.text+0x30): undefined reference to `__shm_directory'
> > collect2: error: ld returned 1 exit status
> > make[1]: *** [Makefile:119: icivau] Error 1
> > make[1]: *** Waiting for unfinished jobs....
> > make: *** [/builds/qemu-project/qemu/tests/Makefile.include:50:
> > build-tcg-tests-aarch64-linux-user] Error 2
> >
> > It looks like this test needs something else.
>
> Maybe:
>
> icivau: LDFLAGS+=-lrt -pthread
That is enough to get it to build, but then in the CI the test
consistently fails:
https://gitlab.com/pm215/qemu/-/jobs/4606447875
TEST icivau on aarch64
make[1]: *** [Makefile:178: run-icivau] Error 1
I'm going to drop this patch from the pullreq until we
can figure out what's going on...
thanks
-- PMM
