Skip to contentSkip to conversation

Heads up: next-20190115 and GCC 9.0 don't play nice.

DISCUSSION [email protected] 1 participant 2 messages Jan 26, 2019

lore
Share & tools
lore
#0 lore raw

So a GCC 9 escaped to Fedora Rawhide in the last few days, and things didn't go well. Fortunately, I had the 8.2.1-7 RPMs still around.

Issue 1: There's a new warning added for taking the address of a member of a packed array. It wasn't *too* noisy.

Issue 2: Looks like it's not ready for prime time.

CC [M] fs/ntfs/super.o
fs/ntfs/super.c: In function 'is_boot_sector_ntfs':
fs/ntfs/super.c:597:3: warning: converting a packed 'const NTFS_BOOT_SECTOR *' {aka 'const struct <anonymous> *'} pointer (alignment 1) to 'le32' {aka 'unsigned int'} (alignment 4) may result in an unaligned pointer value [-Waddress-of-packed-member]

  597 |   for (i = 0, u = (le32*)b; u < (le32*)(&b->checksum); ++u)
      |   ^~~

fs/ntfs/super.c:597:3: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://bugzilla.redhat.com/bugzilla> for instructions.
Preprocessed source stored into /tmp/cc0KDLoH.out file, please attach this to your bugreport. make[2]: *** [scripts/Makefile.build:277: fs/ntfs/super.o] Error 1
make[1]: *** [scripts/Makefile.build:491: fs/ntfs] Error 2
make: *** [Makefile:1044: fs] Error 2

I've reported this to the RedHat/Fedora crew as:
https://bugzilla.redhat.com/show_bug.cgi?id=1669696

#1 lore raw
Quoted 7 lines · On Fri, 25 Jan 2019 22:02:05 -0500, [email protected] said:

So a GCC 9 escaped to Fedora Rawhide in the last few days, and things didn't go well. Fortunately, I had the 8.2.1-7 RPMs still around.

Issue 1: There's a new warning added for taking the address of a member of a packed array. It wasn't *too* noisy.

Issue 2: Looks like it's not ready for prime time.

Jakub Jelinek informed me that gcc-9.0.1-0.1.fc30.x86_64 was out, and I've confirmed that *this* issue is fixed. Off to go see what breaks next. :)