On 11/7/19 9:51 AM, Petr Štetiar wrote: > Hauke Mehrtens <ha...@hauke-m.de> [2019-11-07 00:26:23]: > > Hi, > >>> I didn't looked deeper, but GCC 6,7,8,9 on x86/64 reports following: >>> >>> list.c:626:21: error: variable ‘prev’ might be clobbered by ‘longjmp’ or >>> ‘vfork’ [-Werror=clobbered] >>> 626 | struct uci_option *prev = NULL; >>> | ^~~~ >>> >>> list.c:627:14: error: variable ‘value2’ might be clobbered by ‘longjmp’ or >>> ‘vfork’ [-Werror=clobbered] >>> 627 | const char *value2 = NULL; >>> | ^~~~~~ >> >> I think I also saw these problems in my manual review, but the >> -Wclobbered did not complain about them for me with gcc 8.3 on MIPS. > > Interesting, out of curiosity which MIPS? BTW I see a consistent gcc 8.3 > behaviour[1] across all CI testing targets (ath79, malta be, imx6, armada > a-53) with -Wextra and without your patch applied. You need to click on the > "[x] Failed" button in order to see the build failure. > > 1. https://gitlab.com/ynezz/openwrt-uci/pipelines/94051766/builds
I also saw these and they are fixed in my patch by adding a volatile which is the correct fix for these problems. It was something else where I saw some potential problem in the review, but GCC was not warning about, as far as I know. How do we want to go forward with these patches? Hauke
signature.asc
Description: OpenPGP digital signature
_______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
