On Thu, Feb 9, 2017 at 1:46 PM, Dmitry Vyukov <dvyu...@google.com> wrote: > On Thu, Feb 9, 2017 at 12:55 PM, Arnd Bergmann <a...@arndb.de> wrote: >> On Thu, Feb 9, 2017 at 12:28 PM, Dmitry Vyukov <dvyu...@google.com> wrote: >>> On Wed, Feb 8, 2017 at 10:18 PM, Arnd Bergmann <a...@arndb.de> wrote: >> >> Thanks for the list, that looks very helpful. The ones I found myself >> seem to be a strict (and small) subset of those, using gcc-7.0.1 on x86-64 >> with allmodconfig and a few hundred randconfig builds. Which compiler >> version did you use for your testing? If new versions are better than old >> ones, we could start by fixing the ones that are still present in gcc-6 and >> gcc-7, and making the warning conditionally on the compiler version. >> >> Another idea might be to separate out asan_stack=1 into a separate >> Kconfig option and warn if that is enabled with compilers that are known >> to be relatively bad it keeping the stack small. > > > Mine is gcc version 7.0.0 20161208. Make sure you enable KASAN_INLINE > and I also enabled CONFIG_KCOV. Other than that I just did > allyesconfig + make -k.
Ok, I see my mistake now: On the allmodconfig build, I had KASAN_INLINE disabled, which made the problem go away for almost all files (almost all frame sizes are below 2048 bytes, except for the two issues I posted patches for (hisilicon ethernet driver, and nla_put_* users). On the randconfig build test, I have a long series of patches applied that address all known warnings, including my earlier "kasan: turn off -fsanitize-address-use-after-scope for now" (see https://patchwork.kernel.org/patch/9442323/). Without -fsanitize-address-use-after-scope, the problem is also mostly gone (a few cases show up in drivers/media/, and also in block/sed-opal.c and drivers/tty/vt/keyboard.c). I now have initial patches for all of them to bring the stack size below 2048 bytes. As far as I can see, the remaining problems with scary stack frame sizes you found only show up with the combination of all three: KASAN_INLINE, asan_stack=1 and -fsanitize-address-use-after-scope. If all three were separately configurable and we merge the patches I made, I think we could enable the normal 2048 byte warning in all configurations that have at least one of the turned off, but I don't know which of those combinations would actually be sensible for production kernels. Arnd
