Hi Aaron, On Wed, 2026-09-30 at 15:05 -0400, Aaron Merey wrote: > Check whether the target shdr has already been loaded without using > scn->elf->lock. Otherwise use this lock only for the one-time lazy > loading of the shdr. > > VALGRIND_HG_DISABLE_CHECKING is used on shdr pointers before their > atomic_store_release call to prevent helgrind false positives since > there is no definite order to whether lock-free loads of the shdr > pointers occur before or after this store. > > Signed-off-by: Aaron Merey <[email protected]> > > --- > v2 getshdr: Set section 0 pointer last, add valgrind annotation
Looks good. We went through the whole locking/atomic changes for v1 and I believe that is indeed correct. The setting of sections zero last is a good idea. Kind of the same question about the VALGRIND_HG_DISABLE_CHECKING for scns.data state, is this the right place or could it be put where data is originally allocated? Thanks, Mark
