In the Linux kernel, the following vulnerability has been resolved: btrfs: fix memory ordering between normal and ordered work functions Ordered work functions aren't guaranteed to be handled by the same thread which executed the normal work functions. The only way execution between normal/ordered functions is synchronized is via the WORKDONEBIT, unfortunately the used bitops don't guarantee any ordering whatsoever. This manifested as seemingly inexplicable crashes on ARM64, where asyncchunk::inode is seen as non-null in asynccowsubmit which causes submitcompressedextents to be called and crash occurs because asyncchunk::inode suddenly became NULL. The call trace was similar to: pc : submitcompressedextents+0x38/0x3d0 lr : asynccowsubmit+0x50/0xd0 sp : ffff800015d4bc20 <registers omitted for brevity> Call trace: submitcompressedextents+0x38/0x3d0 asynccowsubmit+0x50/0xd0 runorderedwork+0xc8/0x280 btrfsworkhelper+0x98/0x250 processonework+0x1f0/0x4ac workerthread+0x188/0x504 kthread+0x110/0x114 retfromfork+0x10/0x18 Fix this by adding respective barrier calls which ensure that all accesses preceding setting of WORKDONEBIT are strictly ordered before setting the flag. At the same time add a read barrier after reading of WORKDONEBIT in runorderedwork which ensures all subsequent loads would be strictly ordered after reading the bit. This in turn ensures are all accesses before WORKDONEBIT are going to be strictly ordered before any access that can occur in orderedfunc.