...
I can take this since wrote the code anyway.
Created attachment 303824 [details] Patch
Comment on attachment 303824 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=303824&action=review Comment below, otherwise, LGTM. r- just because of my comment. > Source/JavaScriptCore/wasm/WasmFaultSignalHandler.cpp:104 > + { > + void* faultingAddress = sigInfo->si_addr; > + dataLogLnIf(verbose, "checking faulting address: ", RawPointer(faultingAddress), " is in an active fast memory"); > + LockHolder locker(memoryLock); > + auto& activeFastMemories = viewActiveFastMemories(locker); > + for (void* activeMemory : activeFastMemories) { > + dataLogLnIf(verbose, "checking fast memory at: ", RawPointer(activeMemory)); > + if (activeMemory <= faultingAddress && faultingAddress < static_cast<char*>(activeMemory) + fastMemoryMappedBytes) { > + faultedInActiveFastMemory = true; You should first check that we're in JIT code otherwise we could've crashed while holding the lock.
Comment on attachment 303824 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=303824&action=review >> Source/JavaScriptCore/wasm/WasmFaultSignalHandler.cpp:104 >> + faultedInActiveFastMemory = true; > > You should first check that we're in JIT code otherwise we could've crashed while holding the lock. Ignore me, I'm wrong.
Comment on attachment 303824 [details] Patch r=me
Comment on attachment 303824 [details] Patch Clearing flags on attachment: 303824 Committed r213599: <http://trac.webkit.org/changeset/213599>
All reviewed patches have been landed. Closing bug.