@BruceDawson0xB Hey; regarding https://randomascii.wordpress.com/2018/08/16/24-core-cpu-and-i-cant-type-an-email-part-one/ … - Was this when Chrome was compiled with msvc? Otherwise I'm confused about how CFG was affecting V8 allocations if you were compiling with clang-cl and only using CFG on system libraries....?
Well, the reservation will there if CFG is enabled on anything; but all pages don't get bitmap entries automatically I don't think. (Unless they get everything-is-a-valid-target entries?) 1/2
-
-
You need to specify PAGE_TARGETS_INVALID+SetProcessValidCallTargets for executable VirtualAlloc-ed pages; and you can have CFG on for some libraries but not others...
Thanks. Twitter will use this to make your timeline better. UndoUndo
-
-
-
Ignoring the double negative... Why do you think that all pages don't get bitmap entries automatically? My entire recent post is about hangs caused this happening. Presumably it happens to allow protected modules to call others.
-
Okay; yea, that makes sense. Reusing address space reduces the bitmap size; but so would enabling CFG on more things - JITted regions, libraries - fewer call targets, less bitmap usage.
-
But combining more CFG use with address reuse would be dangerous since prior call targets would still be valid even after you've deallocated and replaced.
-
I don't follow. This would only be dangerous if the old CFG bitmap was retained, which be an odd thing to do. Please run some tests to confirm your ideas. I don't think they match the reality. Start with these and vmmap:https://github.com/randomascii/blogstuff/tree/master/cfg …
End of conversation
New conversation -
Loading seems to be taking a while.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.