That is terrifying.
-
-
Replying to @pervognsen
If the suggested "maybe this helps?" workarounds are 1. turn off SMT 2. turn off uop cache, _and neither appears to work_, I think there
1 reply 0 retweets 9 likes -
Replying to @rygorous @pervognsen
might be a new stepping on the horizon... this sounds bad.
1 reply 0 retweets 12 likes -
Replying to @rygorous @pervognsen
Sounds like OP's system is overclocked though?https://www.asus.com/us/Motherboards/PRIME-B350-PLUS/specifications/ …
1 reply 0 retweets 1 like -
https://www.phoronix.com/forums/forum/phoronix/latest-phoronix-articles/955368-some-ryzen-linux-users-are-facing-issues-with-heavy-compilation-loads?p=955498#post955498 … seems to talk about the same issue without any overclocking?
1 reply 1 retweet 1 like -
Replying to @tehjh @JamesWidman and
and apparently http://gitweb.dragonflybsd.org/dragonfly.git/commitdiff/b48dd28447fc8ef62fbc963accd301557fd9ac20 … is the OS-level workaround - place the userspace stack a bit lower?
1 reply 5 retweets 5 likes -
Ooooh. That would make sense: when IFetch crosses the canonical boundary, everything goes to hell.
1 reply 0 retweets 2 likes -
I don't understand why they don't leave a full unmapped page
1 reply 0 retweets 0 likes -
neither do I; Windows leaves the last 64k unmapped. https://blogs.msdn.microsoft.com/oldnewthing/20141009-00/?p=43883 …
1 reply 1 retweet 2 likes -
Yeah, I was thinking along the same lines. Can't get there without going through an ordinary unmapped range -- avoids all sorts of bugs
1 reply 0 retweets 0 likes
however, relying on such a safety margin in software can lead to brittle code if you're not super-careful: https://bugs.chromium.org/p/project-zero/issues/detail?id=1184 …
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.