-
Show this thread
-
Replying to @mjg59
a downside of dealing with it like that in the PTE zap code is that if someone unmaps a page with an elevated refcount caused by direct IO or whatever, the page will get zeroed mid-IO. and if you're in a fork child, you could zero parent CoW pages. and stuff like that.
2 replies 0 retweets 2 likes -
perhaps a cleaner way would be to add a new PG_* flag in page-flags.h and set that on pages allocated in the VMA. require madvise to happen while the VMA has no pages. and then do the wiping in the put_page() machinery.
1 reply 0 retweets 0 likes -
Replying to @tehjh
So yeah, ideally this is something that propagates down to the pages rather than just being a property of the VMA. Hmm.
1 reply 0 retweets 0 likes -
1 reply 0 retweets 0 likes
added some comments to the commit on github
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.