|
|
Log in / Subscribe / Register

banishing pointers in static domains

banishing pointers in static domains

Posted Feb 7, 2026 0:09 UTC (Sat) by Fowl (subscriber, #65667)
Parent article: Kernel control-flow-integrity support comes to GCC

I’m wondering if an index+switch scheme would have a prohibitive performance impact. Would avoid the problem of having to suppress optimizations and in fact they’d likely be welcomed - devirtualization in the kernel wow!

I’m not sure how to resolve the “C files are compiled separately” thing - *waves hand* linker magic? More hand waving for loadable modules.

Perhaps there would be compensating memory savings - eg. use a u8 instead of a pointer for each function pointer member, modulo alignment/atomicity. 2^8 possible functions per signature ought to be enough for anyone ey?


The LWN site is currently under high scraper load, so comment display has been suppressed for anonymous users. If you are a human, you may read the comments by clicking the button below:

Note: you can avoid this step in the future by logging into your LWN account.


Copyright © 2026, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds