GCC 4.3.0 exposes a kernel bug

blound · Post by **blound** » Sun Mar 09, 2008 7:55 pm

http://lwn.net/SubscriberLink/272048/34349d400065bc89/

interesting read

binutils · Post by **binutils** » Tue Mar 11, 2008 5:34 pm

this issue was foreseen and it was come.

linux kernel developers and gcc developers, they need to talk with each other.
linux kernel doesn't work fine with latest gcc compiler, it happened from time to time.

B.E · Post by **B.E** » Tue Mar 11, 2008 8:04 pm

This bug is no real threat, moreover, by the time this new version of gcc is used to compile programs, the problem will have been fixed, secondly it only effects programs that uses signal handlers, and also use memmove() (or a function that uses and relies on the DF flag), before and within the signal handler. Also there is no escalation of privileges, and will only run at the current users privileges.

Solar · Post by **Solar** » Wed Mar 12, 2008 2:06 am

It should be noted that the kernel bug would be exposed by any compiler correctly implementing the ABI.

I think many kernel gurus are pretty arrogant bas****s, for half a dozen reasons, and can't help but smile whenever they're bitten by that attitude.

Candy · Post by **Candy** » Fri Mar 28, 2008 2:47 pm

Solar wrote:It should be noted that the kernel bug would be exposed by any compiler correctly implementing the ABI.

I think many kernel gurus are pretty arrogant bas****s, for half a dozen reasons, and can't help but smile whenever they're bitten by that attitude.

I still recall them calling GCC 3.* bad because it broke their code ... not to mention that their code was the problem in the first place.