Thanks, I will do more work about sparse and maybe find some answers. Zhang Jianhua 在 2022/5/19 12:22, Eric Biggers 写道:
[Please use reply all, not just reply!] On Thu, May 19, 2022 at 11:54:48AM +0800, zhangjianhua (E) wrote:Hi Eric The warnings in commit message are from the build log in Jan 2022, and I find these sizeof are still here, so I submit these two patches. I build the kernel just now and encounter the same situation with you, there are lots of warnings. Maybe you are right, there should be some mechanism to solve this problem completely.I've updated the commit message and applied this patch, but not the other one, as the other one wasn't actually dealing with a variable length which made it pretty much pointless, as I mentioned. If you'd like to look into making sparse enable this warning by default, I'd certainly encourage you to do so. But it looks like the warning itself could use some more work. It probably should only warn if the sizeof(struct_with_flexible_array) is actually being added to another value, and where that value is not a compile-time constant. - Eric .
![]() |