|
From: Paul F. <pj...@wa...> - 2020-03-28 10:41:46
|
Hi I've started looking at adding aligned_alloc support (this bugzilla item https://bugs.kde.org/show_bug.cgi?id=407589). I expect that this will add a new class tag. Would it be better to add to the end of the class tags (but that would add to the ZONE functions) or after POSIX_MEMALIGN (more logical but more disruptive to the code as I would have to renumber MALLOC_USABLE_SIZE and on? A+ Paul |
|
From: Paul F. <pj...@wa...> - 2020-03-28 13:51:06
|
On second thoughts, this doesn't work so well. aligned_alloc and memalign are aliases: paulf> nm /lib64/libc.so.6 | grep align | grep 0000000000081920 0000000000081920 W aligned_alloc 0000000000081920 t __GI___libc_memalign 0000000000081920 T __libc_memalign 0000000000081920 W memalign 0000000000081920 t __memalign So I just get ignored redirects ==9157== WARNING: new redirection conflicts with existing -- ignoring it --9157-- old: 0x05762920 (memalign ) R-> (1011.0) 0x04c2e718 memalign --9157-- new: 0x05762920 (memalign ) R-> (1017.0) 0x04c2e6c6 aligned_alloc This limits the usefulness to either a libc with separate implementations of memalign/aligned_alloc or a replacement function. I'll update the bugzille with this analysis and a patch, but I'm not sure if it is worth going further. Maybe just update the docs? A+ Paul |