On 05/09/2016 04:20 PM, Dmitry Vyukov wrote: > On Mon, May 9, 2016 at 3:01 PM, Andrey Ryabinin <aryabinin@xxxxxxxxxxxxx> wrote: >> >> >> On 05/09/2016 02:35 PM, Luruo, Kuthonuzo wrote: >>> >>> This patch with atomic bit op is similar in spirit to v1 except that it increases metadata size. >>> >> >> I don't think that this is a big deal. That will slightly increase size of objects <= (128 - 32) bytes. >> And if someone think otherwise, we can completely remove 'alloc_size' >> (we use it only to print size in report - not very useful). > > > Where did 128 come from? > We now should allocate only 32 bytes for 16-byte user object. If not, > there is something to fix. > I just said this wrong. I mean that the patch increases size of objects that have object_size <= (128 - 32). For bigger objects, the new 'struct kasan_[alloc,free]_meta' still fits into optimal redzone. -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@xxxxxxxxx. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>