On 6/26/23 1:40?PM, Linus Torvalds wrote: > > mm/zsmalloc.c: > set_freeobj(zspage, (unsigned int)(-1UL)); [..] > both of which are impressively confused. There's sadly a number of > other cases too. > > That zsmalloc.c case is impressively crazy. First you use the wrong > type, then you use the wrong operation, and then you use a cast to fix > it all up. Absolutely none of which makes any sense, and it should > just have used "-1". That zsmalloc code is gone, you should not see it after you pull MM patches from Andrew.
