On Mon 29-01-18 04:37:45, Matthew Wilcox wrote: > On Tue, Jan 23, 2018 at 01:26:46PM +0100, Michal Hocko wrote: > > On Tue 16-01-18 06:13:54, Matthew Wilcox wrote: > > > 3. Maybe we could rename kvfree() to just free()? Please? There's > > > nothing special about it. One fewer thing for somebody to learn when > > > coming fresh to kernel programming. > > > > I guess one has to learn about kvmalloc already and kvfree is nicely > > symmetric to it. > > I'd really like to get to: > > #define malloc(sz) kvmalloc(sz, GFP_KERNEL) > #define free(p) kvfree(p) > #define realloc(p, sz) kvrealloc(p, sz, GFP_KERNEL) /* Doesn't exist yet */ > #define calloc(n, sz) kvmalloc_array(n, sz, GFP_KERNEL) Considering how many users we already have this is not really feasible. I am also not sure we really want to mimic the userspace API. It is just different with different consequences. You do not want to hide GFP flags because this turned out to be just a bad idea in the past. Just look at pte allocation functions which are unconditioanlly GFP_KERNEL and all the pain that resulted in. We really want people to learn the APIs and understand their limitations. -- Michal Hocko SUSE Labs -- 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>
