Re: [PATCH] streq()

[Denis Vlasenko]

On 24 September 2002 03:40, Ingo Molnar wrote:
> On Tue, 24 Sep 2002, Rusty Russell wrote:
> there's a few more places that tend to cause wasted time, no matter what:
>  - kmalloc(size, flags)/gfp(order, flags) argument ordering. A few months
>    ago i wasted two days on such a bug - since 'size' was very small
>    usually, it never showed up that the allocated buffer was short, until
>    some rare load-test increased the 'size'.

Aughment kmalloc(size, GFP_XXXXX) with kmalloc_XXXXX(size) (inline of course)?
You may use this form in those 90% cases where flags are constant.

This also gets rid of GFP_ prefixes (shorter).

Ingo, Rusty?
--
vda

--- linux-2.5.36/include/linux/slab.h.orig Tue Sep 24 11:00:42 2002
+++ linux-2.5.36/include/linux/slab.h      Tue Sep 24 11:06:32 2002
@@ -61,6 +61,15 @@

 extern void *kmalloc(size_t, int);
 extern void kfree(const void *);
+extern inline void *kmalloc_NOHIGHIO(size_t sz) { return kmalloc(sz, GFP_NOHIGHIO); }
+extern inline void *kmalloc_NOIO    (size_t sz) { return kmalloc(sz, GFP_NOIO    ); }
+extern inline void *kmalloc_NOFS    (size_t sz) { return kmalloc(sz, GFP_NOFS    ); }
+extern inline void *kmalloc_ATOMIC  (size_t sz) { return kmalloc(sz, GFP_ATOMIC  ); }
+extern inline void *kmalloc_USER    (size_t sz) { return kmalloc(sz, GFP_USER    ); }
+extern inline void *kmalloc_HIGHUSER(size_t sz) { return kmalloc(sz, GFP_HIGHUSER); }
+extern inline void *kmalloc_KERNEL  (size_t sz) { return kmalloc(sz, GFP_KERNEL  ); }
+extern inline void *kmalloc_NFS     (size_t sz) { return kmalloc(sz, GFP_NFS     ); }
+extern inline void *kmalloc_KSWAPD  (size_t sz) { return kmalloc(sz, GFP_KSWAPD  ); }

 extern int FASTCALL(kmem_cache_reap(int));