On Wed, 22 Apr 2015 16:33:38 +0800 Gavin Guo <gavin.guo@xxxxxxxxxxxxx> wrote:
> The slub_debug=PU,kmalloc-xx cannot work because in the
> create_kmalloc_caches() the s->name is created after the
> create_kmalloc_cache() is called. The name is NULL in the
> create_kmalloc_cache() so the kmem_cache_flags() would not set the
> slub_debug flags to the s->flags. The fix here set up a kmalloc_names
> string array for the initialization purpose and delete the dynamic
> name creation of kmalloc_caches.
>
> --- a/mm/slab_common.c
> +++ b/mm/slab_common.c
> @@ -793,6 +793,26 @@ void __init create_kmalloc_caches(unsigned long flags)
> int i;
>
> /*
> + * The kmalloc_names is for temporary usage to make
> + * slub_debug=,kmalloc-xx option work in the boot time. The
> + * kmalloc_index() support to 2^26=64MB. So, the final entry of the
> + * table is kmalloc-67108864.
> + */
> + static const char *kmalloc_names[] = {
> + "0", "kmalloc-96", "kmalloc-192",
> + "kmalloc-8", "kmalloc-16", "kmalloc-32",
> + "kmalloc-64", "kmalloc-128", "kmalloc-256",
> + "kmalloc-512", "kmalloc-1024", "kmalloc-2048",
> + "kmalloc-4196", "kmalloc-8192", "kmalloc-16384",
> + "kmalloc-32768", "kmalloc-65536",
> + "kmalloc-131072", "kmalloc-262144",
> + "kmalloc-524288", "kmalloc-1048576",
> + "kmalloc-2097152", "kmalloc-4194304",
> + "kmalloc-8388608", "kmalloc-16777216",
> + "kmalloc-33554432", "kmalloc-67108864"
> + };
> +
> + /*
> * Patch up the size_index table if we have strange large alignment
> * requirements for the kmalloc array. This is only the case for
> * MIPS it seems. The standard arches will not generate any code here.
> @@ -835,7 +855,8 @@ void __init create_kmalloc_caches(unsigned long flags)
> }
> for (i = KMALLOC_SHIFT_LOW; i <= KMALLOC_SHIFT_HIGH; i++) {
> if (!kmalloc_caches[i]) {
> - kmalloc_caches[i] = create_kmalloc_cache(NULL,
> + kmalloc_caches[i] = create_kmalloc_cache(
> + kmalloc_names[i],
> 1 << i, flags);
> }
You could do something like
kmalloc_caches[i] = create_kmalloc_cache(
kmalloc_names[i],
kstrtoul(kmalloc_names[i] + 8),
flags);
here, and remove those weird "96" and "192" cases.
Or if that's considered too messy, make it
static const struct {
const char *name;
unsigned size;
} kmalloc_cache_info[] = {
{ NULL, 0 },
{ "kmalloc-96", 96 },
...
};
but I'm thinking the kstrtoul() trick will be OK.
> - for (i = 0; i <= KMALLOC_SHIFT_HIGH; i++) {
> - struct kmem_cache *s = kmalloc_caches[i];
> - char *n;
> -
> - if (s) {
> - n = kasprintf(GFP_NOWAIT, "kmalloc-%d", kmalloc_size(i));
> -
> - BUG_ON(!n);
> - s->name = n;
> - }
> - }
> -
slab_kmem_cache_release() still does kfree_const(s->name). It will
crash?
--
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>