mm: vmalloc: actually use the in-place vrealloc region

Patch series "mm: vmalloc: Actually use the in-place vrealloc region".

This fixes a performance regression[1] with vrealloc()[1].


The refactoring to not build a new vmalloc region only actually worked
when shrinking.  Actually return the resized area when it grows.  Ugh.

Link: https://lkml.kernel.org/r/[email protected]
Fixes: a0309fa ("mm: vmalloc: support more granular vrealloc() sizing")
Signed-off-by: Kees Cook <[email protected]>
Reported-by: Shung-Hsi Yu <[email protected]>
Closes: https://lore.kernel.org/all/20250515-bpf-verifier-slowdown-vwo2meju4cgp2su5ckj@6gi6ssxbnfqg [1]
Tested-by: Eduard Zingerman <[email protected]>
Tested-by: Pawan Gupta <[email protected]>
Tested-by: Shung-Hsi Yu <[email protected]>
Reviewed-by: "Uladzislau Rezki (Sony)" <[email protected]>
Reviewed-by: Danilo Krummrich <[email protected]>
Cc: "Erhard F." <[email protected]>
Cc: <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>

Author: kees

mm/vmalloc.c

@@ -4111,6 +4111,7 @@ void *vrealloc_noprof(const void *p, size_t size, gfp_t flags)
 		if (want_init_on_alloc(flags))
 			memset((void *)p + old_size, 0, size - old_size);
 		vm->requested_size = size;
+		return (void *)p;
 	}
 
 	/* TODO: Grow the vm_area, i.e. allocate and map additional pages. */