2014-12-19 02:31:06

by Toshi Kikuchi

Subject: [PATCH] lib/genalloc.c: fix the end addr check in addr_in_gen_pool()

Since chunk->end_addr is (chunk->start_addr + size - 1),
the end address to compare should be (start + size - 1).

Signed-off-by: Toshi Kikuchi <[email protected]>
---
lib/genalloc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/genalloc.c b/lib/genalloc.c
index 2e65d20..42a95e9 100644
--- a/lib/genalloc.c
+++ b/lib/genalloc.c
@@ -415,7 +415,7 @@ bool addr_in_gen_pool(struct gen_pool *pool, unsigned long start,
size_t size)
{
bool found = false;
- unsigned long end = start + size;
+ unsigned long end = start + size - 1;
struct gen_pool_chunk *chunk;

rcu_read_lock();
--
2.2.0.rc0.207.ga3a616c

2014-12-22 21:05:41

by Andrew Morton

Subject: Re: [PATCH] lib/genalloc.c: fix the end addr check in addr_in_gen_pool()

On Thu, 18 Dec 2014 18:30:59 -0800 Toshi Kikuchi <[email protected]> wrote:

> Since chunk->end_addr is (chunk->start_addr + size - 1),
> the end address to compare should be (start + size - 1).
>
> Signed-off-by: Toshi Kikuchi <[email protected]>
> ---
> lib/genalloc.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/lib/genalloc.c b/lib/genalloc.c
> index 2e65d20..42a95e9 100644
> --- a/lib/genalloc.c
> +++ b/lib/genalloc.c
> @@ -415,7 +415,7 @@ bool addr_in_gen_pool(struct gen_pool *pool, unsigned long start,
> size_t size)
> {
> bool found = false;
> - unsigned long end = start + size;
> + unsigned long end = start + size - 1;
> struct gen_pool_chunk *chunk;
>
> rcu_read_lock();

urgh. gen_pool_chunk.end_addr should have been made exclusive, not
inclusive. Or switch to start_addr/size. The code would be
considerably nicer that way.

And the struct gen_pool_chunk definition should be moved into
genalloc.c. I'm not sure what drivers/acpi/apei/ghes.c is doing
fiddling around with genalloc internals, but it should stop doing it.

Sigh :(