2022-10-19 08:24:44

by Rolf Eike Beer

Subject: [PATCH] fs/exec.c: simplify initial stack size expansion

I had a hard time trying to understand completely why it is using vm_end in
one side of the expression and vm_start in the other one, and using
something in the "if" clause that is not an exact copy of what is used
below. The whole point is that the stack_size variable that was used in the
"if" clause is the difference between vm_start and vm_end, which is not far
away but makes this thing harder to read than it must be.

Signed-off-by: Rolf Eike Beer <[email protected]>
---
fs/exec.c | 13 +++++--------
1 file changed, 5 insertions(+), 8 deletions(-)

diff --git a/fs/exec.c b/fs/exec.c
index 768843477a49..990891c5d8fe 100644
--- a/fs/exec.c
+++ b/fs/exec.c
@@ -840,16 +840,13 @@ int setup_arg_pages(struct linux_binprm *bprm,
* will align it up.
*/
rlim_stack = bprm->rlim_stack.rlim_cur & PAGE_MASK;
+
+ stack_expand = min(rlim_stack, stack_size + stack_expand);
+
#ifdef CONFIG_STACK_GROWSUP
- if (stack_size + stack_expand > rlim_stack)
- stack_base = vma->vm_start + rlim_stack;
- else
- stack_base = vma->vm_end + stack_expand;
+ stack_base = vma->vm_start + stack_expand;
#else
- if (stack_size + stack_expand > rlim_stack)
- stack_base = vma->vm_end - rlim_stack;
- else
- stack_base = vma->vm_start - stack_expand;
+ stack_base = vma->vm_end - stack_expand;
#endif
current->mm->start_stack = bprm->p;
ret = expand_stack(vma, stack_base);
--
2.37.3


--
Rolf Eike Beer, emlix GmbH, https://www.emlix.com
Fon +49 551 30664-0, Fax +49 551 30664-11
Gothaer Platz 3, 37083 Göttingen, Germany
Sitz der Gesellschaft: Göttingen, Amtsgericht Göttingen HR B 3160
Geschäftsführung: Heike Jordan, Dr. Uwe Kracke – Ust-IdNr.: DE 205 198 055

emlix - smart embedded open source

2022-10-25 22:42:41

by Kees Cook

Subject: Re: [PATCH] fs/exec.c: simplify initial stack size expansion

On Wed, 19 Oct 2022 09:32:35 +0200, Rolf Eike Beer wrote:
> I had a hard time trying to understand completely why it is using vm_end in
> one side of the expression and vm_start in the other one, and using
> something in the "if" clause that is not an exact copy of what is used
> below. The whole point is that the stack_size variable that was used in the
> "if" clause is the difference between vm_start and vm_end, which is not far
> away but makes this thing harder to read than it must be.
>
> [...]

Applied to for-next/execve, thanks!

[1/1] fs/exec.c: simplify initial stack size expansion
https://git.kernel.org/kees/c/bfb4a2b95875

--
Kees Cook