Re: [PATCH v2 1/3] fs/exec: don't force writing memory access

2016-09-29 Thread Oleg Nesterov
On 09/29, Jann Horn wrote: > > @@ -204,7 +204,7 @@ static struct page *get_arg_page(struct linux_binprm > *bprm, unsigned long pos, >* doing the exec and bprm->mm is the new process's mm. >*/ > ret = get_user_pages_remote(current, bprm->mm, pos, 1, write, > -

Re: [PATCH v2 1/3] fs/exec: don't force writing memory access

2016-09-29 Thread Oleg Nesterov
On 09/29, Jann Horn wrote: > > @@ -204,7 +204,7 @@ static struct page *get_arg_page(struct linux_binprm > *bprm, unsigned long pos, >* doing the exec and bprm->mm is the new process's mm. >*/ > ret = get_user_pages_remote(current, bprm->mm, pos, 1, write, > -

[PATCH v2 1/3] fs/exec: don't force writing memory access

2016-09-28 Thread Jann Horn
This shouldn't change behavior in any way - at this point, there should be no non-writable mappings, only the initial stack mapping -, but this change makes it easier to reason about the correctness of the following commits that place restrictions on forced memory writes. Signed-off-by: Jann Horn

[PATCH v2 1/3] fs/exec: don't force writing memory access

2016-09-28 Thread Jann Horn
This shouldn't change behavior in any way - at this point, there should be no non-writable mappings, only the initial stack mapping -, but this change makes it easier to reason about the correctness of the following commits that place restrictions on forced memory writes. Signed-off-by: Jann Horn