diff mbox series

[v2,3/4] selinux: use vma_is_initial_stack() and vma_is_initial_heap()

Message ID 20230719075127.47736-4-wangkefeng.wang@huawei.com (mailing list archive)
State New, archived
Headers show
Series mm: convert to vma_is_initial_heap/stack() | expand

Commit Message

Kefeng Wang July 19, 2023, 7:51 a.m. UTC
Use the helpers to simplify code.

Cc: Paul Moore <paul@paul-moore.com>
Cc: Stephen Smalley <stephen.smalley.work@gmail.com>
Cc: Eric Paris <eparis@parisplace.org>
Acked-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
---
 security/selinux/hooks.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

Comments

David Hildenbrand July 19, 2023, 8:18 a.m. UTC | #1
On 19.07.23 09:51, Kefeng Wang wrote:
> Use the helpers to simplify code.
> 
> Cc: Paul Moore <paul@paul-moore.com>
> Cc: Stephen Smalley <stephen.smalley.work@gmail.com>
> Cc: Eric Paris <eparis@parisplace.org>
> Acked-by: Paul Moore <paul@paul-moore.com>
> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
> ---
>   security/selinux/hooks.c | 7 ++-----
>   1 file changed, 2 insertions(+), 5 deletions(-)
> 
> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> index d06e350fedee..ee8575540a8e 100644
> --- a/security/selinux/hooks.c
> +++ b/security/selinux/hooks.c
> @@ -3762,13 +3762,10 @@ static int selinux_file_mprotect(struct vm_area_struct *vma,
>   	if (default_noexec &&
>   	    (prot & PROT_EXEC) && !(vma->vm_flags & VM_EXEC)) {
>   		int rc = 0;
> -		if (vma->vm_start >= vma->vm_mm->start_brk &&
> -		    vma->vm_end <= vma->vm_mm->brk) {
> +		if (vma_is_initial_heap(vma)) {
>   			rc = avc_has_perm(sid, sid, SECCLASS_PROCESS,
>   					  PROCESS__EXECHEAP, NULL);
> -		} else if (!vma->vm_file &&
> -			   ((vma->vm_start <= vma->vm_mm->start_stack &&
> -			     vma->vm_end >= vma->vm_mm->start_stack) ||
> +		} else if (!vma->vm_file && (vma_is_initial_stack(vma) ||
>   			    vma_is_stack_for_current(vma))) {
>   			rc = avc_has_perm(sid, sid, SECCLASS_PROCESS,
>   					  PROCESS__EXECSTACK, NULL);

Reviewed-by: David Hildenbrand <david@redhat.com>
Christian Göttsche July 19, 2023, 9:02 a.m. UTC | #2
On Wed, 19 Jul 2023 at 09:40, Kefeng Wang <wangkefeng.wang@huawei.com> wrote:
>
> Use the helpers to simplify code.
>
> Cc: Paul Moore <paul@paul-moore.com>
> Cc: Stephen Smalley <stephen.smalley.work@gmail.com>
> Cc: Eric Paris <eparis@parisplace.org>
> Acked-by: Paul Moore <paul@paul-moore.com>
> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
> ---
>  security/selinux/hooks.c | 7 ++-----
>  1 file changed, 2 insertions(+), 5 deletions(-)
>
> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> index d06e350fedee..ee8575540a8e 100644
> --- a/security/selinux/hooks.c
> +++ b/security/selinux/hooks.c
> @@ -3762,13 +3762,10 @@ static int selinux_file_mprotect(struct vm_area_struct *vma,
>         if (default_noexec &&
>             (prot & PROT_EXEC) && !(vma->vm_flags & VM_EXEC)) {
>                 int rc = 0;
> -               if (vma->vm_start >= vma->vm_mm->start_brk &&
> -                   vma->vm_end <= vma->vm_mm->brk) {
> +               if (vma_is_initial_heap(vma)) {

This seems to change the condition from

    vma->vm_start >= vma->vm_mm->start_brk && vma->vm_end <= vma->vm_mm->brk

to

    vma->vm_start <= vma->vm_mm->brk && vma->vm_end >= vma->vm_mm->start_brk

(or AND arguments swapped)

    vma->vm_end >= vma->vm_mm->start_brk && vma->vm_start <= vma->vm_mm->brk

Is this intended?

>                         rc = avc_has_perm(sid, sid, SECCLASS_PROCESS,
>                                           PROCESS__EXECHEAP, NULL);
> -               } else if (!vma->vm_file &&
> -                          ((vma->vm_start <= vma->vm_mm->start_stack &&
> -                            vma->vm_end >= vma->vm_mm->start_stack) ||
> +               } else if (!vma->vm_file && (vma_is_initial_stack(vma) ||
>                             vma_is_stack_for_current(vma))) {
>                         rc = avc_has_perm(sid, sid, SECCLASS_PROCESS,
>                                           PROCESS__EXECSTACK, NULL);
> --
> 2.27.0
>
Kefeng Wang July 19, 2023, 10:22 a.m. UTC | #3
On 2023/7/19 17:02, Christian Göttsche wrote:
> On Wed, 19 Jul 2023 at 09:40, Kefeng Wang <wangkefeng.wang@huawei.com> wrote:
>>
>> Use the helpers to simplify code.
>>
>> Cc: Paul Moore <paul@paul-moore.com>
>> Cc: Stephen Smalley <stephen.smalley.work@gmail.com>
>> Cc: Eric Paris <eparis@parisplace.org>
>> Acked-by: Paul Moore <paul@paul-moore.com>
>> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
>> ---
>>   security/selinux/hooks.c | 7 ++-----
>>   1 file changed, 2 insertions(+), 5 deletions(-)
>>
>> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
>> index d06e350fedee..ee8575540a8e 100644
>> --- a/security/selinux/hooks.c
>> +++ b/security/selinux/hooks.c
>> @@ -3762,13 +3762,10 @@ static int selinux_file_mprotect(struct vm_area_struct *vma,
>>          if (default_noexec &&
>>              (prot & PROT_EXEC) && !(vma->vm_flags & VM_EXEC)) {
>>                  int rc = 0;
>> -               if (vma->vm_start >= vma->vm_mm->start_brk &&
>> -                   vma->vm_end <= vma->vm_mm->brk) {
>> +               if (vma_is_initial_heap(vma)) {
> 
> This seems to change the condition from
> 
>      vma->vm_start >= vma->vm_mm->start_brk && vma->vm_end <= vma->vm_mm->brk
> 
> to
> 
>      vma->vm_start <= vma->vm_mm->brk && vma->vm_end >= vma->vm_mm->start_brk
> 
> (or AND arguments swapped)
> 
>      vma->vm_end >= vma->vm_mm->start_brk && vma->vm_start <= vma->vm_mm->brk
> 
> Is this intended?

The new condition is to check whether there is intersection between
[startbrk,brk] and [vm_start,vm_end], it contains orignal check, so
I think it is ok, but for selinux check, I am not sure if there is
some other problem.

> 
>>                          rc = avc_has_perm(sid, sid, SECCLASS_PROCESS,
>>                                            PROCESS__EXECHEAP, NULL);
>> -               } else if (!vma->vm_file &&
>> -                          ((vma->vm_start <= vma->vm_mm->start_stack &&
>> -                            vma->vm_end >= vma->vm_mm->start_stack) ||
>> +               } else if (!vma->vm_file && (vma_is_initial_stack(vma) ||
>>                              vma_is_stack_for_current(vma))) {
>>                          rc = avc_has_perm(sid, sid, SECCLASS_PROCESS,
>>                                            PROCESS__EXECSTACK, NULL);
>> --
>> 2.27.0
>>
Paul Moore July 19, 2023, 3:25 p.m. UTC | #4
On Wed, Jul 19, 2023 at 6:23 AM Kefeng Wang <wangkefeng.wang@huawei.com> wrote:
> On 2023/7/19 17:02, Christian Göttsche wrote:
> > On Wed, 19 Jul 2023 at 09:40, Kefeng Wang <wangkefeng.wang@huawei.com> wrote:
> >>
> >> Use the helpers to simplify code.
> >>
> >> Cc: Paul Moore <paul@paul-moore.com>
> >> Cc: Stephen Smalley <stephen.smalley.work@gmail.com>
> >> Cc: Eric Paris <eparis@parisplace.org>
> >> Acked-by: Paul Moore <paul@paul-moore.com>
> >> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
> >> ---
> >>   security/selinux/hooks.c | 7 ++-----
> >>   1 file changed, 2 insertions(+), 5 deletions(-)
> >>
> >> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> >> index d06e350fedee..ee8575540a8e 100644
> >> --- a/security/selinux/hooks.c
> >> +++ b/security/selinux/hooks.c
> >> @@ -3762,13 +3762,10 @@ static int selinux_file_mprotect(struct vm_area_struct *vma,
> >>          if (default_noexec &&
> >>              (prot & PROT_EXEC) && !(vma->vm_flags & VM_EXEC)) {
> >>                  int rc = 0;
> >> -               if (vma->vm_start >= vma->vm_mm->start_brk &&
> >> -                   vma->vm_end <= vma->vm_mm->brk) {
> >> +               if (vma_is_initial_heap(vma)) {
> >
> > This seems to change the condition from
> >
> >      vma->vm_start >= vma->vm_mm->start_brk && vma->vm_end <= vma->vm_mm->brk
> >
> > to
> >
> >      vma->vm_start <= vma->vm_mm->brk && vma->vm_end >= vma->vm_mm->start_brk
> >
> > (or AND arguments swapped)
> >
> >      vma->vm_end >= vma->vm_mm->start_brk && vma->vm_start <= vma->vm_mm->brk
> >
> > Is this intended?
>
> The new condition is to check whether there is intersection between
> [startbrk,brk] and [vm_start,vm_end], it contains orignal check, so
> I think it is ok, but for selinux check, I am not sure if there is
> some other problem.

This particular SELinux vma check is see if the vma falls within the
heap; can you confirm that this change preserves this?
Kefeng Wang July 20, 2023, 8:28 a.m. UTC | #5
On 2023/7/19 23:25, Paul Moore wrote:
> On Wed, Jul 19, 2023 at 6:23 AM Kefeng Wang <wangkefeng.wang@huawei.com> wrote:
>> On 2023/7/19 17:02, Christian Göttsche wrote:
>>> On Wed, 19 Jul 2023 at 09:40, Kefeng Wang <wangkefeng.wang@huawei.com> wrote:
>>>>
>>>> Use the helpers to simplify code.
>>>>
>>>> Cc: Paul Moore <paul@paul-moore.com>
>>>> Cc: Stephen Smalley <stephen.smalley.work@gmail.com>
>>>> Cc: Eric Paris <eparis@parisplace.org>
>>>> Acked-by: Paul Moore <paul@paul-moore.com>
>>>> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
>>>> ---
>>>>    security/selinux/hooks.c | 7 ++-----
>>>>    1 file changed, 2 insertions(+), 5 deletions(-)
>>>>
>>>> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
>>>> index d06e350fedee..ee8575540a8e 100644
>>>> --- a/security/selinux/hooks.c
>>>> +++ b/security/selinux/hooks.c
>>>> @@ -3762,13 +3762,10 @@ static int selinux_file_mprotect(struct vm_area_struct *vma,
>>>>           if (default_noexec &&
>>>>               (prot & PROT_EXEC) && !(vma->vm_flags & VM_EXEC)) {
>>>>                   int rc = 0;
>>>> -               if (vma->vm_start >= vma->vm_mm->start_brk &&
>>>> -                   vma->vm_end <= vma->vm_mm->brk) {
>>>> +               if (vma_is_initial_heap(vma)) {
>>>
>>> This seems to change the condition from
>>>
>>>       vma->vm_start >= vma->vm_mm->start_brk && vma->vm_end <= vma->vm_mm->brk
>>>
>>> to
>>>
>>>       vma->vm_start <= vma->vm_mm->brk && vma->vm_end >= vma->vm_mm->start_brk
>>>
>>> (or AND arguments swapped)
>>>
>>>       vma->vm_end >= vma->vm_mm->start_brk && vma->vm_start <= vma->vm_mm->brk
>>>
>>> Is this intended?
>>
>> The new condition is to check whether there is intersection between
>> [startbrk,brk] and [vm_start,vm_end], it contains orignal check, so
>> I think it is ok, but for selinux check, I am not sure if there is
>> some other problem.
> 
> This particular SELinux vma check is see if the vma falls within the
> heap; can you confirm that this change preserves this?

Yes, within is one case of new vma scope check.

>
Paul Moore July 20, 2023, 9:16 p.m. UTC | #6
On Thu, Jul 20, 2023 at 4:28 AM Kefeng Wang <wangkefeng.wang@huawei.com> wrote:
> On 2023/7/19 23:25, Paul Moore wrote:
> > On Wed, Jul 19, 2023 at 6:23 AM Kefeng Wang <wangkefeng.wang@huawei.com> wrote:
> >> On 2023/7/19 17:02, Christian Göttsche wrote:
> >>> On Wed, 19 Jul 2023 at 09:40, Kefeng Wang <wangkefeng.wang@huawei.com> wrote:
> >>>>
> >>>> Use the helpers to simplify code.
> >>>>
> >>>> Cc: Paul Moore <paul@paul-moore.com>
> >>>> Cc: Stephen Smalley <stephen.smalley.work@gmail.com>
> >>>> Cc: Eric Paris <eparis@parisplace.org>
> >>>> Acked-by: Paul Moore <paul@paul-moore.com>
> >>>> Signed-off-by: Kefeng Wang <wangkefeng.wang@huawei.com>
> >>>> ---
> >>>>    security/selinux/hooks.c | 7 ++-----
> >>>>    1 file changed, 2 insertions(+), 5 deletions(-)
> >>>>
> >>>> diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
> >>>> index d06e350fedee..ee8575540a8e 100644
> >>>> --- a/security/selinux/hooks.c
> >>>> +++ b/security/selinux/hooks.c
> >>>> @@ -3762,13 +3762,10 @@ static int selinux_file_mprotect(struct vm_area_struct *vma,
> >>>>           if (default_noexec &&
> >>>>               (prot & PROT_EXEC) && !(vma->vm_flags & VM_EXEC)) {
> >>>>                   int rc = 0;
> >>>> -               if (vma->vm_start >= vma->vm_mm->start_brk &&
> >>>> -                   vma->vm_end <= vma->vm_mm->brk) {
> >>>> +               if (vma_is_initial_heap(vma)) {
> >>>
> >>> This seems to change the condition from
> >>>
> >>>       vma->vm_start >= vma->vm_mm->start_brk && vma->vm_end <= vma->vm_mm->brk
> >>>
> >>> to
> >>>
> >>>       vma->vm_start <= vma->vm_mm->brk && vma->vm_end >= vma->vm_mm->start_brk
> >>>
> >>> (or AND arguments swapped)
> >>>
> >>>       vma->vm_end >= vma->vm_mm->start_brk && vma->vm_start <= vma->vm_mm->brk
> >>>
> >>> Is this intended?
> >>
> >> The new condition is to check whether there is intersection between
> >> [startbrk,brk] and [vm_start,vm_end], it contains orignal check, so
> >> I think it is ok, but for selinux check, I am not sure if there is
> >> some other problem.
> >
> > This particular SELinux vma check is see if the vma falls within the
> > heap; can you confirm that this change preserves this?
>
> Yes, within is one case of new vma scope check.

Thanks for the confirmation.
diff mbox series

Patch

diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c
index d06e350fedee..ee8575540a8e 100644
--- a/security/selinux/hooks.c
+++ b/security/selinux/hooks.c
@@ -3762,13 +3762,10 @@  static int selinux_file_mprotect(struct vm_area_struct *vma,
 	if (default_noexec &&
 	    (prot & PROT_EXEC) && !(vma->vm_flags & VM_EXEC)) {
 		int rc = 0;
-		if (vma->vm_start >= vma->vm_mm->start_brk &&
-		    vma->vm_end <= vma->vm_mm->brk) {
+		if (vma_is_initial_heap(vma)) {
 			rc = avc_has_perm(sid, sid, SECCLASS_PROCESS,
 					  PROCESS__EXECHEAP, NULL);
-		} else if (!vma->vm_file &&
-			   ((vma->vm_start <= vma->vm_mm->start_stack &&
-			     vma->vm_end >= vma->vm_mm->start_stack) ||
+		} else if (!vma->vm_file && (vma_is_initial_stack(vma) ||
 			    vma_is_stack_for_current(vma))) {
 			rc = avc_has_perm(sid, sid, SECCLASS_PROCESS,
 					  PROCESS__EXECSTACK, NULL);