diff mbox series

mm/vma: Return the exact errno for vms_gather_munmap_vmas()

Message ID 20240901145025.2311-1-ice_yangxiao@163.com (mailing list archive)
State New
Headers show
Series mm/vma: Return the exact errno for vms_gather_munmap_vmas() | expand

Commit Message

杨晓 Sept. 1, 2024, 2:50 p.m. UTC
can_modify_vma() in vms_gather_munmap_vmas() returns -EPERM if vma is
sealed so don't always return the fixed -ENOMEM on failure.

Fixes: c2eb22189bbc ("mm/vma: inline munmap operation in mmap_region()")
Fixes: 5887a7ac2383 ("mm/vma: expand mmap_region() munmap call")
Signed-off-by: Xiao Yang <ice_yangxiao@163.com>
Reported-by: kernel test robot <oliver.sang@intel.com>
Closes: https://lore.kernel.org/oe-lkp/202408312155.fd26a58c-oliver.sang@intel.com
---
 mm/mmap.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

Comments

Liam R. Howlett Sept. 3, 2024, 3:58 p.m. UTC | #1
Thanks, but this is already fixed in v8 of my patches.

* Xiao Yang <ice_yangxiao@163.com> [240901 10:51]:
> can_modify_vma() in vms_gather_munmap_vmas() returns -EPERM if vma is
> sealed so don't always return the fixed -ENOMEM on failure.
> 
> Fixes: c2eb22189bbc ("mm/vma: inline munmap operation in mmap_region()")
> Fixes: 5887a7ac2383 ("mm/vma: expand mmap_region() munmap call")
> Signed-off-by: Xiao Yang <ice_yangxiao@163.com>
> Reported-by: kernel test robot <oliver.sang@intel.com>
> Closes: https://lore.kernel.org/oe-lkp/202408312155.fd26a58c-oliver.sang@intel.com
> ---
>  mm/mmap.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/mm/mmap.c b/mm/mmap.c
> index c1781f643046..c9a0dc035819 100644
> --- a/mm/mmap.c
> +++ b/mm/mmap.c
> @@ -1386,8 +1386,9 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
>  		mt_on_stack(mt_detach);
>  		mas_init(&mas_detach, &mt_detach, /* addr = */ 0);
>  		/* Prepare to unmap any existing mapping in the area */
> -		if (vms_gather_munmap_vmas(&vms, &mas_detach))
> -			return -ENOMEM;
> +		error = vms_gather_munmap_vmas(&vms, &mas_detach);
> +		if (error)
> +			return error;
>  
>  		vmg.next = vms.next;
>  		vmg.prev = vms.prev;
> -- 
> 2.44.0
> 
>
杨晓 Sept. 4, 2024, 4:37 a.m. UTC | #2
Hi Liam R


Thanks for your reply.

Could you share the URL of your v8 patches.



Best Regards,
Xiao Yang

At 2024-09-03 23:58:26, "Liam R. Howlett" <Liam.Howlett@oracle.com> wrote:
>
>Thanks, but this is already fixed in v8 of my patches.
>
>* Xiao Yang <ice_yangxiao@163.com> [240901 10:51]:
>> can_modify_vma() in vms_gather_munmap_vmas() returns -EPERM if vma is
>> sealed so don't always return the fixed -ENOMEM on failure.
>> 
>> Fixes: c2eb22189bbc ("mm/vma: inline munmap operation in mmap_region()")
>> Fixes: 5887a7ac2383 ("mm/vma: expand mmap_region() munmap call")
>> Signed-off-by: Xiao Yang <ice_yangxiao@163.com>
>> Reported-by: kernel test robot <oliver.sang@intel.com>
>> Closes: https://lore.kernel.org/oe-lkp/202408312155.fd26a58c-oliver.sang@intel.com
>> ---
>>  mm/mmap.c | 5 +++--
>>  1 file changed, 3 insertions(+), 2 deletions(-)
>> 
>> diff --git a/mm/mmap.c b/mm/mmap.c
>> index c1781f643046..c9a0dc035819 100644
>> --- a/mm/mmap.c
>> +++ b/mm/mmap.c
>> @@ -1386,8 +1386,9 @@ unsigned long mmap_region(struct file *file, unsigned long addr,
>>  		mt_on_stack(mt_detach);
>>  		mas_init(&mas_detach, &mt_detach, /* addr = */ 0);
>>  		/* Prepare to unmap any existing mapping in the area */
>> -		if (vms_gather_munmap_vmas(&vms, &mas_detach))
>> -			return -ENOMEM;
>> +		error = vms_gather_munmap_vmas(&vms, &mas_detach);
>> +		if (error)
>> +			return error;
>>  
>>  		vmg.next = vms.next;
>>  		vmg.prev = vms.prev;
>> -- 
>> 2.44.0
>> 
>>
diff mbox series

Patch

diff --git a/mm/mmap.c b/mm/mmap.c
index c1781f643046..c9a0dc035819 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c
@@ -1386,8 +1386,9 @@  unsigned long mmap_region(struct file *file, unsigned long addr,
 		mt_on_stack(mt_detach);
 		mas_init(&mas_detach, &mt_detach, /* addr = */ 0);
 		/* Prepare to unmap any existing mapping in the area */
-		if (vms_gather_munmap_vmas(&vms, &mas_detach))
-			return -ENOMEM;
+		error = vms_gather_munmap_vmas(&vms, &mas_detach);
+		if (error)
+			return error;
 
 		vmg.next = vms.next;
 		vmg.prev = vms.prev;