Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1577950 > unrolled thread

[PATCH 1/4] ipc/shm: do not check for MAP_POPULATE

Back to article view | Back to linux.kernel

This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by below is the oldest one visible, not the original post.

Contents

  [PATCH 1/4] ipc/shm: do not check for MAP_POPULATE Davidlohr Bueso <dave@stgolabs.net> - 2017-02-09 22:00 +0100
    Re: [PATCH 1/4] ipc/shm: do not check for MAP_POPULATE Hugh Dickins <hughd@google.com> - 2017-02-10 03:30 +0100
      Re: [PATCH 1/4] ipc/shm: do not check for MAP_POPULATE Davidlohr Bueso <dave@stgolabs.net> - 2017-02-13 19:20 +0100

#1577950 — [PATCH 1/4] ipc/shm: do not check for MAP_POPULATE

We do not support prefaulting functionality in sysv shm,
nor MAP_NONBLOCK for that matter. Drop the pointless check
for populate in do_shmat().

Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
---
 ipc/shm.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/ipc/shm.c b/ipc/shm.c
index 06ea9ef7f54a..6b3769967789 100644
--- a/ipc/shm.c
+++ b/ipc/shm.c
@@ -1234,8 +1234,6 @@ long do_shmat(int shmid, char __user *shmaddr, int shmflg,
 		err = (long)addr;
 invalid:
 	up_write(&current->mm->mmap_sem);
-	if (populate)
-		mm_populate(addr, populate);
 
 out_fput:
 	fput(file);
-- 
2.6.6

[toc] | [next] | [standalone]

#1578120

On Thu, 9 Feb 2017, Davidlohr Bueso wrote:

> We do not support prefaulting functionality in sysv shm,
> nor MAP_NONBLOCK for that matter. Drop the pointless check
> for populate in do_shmat().

I haven't checked, but are you sure that "populated" does nothing
when the attacher had previously called mlockall(MCL_FUTURE)?

Hugh

> 
> Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
> ---
>  ipc/shm.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/ipc/shm.c b/ipc/shm.c
> index 06ea9ef7f54a..6b3769967789 100644
> --- a/ipc/shm.c
> +++ b/ipc/shm.c
> @@ -1234,8 +1234,6 @@ long do_shmat(int shmid, char __user *shmaddr, int shmflg,
>  		err = (long)addr;
>  invalid:
>  	up_write(&current->mm->mmap_sem);
> -	if (populate)
> -		mm_populate(addr, populate);
>  
>  out_fput:
>  	fput(file);
> -- 
> 2.6.6

[toc] | [prev] | [next] | [standalone]

#1579980

On Thu, 09 Feb 2017, Hugh Dickins wrote:

>I haven't checked, but are you sure that "populated" does nothing
>when the attacher had previously called mlockall(MCL_FUTURE)?

I checked and you are certainly right. Andrew, please do not
consider this patch, it's bogus.

Thanks,
Davidlohr

[toc] | [prev] | [standalone]

Back to top | Article view | linux.kernel

csiph-web