Re: [RFC 1/1] io_uring: preserve work->mm since actual work processing may need it

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On 4/10/2020 1:03 AM, Bijan Mottahedeh wrote:
> Do not clear work->mm since io_madvise() passes it to do_madvise()
> when the request is actually processed.

As I see, this down_read() from the trace is
down_read(&current->mm->mmap_sem), where current->mm is set by use_mm()
just several lines above your change. So, what do you mean by passing? I
don't see do_madvise() __explicitly__ accepting mm as an argument.

What tree do you use? Extra patches on top?

> 
> Signed-off-by: Bijan Mottahedeh <bijan.mottahedeh@xxxxxxxxxx>
> ---
>  fs/io-wq.c | 2 --
>  1 file changed, 2 deletions(-)
> 
> diff --git a/fs/io-wq.c b/fs/io-wq.c
> index 4023c98..4d20754 100644
> --- a/fs/io-wq.c
> +++ b/fs/io-wq.c
> @@ -431,8 +431,6 @@ static void io_wq_switch_mm(struct io_worker *worker, struct io_wq_work *work)
>  		if (!worker->mm)
>  			set_fs(USER_DS);
>  		worker->mm = work->mm;
> -		/* hang on to this mm */
> -		work->mm = NULL;
>  		return;
>  	}
>  
> 

-- 
Pavel Begunkov



[Index of Archives]     [Linux Samsung SoC]     [Linux Rockchip SoC]     [Linux Actions SoC]     [Linux for Synopsys ARC Processors]     [Linux NFS]     [Linux NILFS]     [Linux USB Devel]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]


  Powered by Linux