Hi,

While looking at this today, I noticed that the 'iov' argument to this
function is totally dead code. Patch 1 gets rid of it, and with this
function and import_ubuf() now being identical, converts the callers
to just use import_ubuf().

No functional changes here, just a cleanup removing some dead code.

 drivers/block/ublk_drv.c         |  9 ++-------
 drivers/char/random.c            |  6 ++----
 fs/aio.c                         |  2 +-
 include/linux/uio.h              |  2 --
 kernel/trace/trace_events_user.c |  4 +---
 lib/iov_iter.c                   | 13 -------------
 net/ipv4/tcp.c                   | 10 ++++------
 net/socket.c                     |  6 ++----
 security/keys/keyctl.c           |  5 ++---
 9 files changed, 14 insertions(+), 43 deletions(-)

-- 
Jens Axboe

On Mon, Dec 04, 2023 at 10:47:48AM -0700, Jens Axboe wrote:
> Hi,
> 
> While looking at this today, I noticed that the 'iov' argument to this
> function is totally dead code. Patch 1 gets rid of it, and with this
> function and import_ubuf() now being identical, converts the callers
> to just use import_ubuf().
> 
> No functional changes here, just a cleanup removing some dead code.

Looks good. I've picked this up into vfs.iov_iter but let me know if
you're already carrying this.

Christian

On 12/5/23 3:57 AM, Christian Brauner wrote:
> On Mon, Dec 04, 2023 at 10:47:48AM -0700, Jens Axboe wrote:
>> Hi,
>>
>> While looking at this today, I noticed that the 'iov' argument to this
>> function is totally dead code. Patch 1 gets rid of it, and with this
>> function and import_ubuf() now being identical, converts the callers
>> to just use import_ubuf().
>>
>> No functional changes here, just a cleanup removing some dead code.
> 
> Looks good. I've picked this up into vfs.iov_iter but let me know if
> you're already carrying this.

That's fine thanks, I don't have them stashed anywhere.

-- 
Jens Axboe