Re: [PATCH 0/4] null_blk: usr memcpy_[to|from]_page()

On 3/30/23 12:56?PM, Chaitanya Kulkarni wrote:

On 3/29/23 15:48, Jens Axboe wrote:

quoted

On 3/29/23 2:46?PM, Chaitanya Kulkarni wrote:

quoted

Hi,

 From :include/linux/highmem.h:
"kmap_atomic - Atomically map a page for temporary usage - Deprecated!"

Use memcpy_from_page() since does the same job of mapping, copying, and
unmaping except it uses non deprecated kmap_local_page() and
kunmap_local(). Following are the differences between kmal_local_page()
and kmap_atomic() :-

Looks fine to me, but I'd fold patches 1-3 rather than split them up.

Sent V2 with above comment, first three patches are from
different code path and they are doing unrelated changes:-

1. WRITE :- copy_to_nullb() only use memcpy_page().
2. READ :- copy_from_nullb() only use memcapy_page() and zero_user().
3. I guess zoned read beyond write pointer null_fill_pattern()
    memset_page().

if anything goes wrong in any of 3 code paths we will have to
entire change which we shouldn't, that's why kept it separate,
I'm fine with whatever you decide ...

It doesn't matter... It's not like this is a hugely complicated thing.
It's just a straight forward conversion. They all just switch to using a
helper, which is the right thing to do as it reduces the stuff we have
to do in there. Only reason why I excluded patch 4 is that it is a bit
different than the others. But it really could just be one patch.

-- 
Jens Axboe