linux-kernel - Re: [PATCH] drivers/block/null_blk: Switch from radix tree api to xarrays

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date: Fri, 29 Dec 2023 09:52:47 -0700
From: Jens Axboe <axboe@...nel.dk>
To: Gautam Menghani <gautam@...ux.ibm.com>, kch@...dia.com,
 ming.lei@...hat.com, damien.lemoal@...nsource.wdc.com,
 zhouchengming@...edance.com, nj.shetty@...sung.com, akinobu.mita@...il.com
Cc: linux-block@...r.kernel.org, linux-kernel@...r.kernel.org,
 riteshh@...ux.ibm.com
Subject: Re: [PATCH] drivers/block/null_blk: Switch from radix tree api to
 xarrays

On 12/29/23 9:41 AM, Gautam Menghani wrote:
> Convert the null_blk driver to use the xarray API instead of radix tree 
> API.
> 
> Testing:
> Used blktests test suite (block and zbd suites) to test the current 
> null_blk driver and null_blk driver with this patch applied. The tests 
> results in both the instances were the same.

What's the purpose of the change?

One thing that always annoys me slightly with xarray is the implied
locking. So now you're grabbing two locks rather than just utilizing the
lock that was already held. I think a better transformation would be to
first change the locking to be closer to the lookup and deletion, and
then convert to xarray and now being able to drop that other lock. Just
doing a blind conversion like this without potentially understanding the
details of it is not a good idea, imho.

-- 
Jens Axboe