linux-kernel - Re: [PATCH v3 1/9] iov_iter: Add a function to extract an iter's buffers to a bvec iter

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] [day] [month] [year] [list]

Message-ID: <YpWvolJHuCWAnL62@infradead.org>
Date:   Mon, 30 May 2022 23:03:14 -0700
From:   Christoph Hellwig <hch@...radead.org>
To:     David Howells <dhowells@...hat.com>
Cc:     Steve French <smfrench@...il.com>,
        Shyam Prasad N <nspmangalore@...il.com>,
        Rohith Surabattula <rohiths.msft@...il.com>,
        Jeff Layton <jlayton@...nel.org>,
        Al Viro <viro@...iv.linux.org.uk>, linux-cifs@...r.kernel.org,
        linux-fsdevel@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v3 1/9] iov_iter: Add a function to extract an iter's
 buffers to a bvec iter

On Fri, May 27, 2022 at 11:44:02AM +0100, David Howells wrote:
> Copy cifs's setup_aio_ctx_iter() and to lib/iov_iter.c and generalise it as
> extract_iter_to_iter().  This allocates and sets up an array of bio_vecs
> for all the page fragments in an I/O iterator and sets a second supplied
> iterator to bvec-type pointing to the array.
> 
> This is can be used when setting up for a direct I/O or an asynchronous I/O
> to set up a record of the page fragments that are going to contribute to
> the buffer, paging them all in to prevent DIO->mmap loops and allowing the
> original iterator to be deallocated (it may be on the stack of the caller).
> 
> Note that extract_iter_to_iter() doesn't actually need to make a separate
> allocation for the page array.  It can place the page array at the end of
> the bvec array storage, provided it traverses both arrays from the 0th
> element forwards.

I really do not like this as a general purpose helper.  This is an odd
quirk that we really generally should not needed unless you have very
convoluted locking.  So please keep it inside of cifs.