linux-kernel - Re: [PATCH] iov_iter: Four fixes for 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] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 26 Apr 2021 15:02:46 -0400
From:   Jeff Layton <jlayton@...hat.com>
To:     David Howells <dhowells@...hat.com>,
        Alexander Viro <viro@...iv.linux.org.uk>
Cc:     linux-fsdevel@...r.kernel.org,
        Dave Wysochanski <dwysocha@...hat.com>,
        Marc Dionne <marc.dionne@...istor.com>,
        "Matthew Wilcox (Oracle)" <willy@...radead.org>,
        Christoph Hellwig <hch@....de>, linux-mm@...ck.org,
        linux-cachefs@...hat.com, linux-afs@...ts.infradead.org,
        linux-nfs@...r.kernel.org, linux-cifs@...r.kernel.org,
        ceph-devel@...r.kernel.org, v9fs-developer@...ts.sourceforge.net,
        Trond Myklebust <trond.myklebust@...merspace.com>,
        Anna Schumaker <anna.schumaker@...app.com>,
        Steve French <sfrench@...ba.org>,
        Dominique Martinet <asmadeus@...ewreck.org>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] iov_iter: Four fixes for ITER_XARRAY

On Mon, 2021-04-26 at 00:14 +0100, David Howells wrote:
> Hi Al,
> 
> I think this patch should include all the fixes necessary.  I could merge
> it in, but I think it might be better to tag it on the end as an additional
> patch.
> 
> David
> ---
> iov_iter: Four fixes for ITER_XARRAY
> 
> Fix four things[1] in the patch that adds ITER_XARRAY[2]:
> 
>  (1) Remove the address_space struct predeclaration.  This is a holdover
>      from when it was ITER_MAPPING.
> 
>  (2) Fix _copy_mc_to_iter() so that the xarray segment updates count and
>      iov_offset in the iterator before returning.
> 
>  (3) Fix iov_iter_alignment() to not loop in the xarray case.  Because the
>      middle pages are all whole pages, only the end pages need be
>      considered - and this can be reduced to just looking at the start
>      position in the xarray and the iteration size.
> 
>  (4) Fix iov_iter_advance() to limit the size of the advance to no more
>      than the remaining iteration size.
> 
> Reported-by: Al Viro <viro@...iv.linux.org.uk>
> Signed-off-by: David Howells <dhowells@...hat.com>
> Link: https://lore.kernel.org/r/YIVrJT8GwLI0Wlgx@zeniv-ca.linux.org.uk [1]
> Link: https://lore.kernel.org/r/161918448151.3145707.11541538916600921083.stgit@warthog.procyon.org.uk [2]
> ---
>  include/linux/uio.h |    1 -
>  lib/iov_iter.c      |    5 +++++
>  2 files changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/include/linux/uio.h b/include/linux/uio.h
> index 5f5ffc45d4aa..d3ec87706d75 100644
> --- a/include/linux/uio.h
> +++ b/include/linux/uio.h
> @@ -10,7 +10,6 @@
>  #include <uapi/linux/uio.h>
>  
> 
> 
> 
>  struct page;
> -struct address_space;
>  struct pipe_inode_info;
>  
> 
> 
> 
>  struct kvec {
> diff --git a/lib/iov_iter.c b/lib/iov_iter.c
> index 44fa726a8323..61228a6c69f8 100644
> --- a/lib/iov_iter.c
> +++ b/lib/iov_iter.c
> @@ -791,6 +791,8 @@ size_t _copy_mc_to_iter(const void *addr, size_t bytes, struct iov_iter *i)
>  			curr_addr = (unsigned long) from;
>  			bytes = curr_addr - s_addr - rem;
>  			rcu_read_unlock();
> +			i->iov_offset += bytes;
> +			i->count -= bytes;
>  			return bytes;
>  		}
>  		})
> @@ -1147,6 +1149,7 @@ void iov_iter_advance(struct iov_iter *i, size_t size)
>  		return;
>  	}
>  	if (unlikely(iov_iter_is_xarray(i))) {
> +		size = min(size, i->count);
>  		i->iov_offset += size;
>  		i->count -= size;
>  		return;
> @@ -1346,6 +1349,8 @@ unsigned long iov_iter_alignment(const struct iov_iter *i)
>  			return size | i->iov_offset;
>  		return size;
>  	}
> +	if (unlikely(iov_iter_is_xarray(i)))
> +		return (i->xarray_start + i->iov_offset) | i->count;
>  	iterate_all_kinds(i, size, v,
>  		(res |= (unsigned long)v.iov_base | v.iov_len, 0),
>  		res |= v.bv_offset | v.bv_len,
> 

I did a test run with your v7 pile, this patch, and my ceph fscache
rework patches and it did fine. You can add:

Tested-by: Jeff Layton <jlayton@...hat.com>