linux-kernel - Re: [PATCH 2/6] iov_iter: new iov_iter_pin_pages*(), for FOLL

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]

Message-ID: <0c5bcd23-607f-2ef9-daa0-11557c9f8e8f@nvidia.com>
Date:   Mon, 28 Feb 2022 14:49:11 -0800
From:   John Hubbard <jhubbard@...dia.com>
To:     Jens Axboe <axboe@...nel.dk>, jhubbard.send.patches@...il.com,
        Jan Kara <jack@...e.cz>, Christoph Hellwig <hch@...radead.org>,
        Dave Chinner <dchinner@...hat.com>,
        "Darrick J . Wong" <djwong@...nel.org>,
        Theodore Ts'o <tytso@....edu>,
        Alexander Viro <viro@...iv.linux.org.uk>,
        Miklos Szeredi <miklos@...redi.hu>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Chaitanya Kulkarni <kch@...dia.com>
Cc:     linux-block@...r.kernel.org, linux-fsdevel@...r.kernel.org,
        linux-xfs@...r.kernel.org, linux-mm@...ck.org,
        LKML <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 2/6] iov_iter: new iov_iter_pin_pages*(), for FOLL_PIN
 pages

On 2/27/22 13:57, Jens Axboe wrote:
>> +ssize_t iov_iter_pin_pages(struct iov_iter *i,
>> +		   struct page **pages, size_t maxsize, unsigned int maxpages,
>> +		   size_t *start)
>> +{
>> +	size_t len;
>> +	int n, res;
>> +
>> +	if (maxsize > i->count)
>> +		maxsize = i->count;
>> +	if (!maxsize)
>> +		return 0;
>> +
>> +	WARN_ON_ONCE(!iter_is_iovec(i));
>> +
>> +	if (likely(iter_is_iovec(i))) {
>> +		unsigned int gup_flags = 0;
>> +		unsigned long addr;
>> +
>> +		if (iov_iter_rw(i) != WRITE)
>> +			gup_flags |= FOLL_WRITE;
>> +		if (i->nofault)
>> +			gup_flags |= FOLL_NOFAULT;
>> +
>> +		addr = first_iovec_segment(i, &len, start, maxsize, maxpages);
>> +		n = DIV_ROUND_UP(len, PAGE_SIZE);
>> +		res = pin_user_pages_fast(addr, n, gup_flags, pages);
>> +		if (unlikely(res <= 0))
>> +			return res;
>> +		return (res == n ? len : res * PAGE_SIZE) - *start;
> 
> Trying to be clever like that just makes the code a lot less readable. I
> should not have to reason about a return value. Same in the other
> function.
> 

Here is a differential patch on top of this one, and only showing one of
the two routines. How does this direction look to you?


diff --git a/lib/iov_iter.c b/lib/iov_iter.c
index e64e8e4edd0c..8e96f1e9ebc6 100644
--- a/lib/iov_iter.c
+++ b/lib/iov_iter.c
@@ -1588,7 +1588,17 @@ ssize_t iov_iter_pin_pages(struct iov_iter *i,
  		res = pin_user_pages_fast(addr, n, gup_flags, pages);
  		if (unlikely(res <= 0))
  			return res;
-		return (res == n ? len : res * PAGE_SIZE) - *start;
+
+		/* Cap len at the number of pages that were actually pinned: */
+		if (res < n)
+			len = res * PAGE_SIZE;
+
+		/*
+		 * The return value is the amount pinned in bytes that the
+		 * caller will actually use. So, reduce it by the offset into
+		 * the first page:
+		 */
+		return len - *start;
  	}

  	return -EFAULT;

thanks,
-- 
John Hubbard
NVIDIA