From: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
To: Sascha Hauer <s.hauer@pengutronix.de>
Cc: barebox@lists.infradead.org
Subject: Re: [PATCH] fs: Fix memcpy_sz for remaining count/rwsize
Date: Mon, 12 Oct 2015 09:36:37 +0200 [thread overview]
Message-ID: <561B6305.40306@gmail.com> (raw)
In-Reply-To: <20151012061112.GJ7858@pengutronix.de>
On 12.10.2015 08:11, Sascha Hauer wrote:
> On Thu, Oct 08, 2015 at 11:19:45PM +0200, Sebastian Hesselbarth wrote:
>> When using memcpy_sz with rwsize != 1 integer division of
>> count/rwsize may leave some bytes of the request uncopied if
>> count is not a multiple of rwsize.
>>
>> Fix this behavior by decrementing count by rwsize instead of
>> integer division and use plain memcpy for the remaining bytes.
>>
>> Signed-off-by: Sebastian Hesselbarth <sebastian.hesselbarth@gmail.com>
>> ---
>> Cc: barebox@lists.infradead.org
>> ---
>> fs/fs.c | 9 ++++++---
>> 1 file changed, 6 insertions(+), 3 deletions(-)
>>
>> diff --git a/fs/fs.c b/fs/fs.c
>> index c041e41bb51b..ccbda22d2692 100644
>> --- a/fs/fs.c
>> +++ b/fs/fs.c
>> @@ -1580,9 +1580,7 @@ static void memcpy_sz(void *dst, const void *src, size_t count, int rwsize)
>>
>> rwsize = rwsize >> O_RWSIZE_SHIFT;
>>
>> - count /= rwsize;
>> -
>> - while (count-- > 0) {
>> + while (count > 0) {
>> switch (rwsize) {
>> case 1:
>> *((u8 *)dst) = *((u8 *)src);
>> @@ -1599,7 +1597,12 @@ static void memcpy_sz(void *dst, const void *src, size_t count, int rwsize)
>> }
>> dst += rwsize;
>> src += rwsize;
>> + count -= rwsize;
>> }
>
> This doesn't look correct. When count > 0 you are inside the loop, so
>
>> +
>> + /* copy remaining bytes with plain memcpy */
>> + if (count)
>> + memcpy(dst, src, count);
>
> here count <= 0 which is no meaningful argument for the copy size.
>
> Should the loop start with while (count >= rwsize) instead?
Dammit, last minute cosmetic change including breaking the
whole patch. Sorry for that.
> I wonder if the behaviour shouldn't rather be:
> - let memcpy_sz return the number of bytes copied and not copy the
> remaining partial word.
> - return error from memcpy_sz when input count < rwsize
>
> This would allow us to catch wrongly aligned sizes.
I am open for any different resolution. I stumbled upon the odd
behavior of memcpy_sz while writing to NAND using memcpy. Maybe
it would be also good to always pick byte size for memcpy when
no specific size has been passed. It took me a while until I
realized it is not the NAND controller but memcpy that breaks
the data written by leaving some bytes uncopied.
Sebastian
_______________________________________________
barebox mailing list
barebox@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/barebox
next prev parent reply other threads:[~2015-10-12 7:37 UTC|newest]
Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-10-08 21:19 Sebastian Hesselbarth
2015-10-12 6:11 ` Sascha Hauer
2015-10-12 7:36 ` Sebastian Hesselbarth [this message]
2015-10-12 18:51 ` Sebastian Hesselbarth
2015-10-13 8:00 ` Sascha Hauer
2015-10-13 8:09 ` Sebastian Hesselbarth
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=561B6305.40306@gmail.com \
--to=sebastian.hesselbarth@gmail.com \
--cc=barebox@lists.infradead.org \
--cc=s.hauer@pengutronix.de \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox