Re: [PATCH 03/18] Blackfin Serial Driver: fix a in dma circle rx buffer handling

[Andrew Morton <akpm@linux-foundation.org>]

On Fri,  6 Mar 2009 14:42:45 +0800
Bryan Wu <cooloney@kernel.org> wrote:

> Subject: [PATCH 03/18] Blackfin Serial Driver: fix a in dma circle rx buffer handling

I changed this to "fix a bug"

> Date: 	Fri,  6 Mar 2009 14:42:45 +0800
> Sender: linux-kernel-owner@vger.kernel.org
> X-Mailer: git-send-email 1.5.6.3
> 
> From: Sonic Zhang <sonic.zhang@analog.com>
> 
> Reported-by: Qian Zhang <zhangq@sansitech.com>
> Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
> Signed-off-by: Bryan Wu <cooloney@kernel.org>

No changelog.

>  drivers/serial/bfin_5xx.c |    4 +++-
>  1 files changed, 3 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/serial/bfin_5xx.c b/drivers/serial/bfin_5xx.c
> index 350bfc4..1e027e3 100644
> --- a/drivers/serial/bfin_5xx.c
> +++ b/drivers/serial/bfin_5xx.c
> @@ -402,9 +402,11 @@ static void bfin_serial_dma_rx_chars(struct bfin_serial_port *uart)
>  	else
>  		flg = TTY_NORMAL;
>  
> -	for (i = uart->rx_dma_buf.tail; i != uart->rx_dma_buf.head; i++) {
> +	for (i = uart->rx_dma_buf.tail; ; i++) {
>  		if (i >= UART_XMIT_SIZE)
>  			i = 0;
> +		if (i == uart->rx_dma_buf.head)
> +			break;
>  		if (!uart_handle_sysrq_char(&uart->port, uart->rx_dma_buf.buf[i]))
>  			uart_insert_char(&uart->port, status, OE,
>  				uart->rx_dma_buf.buf[i], flg);

The fix is not sufficiently obvious to justify the omission of a
changelog.

Please always include a changelog unless the patch is quite obvious.

Please send a changelog entry for this patch.  This one should contain
a description of the bug and a description of how it was fixed.  A
bugfix's changelog should also provide the reviewer with an
understanding of the seriousness of the bug - what impact does it have
upon users?  Because without this information we cannot decide into
which kernel version(s) the code should be merged.

It is alse very useful if the maintainer and/or patch submitter provide
information about which kernel versions are affected and whether the
fix should be backported into -stable kerenls and if so, why.


Bryan, I should not need to explain all this stuff to you.  Please
don't let this happen again.