From mboxrd@z Thu Jan 1 00:00:00 1970 From: Paul Mundt Date: Wed, 27 Aug 2008 14:40:48 +0000 Subject: Re: [PATCH] i2c: fix i2c-sh_mobile timing issues Message-Id: <20080827144048.GA32530@linux-sh.org> List-Id: References: <20080827093356.9145.33751.sendpatchset@rx1.opensource.se> In-Reply-To: <20080827093356.9145.33751.sendpatchset@rx1.opensource.se> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Magnus Damm Cc: i2c@lm-sensors.org, khali@linux-fr.org, linux-sh@vger.kernel.org, ben-linux@fluff.org On Wed, Aug 27, 2008 at 06:33:56PM +0900, Magnus Damm wrote: > From: Magnus Damm > > This patch teaches the i2c-sh_mobile driver to make use of wait irqs. > Without this patch only dte irqs are used which may lead to overruns > and cases of missing stop and extra bytes being read on the i2c bus. > > Use of wait irqs forces the hardware to pause and wait until the cpu > is ready. Polling is also reworked in this patch to fix ms delay issues. > > Verified with bus analyzer and tested on MigoR and AP325RXA boards. > > Signed-off-by: Magnus Damm > --- > > Paul, can you merge this for 2.6.27? > > drivers/i2c/busses/i2c-sh_mobile.c | 271 +++++++++++++++++++++++++++--------- > 1 file changed, 208 insertions(+), 63 deletions(-) > I'll add it to my queue if Jean and Ben don't have any issues with the patch. If the polling works as is, I would suggest that it is not 2.6.27 material, but if it's fixing operational bugs then it doesn't make sense to leave 2.6.27 in a known broken state.