From mboxrd@z Thu Jan 1 00:00:00 1970 From: Wolfgang Grandegger Date: Wed, 16 Sep 2009 13:45:03 +0200 Subject: [U-Boot] [PATCH 2/2] fsl_i2c: Add CONFIG_FSL_I2C_CUSTOM_DFSR In-Reply-To: References: <1253040358-30817-1-git-send-email-Joakim.Tjernlund@transmode.se> <1253040358-30817-2-git-send-email-Joakim.Tjernlund@transmode.se> <4AAFE249.9080604@freescale.com> <4AAFE54F.7000501@freescale.com> <4AB0BC4D.1030007@denx.de> Message-ID: <4AB0CFBF.2050009@grandegger.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: u-boot@lists.denx.de Joakim Tjernlund wrote: > Wolfgang Grandegger wrote on 16/09/2009 12:22:05: >> Joakim Tjernlund wrote: >>>> Timur Tabi wrote on 15/09/2009 21:04:47: >>>>> Joakim Tjernlund wrote: >>>>> >>>>>> No, the impact on speed from DFSR is pretty small so it will >>>>>> be close enough. >>>>> How small? >>>> From the app note: >>>> divisor = B * (A + ((3*C)/B)*2); >>>> >>>> C is dfsr and 10 <= A <= 30, 16 <= B <= 2048 >>>> Considering the actual speed may be way lower the requested speed >>>> I think this is small enough. >>> Once we have the new procedure in place, we can calculate the exact >>> divisor so the need for extra CONFIG_ options goes away. >> As Timur pointed out, a new table/algorithm would require some real >> testing and also some feedback from the users. Who knows if "your" >> values do not make trouble. Therefore I vote for using custom >> settings for maximum flexibility: >> >> CONFIG_FSL_I2C_CUSTOM_FDR >> CONFIG_FSL_I2C_CUSTOM_DFSR > > Oh well, since you both wanted it I added it. > Sent 3 patches, the last patch impl. the latest AN2819 spec. > > Would you mind test it a little? OK, I will do some tests later this week. What CPU do you use and at what I2C bus frequency do you test? Wolfgang.