Hi Michal, On Thu, Aug 6, 2015 at 11:18 AM, Michal Simek <michal.simek@xxxxxxxxxx> wrote: > Every patch like this has to fix something. If some certain IP versions > are affected then it should be listed at least one version. Please re-read my email. I said: - I do NOT know of any bug. - all information about a possible bug come from that three lines of commentary which are in the _original_ mainline driver. - there is no errata mentioned in any of the datasheets I read. > Huh. Don't get what you are saying here. Based on your description it is > HW bug and it is completely fine to have dt property to avoid this bug. > Having another config property is just not acceptable solution now. You misunderstood. To recapitulate: - I have not seen this bug in action.- - I have not observed this bug in simulation. - I don't think it exists in current IP. - I personally think any odd behavior the original committers might have observed came down to bugs in their ISR. If any errata had existed in previous IP, I would expect the datasheets to mention this-- they don't. I have looked at datasheets axi_iic_ds756 v1.01a, axi_iic_ds756 v1.01b, axi_iic_ds756 v1.02a and xps_iic_ds606 v2.03a. I have simulated axi_iic v1.02a and successfully use axi_iic v1.02a in hardware with my patched driver without the CONFIG_I2C_XILINX_ERRATA configuration option configured, i.e. # CONFIG_I2C_XILINX_ERRATA is not set. On the off-chance this bug _does_ exist in previous or current IP, I provided this compile-time configurable option to enable the _old_ mainline behavior of resetting the IP after every TX error so as to not break somebody's platform. And that's why it isn't a DT property. It can always be promoted to a DT property when the existence of any errata were actually verified, but I have neither time nor inclination to do so. Therefore, I _don't know_ which version of the IP _if any_ is affected. If you require more information, I suggest you inquire with Richard Röjfors or possibly Ben Dooks about it. Richard Röjfors is to git blame for the reset-after-Tx-error functionality and the comment mentioning a possible errata. Regards, Robert -- To unsubscribe from this list: send the line "unsubscribe linux-i2c" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
