Hi, > Cool. Here's a version that gets rid of the lookup table by encoding the > posted write pending bit into the reg offset. This should be OK, as the > functions are used within dmtimer.c only. Ok, that looks nice and will generate a bit better code. I thought briefly on that it but wanted the change to be small and was a little worried someone might use the offset somewhere else. But, they way you did it makes the 2nd aspect go away. * You have defined WPSHIFT to 8. However, that bit is currently taken by WP_TOCR. I chose 15 hoping if the register expanded it would be the last one in a u_int16 and still can be encoded in an op code as a shift value. A 16 would be a good codegen value, but always assume u_int32 register. As the registers are u_int32 in current implementations, probably 16 is a better value. > The init of other timers into posted mode is not done yet, so I changed > the timer->posted handling too. Not inited in software, BUT the power on reset of the timer module is to posted mode in reset. Its better to have someone shut it off in both places as its on by default. I'll tweak the points mentioned and fix a typo in the description and send a version. Regards, Richard W. -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
