On Wednesday, October 31, 2012 03:27:41 PM Johannes Berg wrote: > On Wed, 2012-10-31 at 14:57 +0200, Vladimir Kondratiev wrote: > > > You didn't also address my comments regarding WIL6210_ISR_COR #ifdef > > > code. > > > > I added string with short explanation in Kconfig. > > Longer explanation: while clear-on-read is good for production mode, it > > makes debugging much harder - reading ISR registers clears interrupt, and > > one can no more monitor ISR with debugfs. So, when debugging ISR flows - > > and they still need some debugging - one have to use W1C mode. That's why > > it is still present. > > Since I'm CC'ed on your patches (not sure why) I'll offer the opinion > that you should add interrupt and IO access tracing rather than reading > registers with some debug thing ;) Johannes, you just proved reason to be CC'ed: you give valuable feedback :-) Oh, no! Usually, you start digging into ISR registers when something stalled when it should not. At this point, you want to discover that, for example, some interrupt bit not came up. Then, you may try to manually set it writing into register. If you trace, it will flood dmesg, impacting performance, in normal situation; and will not let you analyze and recover from abnormal one. I need "manual control" here. > > johannes -- To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
