On 06.07.2012 00:30, Mike Giacomelli wrote: > 1) It uses the register_storage_idle_func to write log files to disk > during spin up. This is similar to how config files and battery bench > logs are handled. Thus, there is no loss of battery life from > logging. Can this be turned off? I am hunting issues which relate to disk access, and I get freezes upon writing / reading in some cases.
> Once this is committed, I will probably redefine all DEBUGF and LOGF > statements to the new system with a 'warning' priority so that they do > not get logged by default. It will then be up to developers going > forward to decide which conditions are critical enough to warrant > logging to disk. Have you made some testing on the performance and the executable size if you really compile in all DEBUGF and LOGF calls? Related to your gerrit, can one see the whole diff of all patchsets together in gerrit somehow? cheers, Jonas
