On Fri, 14 May 2021 05:51:24 +0300 Tzvetomir Stoyanov <tz.stoyanov@xxxxxxxxx> wrote: > On Fri, May 14, 2021 at 12:23 AM Steven Rostedt <rostedt@xxxxxxxxxxx> wrote: > > > > On Fri, 7 May 2021 12:53:28 +0300 > > "Tzvetomir Stoyanov (VMware)" <tz.stoyanov@xxxxxxxxx> wrote: > > > > > Removed the "weak" definition of the library log functions. Setting > > > the desired log level can be used to silence the library logs, instead > > > of overwriting the log functions. > > > > That's not the purpose of the weak definition. It is so that a GUI (like > > kernelshark) can turn it into a pop up if need be or show in a status > > window. It was not for quieting the function. > > The description is not correct. All these log functions use > tep_vprint(), which is implemented as weak. This function has > information for the library and log severity and should be used by > kernelshark for that purpose. That's why I removed weak library > specific functions. Yes, tep_vprintk() could offer this as well, but there's still no reason to not keep them as weak. Adding log levels is still unrelated to removing the weak attribute. I still do not see the rationale for removing the weak attribute of these functions. It gives more flexibility for the users, without any downsides. Or is there a downside that you see for keeping them weak? -- Steve
|