On Wed, Jun 29, 2016 at 03:48:25PM +0300, Andy Shevchenko wrote: > On Wed, 2016-06-29 at 17:46 +0530, Rajneesh Bhardwaj wrote: > > On Wed, Jun 29, 2016 at 02:49:45PM +0300, Andy Shevchenko wrote: > > > On Wed, 2016-06-29 at 13:33 +0530, Rajneesh Bhardwaj wrote: > > > > On Tue, Jun 28, 2016 at 01:05:07PM +0300, Andy Shevchenko wrote: > > > > > While this is good for simple attributes i see a potential issue > > > > with > > > > this > > > > approch when we want to display multiple attributes for a single > > > > debugfs > > > > entry. > > > > We typically use seq_printf in a loop and display multiline > > > > output > > > > for a single debugfs entry. How do we go about the scenarios where > > > > we > > > > want > > > > to display formatted multiline debug information? > > > > > > We are talking about abstract case or is it change coming to this > > > very > > > driver? > > > > > > > We will be sending few patches for this very driver shortly and one > > such > > feature would require multiline formatted output for a single debugfs > > entry. > > If it touches the same attribute it needs to be seen first before > applying this one. > No, this attribute wont be touched. > > > > Can we have seq_printf for such cases and use DEFINE_DEBUGFS_ATTRIBUTE > > for > > other simple attributes? Doing this might raise some coding style > > consistency > > concerns though so lets wait to hear Darren's opinion on this patch. > > For the rest we have to see the implementation and decide how to proceed > (since it's a debugfs dedicated for _debugging_) we might change it in > the way we like. Though I don't encourage people to do this. I like > interfaces on which people thought before implementing. > Sure, i am fine with DEFINE_DEBUGFS_ATTRIBUTE but would like to leave the scope for having seq_printf for future patches i.e. linux/seq_file.h and thus would have a dependency on struct file_operations for seq_read. > -- > > Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx> > Intel Finland Oy -- Best Regards, Rajneesh -- To unsubscribe from this list: send the line "unsubscribe platform-driver-x86" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
