On Wed, 2014-07-16 at 15:57 -0700, Greg KH wrote: > On Wed, Jul 16, 2014 at 03:46:09PM -0700, Joe Perches wrote: > > On Wed, 2014-07-16 at 15:10 -0700, Greg KH wrote: > > > On Wed, Jul 16, 2014 at 11:03:06PM +0100, Michalis Pappas wrote: > > [] > > > > Ok, I agree on the ifdef stuff. My question was regarding your > > > > suggestion above to replace print_hex_debug() with dev_dbg() > > > > > > You want the device name/id/label to show up as well, that is why you > > > should use the dev_dbg() version, print_hex_dump() does not take a > > > struct device *, so the user has no idea what device this data was > > > coming from. > > > > But Michalis could alway add something like: > > dev_hex_dump() > > and > > dev_dbg_hex_dump() > > > > With the built-in "hex dump primitive" in printk(), why would you want > to do that? You shouldn't be putting more than 64 bytes in a single > printk message in a hex dump, if you want to do more, use debugfs. %p*h doesn't also emit "ascii or dots" for one. I agree %p*h should be preferred but people do all sorts of strange things in debugging printks. _______________________________________________ devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxx http://driverdev.linuxdriverproject.org/mailman/listinfo/driverdev-devel
