Messages in this thread | | | Date | Wed, 24 Sep 2008 12:51:01 +0200 | From | Stefan Richter <> | Subject | Re: [PATCH 2/3] e1000e: Useset_memory_ro()/set_memory_rw() to protect flash memory |
| |
David Miller wrote: > From: Jiri Kosina <jkosina@suse.cz> > Date: Wed, 24 Sep 2008 01:45:53 +0200 (CEST) > >> Yes, probably not a vanilla kernel material, but could be very well used >> for testing to further isolate whether it is kernel or userspace >> corrupting the memory mapped eeprom. > > Absolutely.
Would it help and would it be feasible to drastically shorten the lifetime of the MMIO mapping? I.e. only have it mapped briefly during the driver probe, and temporarily whenever a userspace tool like ethtool needs it. -- Stefan Richter -=====-==--- =--= ==--- http://arcgraph.de/sr/
| |