Christoph Hellwig wrote:
On Mon, Oct 27, 2008 at 05:02:38PM -0700, David Daney wrote:
Signed-off-by: Tomaso Paoletti <tpaoletti@xxxxxxxxxxxxxxxxxx>
Signed-off-by: David Daney <ddaney@xxxxxxxxxxxxxxxxxx>
---
.../cavium-octeon/executive/cvmx-csr-addresses.h | 8391 ++++++
arch/mips/cavium-octeon/executive/cvmx-csr-enums.h | 86 +
.../cavium-octeon/executive/cvmx-csr-typedefs.h |27517 ++++++++++++++++++++
27517 lines in a header and it's all junk?
I'm glad you asked. No it is not all junk.
That file contains the bit definitions for all on-chip registers.
We are interested in transforming this information into a form suitable
for inclusion in the kernel. Any specific suggestions as to improve the
patch will be considered.
Several possibilities are:
1) Don't typedef all the unions in cvmx-csr-typedefs.h. An rename the
file so it doesn't contain the reprehensible word 'typedef'
2) Break cvmx-csr-addresses.h and cvmx-csr-typedefs.h into several
parts, one for each functional block in the processor.
There are obviously other options as well...
Thanks,
David Daney
