This is the mail archive of the
ecos-patches@sources.redhat.com
mailing list for the eCos project.
Re: RealTek 8139 driver, 3. revision
- From: Gary Thomas <gary at mlbassoc dot com>
- To: Eric Doenges <Eric dot Doenges at DynaPel dot de>
- Cc: eCos patches <ecos-patches at sources dot redhat dot com>
- Date: 22 Aug 2003 06:44:40 -0600
- Subject: Re: RealTek 8139 driver, 3. revision
- Organization: MLB Associates
- References: <3F3A5255.4080701@DynaPel.de>
On Wed, 2003-08-13 at 08:59, Eric Doenges wrote:
> Hi,
>
>
> this is the third revision of my 8139 driver, with changes suggested
> and/or supplied by Gary Thomas and Andy Dyer. The differences to the
> second revision are:
>
> 1. The '_init' function calls cyg_pci_configure_device to take care
> of platforms where the HAL does not configure PCI devices at
> startup.
>
> 2. Used the new 'CYGARC_PCI_DMA_ADDRESS' macro instead of
> CYGARC_PHYSICAL_ADDRESS.
>
> 3. Added the use of CYGARC_UNCACHED_ADDRESS for those platforms
> where this makes sense. Since not all platforms support this
> (mine doesn't, for instance), also added a
> CYGPKG_DEVS_ETH_RLTK_8139_SOFTWARE_CACHE_COHERENCY that can be set
> in the platform-specific .inl file. If defined, HAL_DCACHE_FLUSH
> is called after copying data to the transmit buffer, and
> HAL_DCACHE_INVALIDATE is called before retrieving data from the
> receive buffer. How to create the buffers so that cache line
> invalidation does not trash adjacent data structures is left as
> an exercise to the reader (since it will be platform specific).
>
> 4. Added a function that dumps 8139 register contents when debugging
> the driver.
I finally had time to test this with my hardware and it worked just
fine. ... now just to get the paperwork, so we can merge it into CVS!
--
Gary Thomas <gary@mlbassoc.com>
MLB Associates