FCC patch..really
From: Allen Curtis <hidden>
Date: 2002-08-28 00:08:04
-- All things come to those who wait. Some of us just have to wait a little longer...
5 messages, 2 authors, 2002-08-28 · open the first message on its own page
From: Allen Curtis <hidden>
Date: 2002-08-28 00:08:04
-- All things come to those who wait. Some of us just have to wait a little longer...
From: Dan Malek <hidden>
Date: 2002-08-28 03:52:36
Allen Curtis wrote:
+ if [ "$CONFIG_FCC1_ENET" = "y" ]; then + choice 'RX Clk Source' \
Please don't do this.......it doesn't belong in the configuration file. There are more things unique to a board configuration than the clock routing. This belongs as a structure in a C file, selectable by board type (along with other board specific selections). Thanks. -- Dan ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
From: Allen Curtis <hidden>
Date: 2002-08-28 04:11:25
quoted
+ if [ "$CONFIG_FCC1_ENET" = "y" ]; then + choice 'RX Clk Source' \Please don't do this.......it doesn't belong in the configuration file. There are more things unique to a board configuration than the clock routing. This belongs as a structure in a C file, selectable by board type (along with other board specific selections).
This may not be the right implementation but I believe it is the right direction. Personally I am tired of looking at all the different board specific implementations and conditional compile statements. How can we accomplish the same thing without cluttering the code with platform stuff? I think there needs to be an effort to define the kernel features and have the configuration translate platform features into generic kernel features that can be used by everyone. It doesn't do anyone any good to make them splunk though the Linux source to find platform tweaks. (although it does help to keep us contractors employed) I have another update that does the same thing with the MDIO lines per PHY ;) ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
From: Dan Malek <hidden>
Date: 2002-08-28 05:29:38
Allen Curtis wrote:
I have another update that does the same thing with the MDIO lines per PHY ;)
Please don't. This stuff doesn't belong in configuration files. There is nothing wrong with creating platform specific configuration in files that can be compiled and using a single configuration option and #define to make this happen. Cluttered configuration scripts are a source of error and headache. -- Dan ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/
From: Allen Curtis <hidden>
Date: 2002-08-28 14:43:34
quoted
I have another update that does the same thing with the MDIOlines per PHYquoted
;)Please don't. This stuff doesn't belong in configuration files. There is nothing wrong with creating platform specific configuration in files that can be compiled and using a single configuration option and #define to make this happen. Cluttered configuration scripts are a source of error and headache.
Ok but lets keep the #defines that control the FCC setup and put them into the platform headers. If the defines are not present then they will default to the "normal" configuration. Unfortunately with this method you are required to modify each platform configuration file instead of having it done as part of the configuration process. I am not sure which would be more error prone. ** Sent via the linuxppc-embedded mail list. See http://lists.linuxppc.org/