Hi everybody...
v2009.08-rc1 has just been released.
Please help testing, and submit fixes to the remaining problems. Things still missing should be submitted ASAP.
v2009.08-rc1 will also be the base for the "next" branch which I will create soon.
Please use the "next" branch as base for any new work. Note: Please keep in mind that it may be rebased against master every now and then without warning, but I will usually try to avoid this.
Summary of current status:
* Many boards now throw the (expected)
#warning Ethernet driver is deprecated. Please update to use CONFIG_NET_MULTI
messages.
* ARM (as far as present [see below]) shows this:
===== actux4 ===== actux4.c: In function 'board_init': actux4.c:83: warning: left shift count >= width of type actux4.c:83: warning: left shift count >= width of type ===== mv88f6281gtw_ge ===== mv88e61xx.c: In function 'mv88e61xx_switch_initialize': mv88e61xx.c:346: warning: 'idstr' may be used uninitialized in this function ===== rd6281a ===== mv88e61xx.c: In function 'mv88e61xx_busychk_multic': mv88e61xx.c:46: warning: passing argument 4 of 'miiphy_read' from incompatible pointer type mv88e61xx.c: In function 'mv88e61xx_wr_phy': mv88e61xx.c:61: warning: passing argument 4 of 'miiphy_read' from incompatible pointer type mv88e61xx.c:57: warning: unused variable 'reg' mv88e61xx.c: In function 'mv88e61xx_rd_phy': mv88e61xx.c:80: warning: passing argument 4 of 'miiphy_read' from incompatible pointer type mv88e61xx.c:76: warning: unused variable 'reg' mv88e61xx.c: In function 'mv88e61xx_switch_initialize': mv88e61xx.c:346: warning: 'idstr' may be used uninitialized in this function nand_util.c:45:2: warning: #warning Please define CONFIG_SYS_64BIT_VSPRINTF for correct output! ===== sheevaplug ===== nand_util.c:45:2: warning: #warning Please define CONFIG_SYS_64BIT_VSPRINTF for correct output!
* ARM:
many things missing, still waiting for pull request from Jean-Christophe
From my point of view, ARM is currently the biggest problem. Discus-
sion of related patches often happens with big delays, pull requests always come very late. It seems obvious that Jean-Christophe is overloaded and cannot cope any more with all the ARM related work. This is actually no big surprise, as ARM related activites have significantly grown in the last months. Maybe we should split off the most active CPU families and install separate custodians for these, like we did for PowerPC long ago. OMAP and friends (TI) could be one such group, Kirkwood and friends (Marvell) could be another one. Of course we would need to find custodians for these first.
Comments and ideas welcome.
Best regards,
Wolfgang Denk