Message ID | 4EEA0DF6.3050408@atmel.com |
---|---|
State | New |
Headers | show |
Hi, On Thu, Dec 15, 2011 at 7:10 AM, Nicolas Ferre <nicolas.ferre@atmel.com> wrote: > Arnd, Olof, > > This is a first "drivers" part of our material for 3.3 merge window. It collects > Ethernet drivers changes. > One thing to note is that the macb.c driver is modified to support > device tree. We agreed with David to drag those modifications into > arm-soc git tree because they go on top of Jamie's work on macb-gem driver > already in your drivers/macb-gem branch. > > This pull request goes on top of Linus' 3.2-rc5 commit but with additional > dependencies with two of your branches: > 1/ fixes (which seems already included in 3.2-rc5) > 2/ drivers/macb-gem > > I had to resolve a conflict while merging the macb-gem branche with current 3.2-rc5. The > resolution is commited into this branch. > Tell me if you need more precision. The history of this branch due to the double dependencies gets a bit awkward. Also, it seems like you in one of your merges kept the 'default y' in NET_CADENCE that should have been removed. Would you mind trying this instead? * Base this new branch on the previous macb-gem branch * Cherry-pick the fixes from 'fixes' into the new branch if truly needed -- I tested here and all of the patches applied cleanly. * Then apply these patches * For build testing before sending, merge into for-next to aggregate with the rest of the 'fixes' changes -- it's OK that we have to do a fixup here. That should give a clean branch with neat history for just these changes. About 'fixes': Since they are in for-next, and will be in 3.2, that dependency will be taken care of that way. Then send a fresh pull request for that new branch, please. Thanks! -Olof
On 12/16/2011 06:53 AM, Olof Johansson : > Hi, > > On Thu, Dec 15, 2011 at 7:10 AM, Nicolas Ferre <nicolas.ferre@atmel.com> wrote: >> Arnd, Olof, >> >> This is a first "drivers" part of our material for 3.3 merge window. It collects >> Ethernet drivers changes. >> One thing to note is that the macb.c driver is modified to support >> device tree. We agreed with David to drag those modifications into >> arm-soc git tree because they go on top of Jamie's work on macb-gem driver >> already in your drivers/macb-gem branch. >> >> This pull request goes on top of Linus' 3.2-rc5 commit but with additional >> dependencies with two of your branches: >> 1/ fixes (which seems already included in 3.2-rc5) >> 2/ drivers/macb-gem >> >> I had to resolve a conflict while merging the macb-gem branche with current 3.2-rc5. The >> resolution is commited into this branch. >> Tell me if you need more precision. > > The history of this branch due to the double dependencies gets a bit > awkward. Also, it seems like you in one of your merges kept the > 'default y' in NET_CADENCE that should have been removed. Well, no, this line has been added into "eccab1ec" and is in mainline since 3.2-rc3. This line have to be kept and the merge conflict is triggered during the merge of macb-gem on top of Linus' tree. BTW, this merge is already performed properly in linux-next. > Would you mind trying this instead? > > * Base this new branch on the previous macb-gem branch > * Cherry-pick the fixes from 'fixes' into the new branch if truly > needed -- I tested here and all of the patches applied cleanly. > * Then apply these patches > * For build testing before sending, merge into for-next to aggregate > with the rest of the 'fixes' changes -- it's OK that we have to do a > fixup here. > > That should give a clean branch with neat history for just these > changes. About 'fixes': Since they are in for-next, and will be in > 3.2, that dependency will be taken care of that way. > > Then send a fresh pull request for that new branch, please. Ok, I will base my "drivers" branch on top of macb-gem one and send you another pull request. Thanks, bye,
On Friday 16 December 2011, Nicolas Ferre wrote: > This is the version 2 of the pull request related to Ethernet drivers for AT91. > As you requested, it is based on arm-soc/drivers/macb-gem branch: > > The following changes since commit 0116da4fcc1ae8a80d9002441e98768f2a6fa2fe: > > macb: allow GEM to have configurable receive buffer size (2011-11-22 15:21:20 +0000) > > are available in the git repository at: > git://github.com/at91linux/linux-at91.git for-arnd-3.3-driversV2 > > Jean-Christophe PLAGNIOL-VILLARD (1): > net/macb: add DT support for Cadence macb/gem driver > > Nicolas Ferre (2): > ARM: at91/net: add macb ethernet controller in 9g45/9g20 DT > net/at91_ether: use gpio_is_valid for phy IRQ line Pulled into next/drivers, thanks for the quick update! Arnd
Nicolas, Jean-Christophe, On Fri, Dec 16, 2011 at 4:24 AM, Nicolas Ferre <nicolas.ferre@atmel.com> wrote: > Jean-Christophe PLAGNIOL-VILLARD (1): > net/macb: add DT support for Cadence macb/gem driver This patch breaks build for non-OF configs. I hit this at least with afeb9260_defconfig, at91cap9_defconfig and at91sam9263_defconfig. Still building other defconfigs: drivers/net/ethernet/cadence/macb.c:1503: error: implicit declaration of function 'of_match_ptr' drivers/net/ethernet/cadence/macb.c:1503: error: 'macb_dt_ids' undeclared here (not in a function) drivers/net/ethernet/cadence/macb.c:1503: error: initializer element is not constant drivers/net/ethernet/cadence/macb.c:1503: error: (near initialization for 'macb_driver.driver.of_match_table') -Olof