On 11/29/2012 12:50 PM, Tom Warren wrote: > Stephen, > >> -----Original Message----- >> From: Stephen Warren [mailto:swar...@wwwdotorg.org] >> Sent: Thursday, November 29, 2012 12:10 PM >> To: Tom Warren >> Cc: Simon Glass; u-boot@lists.denx.de; Stephen Warren; Marc Dietrich; >> Thierry Reding >> Subject: Re: [PATCH 1/3] tegra: only define TEGRA_DEVICE_SETTINGS if not >> already defined >> >> On 11/29/2012 11:40 AM, Tom Warren wrote: >>> Stephen, >>> >>>> -----Original Message----- >>>> From: Stephen Warren [mailto:swar...@wwwdotorg.org] >>>> Sent: Wednesday, November 28, 2012 2:03 PM >>>> To: Simon Glass >>>> Cc: u-boot@lists.denx.de; Tom Warren; Stephen Warren; Marc Dietrich; >>>> Thierry Reding >>>> Subject: Re: [PATCH 1/3] tegra: only define TEGRA_DEVICE_SETTINGS if >>>> not already defined >>>> >>>> On 11/28/2012 02:01 PM, Simon Glass wrote: >>>>> Hi Stephen, >>>>> >>>>> On Wed, Nov 28, 2012 at 9:50 AM, Stephen Warren >>>>> <swar...@wwwdotorg.org> >>>> wrote: >>>>>> From: Stephen Warren <swar...@nvidia.com> >>>>>> >>>>>> seaboard.h attempts to undefine TEGRA_DEVICE_SETTINGS and provide a >>>>>> custom value. This worked when the "pre" included tegra20-common.h >>>>>> provided the default. However, changes in the main U-Boot repo >>>>>> removed this default from the "pre" included tegra20-common.h to >>>>>> the >>>> "post" >>>>>> included tegra-common-post.h, which uncondtionally provides the value. >>>>>> This causes the following compile warnings: >>>>>> >>>>>> In file included from /home/swarren/shared/git_wa/u- >>>> boot/include/configs/seaboard.h:129:0, >>>>>> from /home/swarren/shared/git_wa/u- >>>> boot/include/config.h:10, >>>>>> from /home/swarren/shared/git_wa/u- >>>> boot/include/common.h:37, >>>>>> from lib/asm-offsets.c:18: >>>>>> /home/swarren/shared/git_wa/u-boot/include/configs/tegra-common-post. >>>>>> h:163:0: warning: "TEGRA_DEVICE_SETTINGS" redefined >>>>>> /home/swarren/shared/git_wa/u-boot/include/configs/seaboard.h:110:0: >>>>>> note: this is the location of the previous definition >>>>>> >>>>>> Solve this by modifying tegra-common-post.h to only provide a value >>>>>> for TEGRA_DEVICE_SETTINGS if the board-specific header has not >>>>>> already provided a custom value. >>>>>> >>>>>> Signed-off-by: Stephen Warren <swar...@nvidia.com> >>>>> >>>>> These series doesn't apply to u-boot-tegra/master or /next for me, >>>>> and the last one doesn't seem to apply to u-boot/master either. Can >>>>> you please take a look, may be a timing issue. >>>> >>>> Yes, as I mentioned this problem will only exist once u-boot/master >>>> and u- boot-arm/master are merged together, so this patch series >>>> applies to the result of the merge, which will be (at least part of) >>>> the state of >>>> u-boot-tegra/* at some unspecified future time:-) >>> >>> Allen's TEGRA_DEVICE_SETTINGS patch is already in u-boot/master. >> >> Yup. >> >>> I applied it to u-boot-tegra/master, then applied your 1/3 & 2/3 >>> patches >>> (3/3 isn't needed if Allen's patch is already in). >> >> Maybe. It depends how Tom Rini does the u-boot-arm/master -> u-boot/master >> merge; u-boot-arm/master contains a patch that edits TEGRA_DEVICE_SETTINGS >> in seaboard.h (Simon's to add LCD support), and it's not obvious when >> merging those two branches that the merge result should be to remove that >> change from seaboard.h. No actually, that shouldn't be the merge result; if >> it was, the LCD additions would be removed from TEGRA_DEVICE_SETTINGS >> without patch 2/3 that I posted. That is, unless patch 2/3 of mine gets into >> u-boot-arm/master before Albert sends a pull request to Tom Rini... >> >> So, I think that once you rebase onto an upstream branch that itself >> includes all these patches rather than manually applying them, you will >> probably find patch 3/3 is required. Admittedly it's not if you just apply >> Allen's patch to the current u-boot-arm/master. > > I'm not anticipating doing a new pull request for u-boot-tegra/master > any time real soon, so we can wait for upstream (u-boot-arm & > u-boot/master) to settle a bit. But regardless, Albert has said that > he's fine with custodians submitting pull requests that contain/depend > upon patches that are already in an upstream repo, or with the custodian > stating explicitly which upstream patches are needed for a pull > request to succeed/build/work.
I believe he said that about git commits (i.e. you can branch from a commit in u-boot/master just fine) not about patches (i.e. you can't manually apply a patch that's already upstream, since that will cause the patch itself to be duplicated in git). So, if you apply Allen's patch manually to u-boot-tegra/master, you'll need to rebase u-boot-tegra/master onto u-boot/master or u-boot-arm/master once Allen's patch is there, to remove the duplicate patch, rather than just sending a pull request containing a duplicate patch. BTW, your messages have suddenly stopped being word-wrapped, which makes quoting/reading them a bit hard. _______________________________________________ U-Boot mailing list U-Boot@lists.denx.de http://lists.denx.de/mailman/listinfo/u-boot