| Message ID | 20200510174013.145466-1-sjg@chromium.org |
|---|---|
| Headers | show |
| Series | common: Further reduce the size of common.h | expand |
Hi Tom, On Sun, 10 May 2020 at 11:40, Simon Glass <sjg at chromium.org> wrote: > > This is yet another series to remove the contents of common.h so that > individual files include the headers they need. > > With this series there are only 14 includes left. > > In order to get this far the commits include a number of small clean-ups: > > - Fixing code style like 'udelay (2)' to 'udelay(1) > - Dropping typedefs so that forward struct declarations can be used > - Actually adding common.h in a few files where it is missing (thus > causing inconsistent blk.h config in EFI, for example) > > Unfortunately removing a header from common.h does not always just affect > the files that rely on that header. For example, removing net.h from > common.h created problems since net.h itself includes a lots of headers. > Rather than blindly add net.h to each of those files, I chose to add what > those files actually needed. > > Changes in v3: > - Update imxrt1020-evk, etc. > - Update stm32mp15_optee, etc. > - Update vocore2, etc. > - Rebase to master > - Addition fix for xilinx_zynqmp_mini etc. > Just wondering about the fate of this series? Regards, Simon
On Fri, May 15, 2020 at 03:02:48PM -0600, Simon Glass wrote: > Hi Tom, > > On Sun, 10 May 2020 at 11:40, Simon Glass <sjg at chromium.org> wrote: > > > > This is yet another series to remove the contents of common.h so that > > individual files include the headers they need. > > > > With this series there are only 14 includes left. > > > > In order to get this far the commits include a number of small clean-ups: > > > > - Fixing code style like 'udelay (2)' to 'udelay(1) > > - Dropping typedefs so that forward struct declarations can be used > > - Actually adding common.h in a few files where it is missing (thus > > causing inconsistent blk.h config in EFI, for example) > > > > Unfortunately removing a header from common.h does not always just affect > > the files that rely on that header. For example, removing net.h from > > common.h created problems since net.h itself includes a lots of headers. > > Rather than blindly add net.h to each of those files, I chose to add what > > those files actually needed. > > > > Changes in v3: > > - Update imxrt1020-evk, etc. > > - Update stm32mp15_optee, etc. > > - Update vocore2, etc. > > - Rebase to master > > - Addition fix for xilinx_zynqmp_mini etc. > > > > Just wondering about the fate of this series? Working my way back to it again, thanks!