linux-next: ahci_platform broken build

linux-next: ahci_platform broken build

[Tim Gardner]

Gents  - while reviewing some patches for arm64 in Ubuntu I noticed that 
there is at least one patch sequence in linux-next that breaks the build.

156c5887948cd191417f18026aab9ce26e5a95da ahci-platform: Add support for 
devices with more then 1 clock
039ece38da45f5e6a94be3aa7611cf3634bc2461 libahci: Allow drivers to 
override start_engine

For example:

git reset --hard 156c5887948cd191417f18026aab9ce26e5a95da
echo "CONFIG_SATA_AHCI_PLATFORM=y" >> .config
make oldconfig scripts prepare
make M=drivers/ata

   WARNING: Symbol version dump /home/rtg/linux/linux-next/Module.symvers
            is missing; modules will have no dependencies and modversions.

   CC      drivers/ata/libata-core.o
   CC      drivers/ata/libata-scsi.o
   CC      drivers/ata/libata-eh.o
   CC      drivers/ata/libata-transport.o
   CC      drivers/ata/libata-sff.o
   CC      drivers/ata/libata-pmp.o
   CC      drivers/ata/libata-acpi.o
   LD      drivers/ata/libata.o
   CC      drivers/ata/ahci.o
   CC      drivers/ata/libahci.o
   CC      drivers/ata/ahci_platform.o
drivers/ata/ahci_platform.c: In function ‘ahci_probe’:
drivers/ata/ahci_platform.c:209:2: error: implicit declaration of 
function ‘ahci_enable_clks’ [-Werror=implicit-function-declaration]
drivers/ata/ahci_platform.c:293:2: error: implicit declaration of 
function ‘ahci_disable_clks’ [-Werror=implicit-function-declaration]
cc1: some warnings being treated as errors
make[1]: *** [drivers/ata/ahci_platform.o] Error 1
make: *** [_module_drivers/ata] Error 2

The compile problem is then fixed in 
96a01ba52c60fdd74dd6e8cf06645d06515b1396 (ahci-platform: Add enable_ / 
disable_resources helper functions) which is not really kosher.

It seems like you ought to fix this before the 3.15 merge window opens.

rtg
-- 
Tim Gardner tim.gardner@canonical.com

Re: linux-next: ahci_platform broken build

[Hans de Goede]

Hi,

On 03/16/2014 03:05 PM, Tim Gardner wrote:
> Gents  - while reviewing some patches for arm64 in Ubuntu I noticed that there is at least one patch sequence in linux-next that breaks the build.
> 
> 156c5887948cd191417f18026aab9ce26e5a95da ahci-platform: Add support for devices with more then 1 clock
> 039ece38da45f5e6a94be3aa7611cf3634bc2461 libahci: Allow drivers to override start_engine
> 
> For example:
> 
> git reset --hard 156c5887948cd191417f18026aab9ce26e5a95da
> echo "CONFIG_SATA_AHCI_PLATFORM=y" >> .config
> make oldconfig scripts prepare
> make M=drivers/ata
> 
>   WARNING: Symbol version dump /home/rtg/linux/linux-next/Module.symvers
>            is missing; modules will have no dependencies and modversions.
> 
>   CC      drivers/ata/libata-core.o
>   CC      drivers/ata/libata-scsi.o
>   CC      drivers/ata/libata-eh.o
>   CC      drivers/ata/libata-transport.o
>   CC      drivers/ata/libata-sff.o
>   CC      drivers/ata/libata-pmp.o
>   CC      drivers/ata/libata-acpi.o
>   LD      drivers/ata/libata.o
>   CC      drivers/ata/ahci.o
>   CC      drivers/ata/libahci.o
>   CC      drivers/ata/ahci_platform.o
> drivers/ata/ahci_platform.c: In function ‘ahci_probe’:
> drivers/ata/ahci_platform.c:209:2: error: implicit declaration of function ‘ahci_enable_clks’ [-Werror=implicit-function-declaration]
> drivers/ata/ahci_platform.c:293:2: error: implicit declaration of function ‘ahci_disable_clks’ [-Werror=implicit-function-declaration]
> cc1: some warnings being treated as errors
> make[1]: *** [drivers/ata/ahci_platform.o] Error 1
> make: *** [_module_drivers/ata] Error 2
> 
> The compile problem is then fixed in 96a01ba52c60fdd74dd6e8cf06645d06515b1396 (ahci-platform: Add enable_ / disable_resources helper functions) which is not really kosher.

Ah good catch, at some point I decided to prefix all the ahci_* functions defined in ahci_platform.c
with ahci_platform_ instead of just ahci_. Somehow the functions got added with their new names
directly in the "ahci-platform: Add support for devices with more then 1 clock" patch (this is
intentional), but the calls added in that patch kept using the old names. Then in the
"ahci-platform: Add enable_ / disable_resources helper functions" patch
the calls get moved to another place and now show up with the proper ahci_platform_*
names. My bad.

Tejun it is probably easiest if you fix this up directly in your tree, let me know if you
want a fixed up resend of the 2 patches in question instead.

Regards,

Hans

Re: linux-next: ahci_platform broken build

[Tejun Heo]

On Sun, Mar 16, 2014 at 03:13:55PM +0100, Hans de Goede wrote:
> Tejun it is probably easiest if you fix this up directly in your tree, let me know if you
> want a fixed up resend of the 2 patches in question instead.

It already happened and got resolved in the span of few patches.  Shit
happens.  Let's just try to test each step better next time.

Thanks.

-- 
tejun