Re: linux-next: build failure after merge of the amdgpu tree

From: Greg KH
Date: Thu Jun 23 2022 - 07:02:12 EST


On Thu, Jun 23, 2022 at 04:07:23PM +1000, Stephen Rothwell wrote:
> Hi all,
>
> After merging the amdgpu tree, today's linux-next build (x86_64
> allmodconfig) failed like this:
>
> In file included from include/linux/radix-tree.h:12,
> from include/linux/idr.h:15,
> from include/linux/kernfs.h:12,
> from include/linux/sysfs.h:16,
> from include/linux/kobject.h:20,
> from include/linux/energy_model.h:7,
> from include/linux/device.h:16,
> from include/linux/power_supply.h:15,
> from drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:28:
> drivers/gpu/drm/amd/amdgpu/amdgpu_device.c: In function 'amdgpu_reset_capture_coredumpm':
> include/linux/gfp.h:337:25: error: passing argument 5 of 'dev_coredumpm' makes pointer from integer without a cast [-Werror=int-conversion]
> 337 | #define GFP_KERNEL (__GFP_RECLAIM | __GFP_IO | __GFP_FS)
> | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> | |
> | unsigned int
> drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:4726:55: note: in expansion of macro 'GFP_KERNEL'
> 4726 | dev_coredumpm(dev->dev, THIS_MODULE, adev, 0, GFP_KERNEL,
> | ^~~~~~~~~~
> In file included from drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:35:
> include/linux/devcoredump.h:59:30: note: expected 'ssize_t (*)(char *, loff_t, size_t, void *, size_t)' {aka 'long int (*)(char *, long long int, long unsigned int, void *, long unsigned int)'} but argument is of type 'unsigned int'
> 59 | ssize_t (*read)(char *buffer, loff_t offset, size_t count,
> | ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> 60 | void *data, size_t datalen),
> | ~~~~~~~~~~~~~~~~~~~~~~~~~~~
> drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:4727:23: error: passing argument 6 of 'dev_coredumpm' from incompatible pointer type [-Werror=incompatible-pointer-types]
> 4727 | amdgpu_devcoredump_read, amdgpu_devcoredump_free);
> | ^~~~~~~~~~~~~~~~~~~~~~~
> | |
> | ssize_t (*)(char *, loff_t, size_t, void *, size_t) {aka long int (*)(char *, long long int, long unsigned int, void *, long unsigned int)}
> In file included from drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:35:
> include/linux/devcoredump.h:61:27: note: expected 'void (*)(void *)' but argument is of type 'ssize_t (*)(char *, loff_t, size_t, void *, size_t)' {aka 'long int (*)(char *, long long int, long unsigned int, void *, long unsigned int)'}
> 61 | void (*free)(void *data));
> | ~~~~~~~^~~~~~~~~~~~~~~~~
> drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:4726:9: error: too many arguments to function 'dev_coredumpm'
> 4726 | dev_coredumpm(dev->dev, THIS_MODULE, adev, 0, GFP_KERNEL,
> | ^~~~~~~~~~~~~
> In file included from drivers/gpu/drm/amd/amdgpu/amdgpu_device.c:35:
> include/linux/devcoredump.h:57:6: note: declared here
> 57 | void dev_coredumpm(struct device *dev, struct module *owner,
> | ^~~~~~~~~~~~~
> cc1: all warnings being treated as errors
>
> Caused by commit
>
> 3d8785f6c04a ("drm/amdgpu: adding device coredump support")
>
> interacting with commit
>
> 77515ebaf019 ("devcoredump: remove the useless gfp_t parameter in dev_coredumpv and dev_coredumpm")
>
> from the driver-core tree.
>
> I have applied the following merge resolution patch for today.
>
> From: Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx>
> Date: Thu, 23 Jun 2022 15:56:22 +1000
> Subject: [PATCH] fix up for "devcoredump: remove the useless gfp_t parameter in dev_coredumpv and dev_coredumpm"
>
> interacting with
>
> 3d8785f6c04a ("drm/amdgpu: adding device coredump support")
>
> Signed-off-by: Stephen Rothwell <sfr@xxxxxxxxxxxxxxxx>
> ---
> drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> index f2a4c268ac72..9d6418bb963e 100644
> --- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> +++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
> @@ -4723,7 +4723,7 @@ static void amdgpu_reset_capture_coredumpm(struct amdgpu_device *adev)
> struct drm_device *dev = adev_to_drm(adev);
>
> ktime_get_ts64(&adev->reset_time);
> - dev_coredumpm(dev->dev, THIS_MODULE, adev, 0, GFP_KERNEL,
> + dev_coredumpm(dev->dev, THIS_MODULE, adev, 0,
> amdgpu_devcoredump_read, amdgpu_devcoredump_free);
> }
> #endif

Ok, this isn't going to work as-is. I'll go revert this patch (and the
other one) from my tree and then request this to be done in a way that
does not break anyone going forward, if possible...

thanks,

greg k-h