Hi
Am 05.06.20 um 18:22 schrieb Daniel Vetter:
> On Fri, Jun 05, 2020 at 03:57:56PM +0200, Thomas Zimmermann wrote:
>> The memory-management code now cleans up automatically as part of
>> device destruction.
>>
>> Signed-off-by: Thomas Zimmermann <tzimmerm...@suse.de>
>> ---
>> drivers/gpu/drm/mgag200/mgag200_drv.h | 1 -
>> drivers/gpu/drm/mgag200/mgag200_main.c | 5 +----
>> drivers/gpu/drm/mgag200/mgag200_mm.c | 28 ++++++++++++++------------
>> 3 files changed, 16 insertions(+), 18 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/mgag200/mgag200_drv.h
>> b/drivers/gpu/drm/mgag200/mgag200_drv.h
>> index cd786ffe319b8..7b6e6827a9a21 100644
>> --- a/drivers/gpu/drm/mgag200/mgag200_drv.h
>> +++ b/drivers/gpu/drm/mgag200/mgag200_drv.h
>> @@ -198,6 +198,5 @@ void mgag200_i2c_destroy(struct mga_i2c_chan *i2c);
>>
>> /* mgag200_mm.c */
>> int mgag200_mm_init(struct mga_device *mdev);
>> -void mgag200_mm_fini(struct mga_device *mdev);
>>
>> #endif /* __MGAG200_DRV_H__ */
>> diff --git a/drivers/gpu/drm/mgag200/mgag200_main.c
>> b/drivers/gpu/drm/mgag200/mgag200_main.c
>> index e9ad783c2b44d..49bcdfcb40a4e 100644
>> --- a/drivers/gpu/drm/mgag200/mgag200_main.c
>> +++ b/drivers/gpu/drm/mgag200/mgag200_main.c
>> @@ -57,13 +57,11 @@ int mgag200_driver_load(struct drm_device *dev, unsigned
>> long flags)
>> ret = mgag200_modeset_init(mdev);
>> if (ret) {
>> drm_err(dev, "Fatal error during modeset init: %d\n", ret);
>> - goto err_mgag200_mm_fini;
>> + goto err_mm;
>> }
>>
>> return 0;
>>
>> -err_mgag200_mm_fini:
>> - mgag200_mm_fini(mdev);
>> err_mm:
>> dev->dev_private = NULL;
>> return ret;
>> @@ -75,6 +73,5 @@ void mgag200_driver_unload(struct drm_device *dev)
>>
>> if (mdev == NULL)
>> return;
>> - mgag200_mm_fini(mdev);
>> dev->dev_private = NULL;
>> }
>> diff --git a/drivers/gpu/drm/mgag200/mgag200_mm.c
>> b/drivers/gpu/drm/mgag200/mgag200_mm.c
>> index f56b0456994f4..1b1e5ec5d1ceb 100644
>> --- a/drivers/gpu/drm/mgag200/mgag200_mm.c
>> +++ b/drivers/gpu/drm/mgag200/mgag200_mm.c
>> @@ -28,6 +28,8 @@
>>
>> #include <linux/pci.h>
>>
>> +#include <drm/drm_managed.h>
>> +
>> #include "mgag200_drv.h"
>>
>> static size_t mgag200_probe_vram(struct mga_device *mdev, void __iomem *mem,
>> @@ -73,6 +75,18 @@ static size_t mgag200_probe_vram(struct mga_device *mdev,
>> void __iomem *mem,
>> return offset - 65536;
>> }
>>
>> +static void mgag200_mm_release(struct drm_device *dev, void *ptr)
>> +{
>> + struct mga_device *mdev = to_mga_device(dev);
>> +
>> + mdev->vram_fb_available = 0;
>> + iounmap(mdev->vram);
>> + arch_io_free_memtype_wc(pci_resource_start(dev->pdev, 0),
>> + pci_resource_len(dev->pdev, 0));
>> + arch_phys_wc_del(mdev->fb_mtrr);
>> + mdev->fb_mtrr = 0;
>> +}
>> +
>> int mgag200_mm_init(struct mga_device *mdev)
>> {
>> struct drm_device *dev = mdev->dev;
>> @@ -104,22 +118,10 @@ int mgag200_mm_init(struct mga_device *mdev)
>>
>> mdev->vram_fb_available = mdev->mc.vram_size;
>>
>> - return 0;
>> + return drmm_add_action_or_reset(dev, mgag200_mm_release, NULL);
>>
>> err_arch_phys_wc_del:
>> arch_phys_wc_del(mdev->fb_mtrr);
>> arch_io_free_memtype_wc(start, len);
>
> Btw I think devm versions of these two would benefit a bunch of drivers in
> their cleanup code. devm_ not drmm_ since it's hw resource cleanup. In
Yup, I was looking and couldn't find them. Their non-existence is the
reason for the release function. Having them would make sense.
> case you ever run out of ideas :-)
I already have patches for further mgag200 mode-setting cleanups and ast
managed initialization. So no shortage of idea here. :)
Best regards
Thomas
>
> Cheeres, Daniel
>
>> return ret;
>> }
>> -
>> -void mgag200_mm_fini(struct mga_device *mdev)
>> -{
>> - struct drm_device *dev = mdev->dev;
>> -
>> - mdev->vram_fb_available = 0;
>> - iounmap(mdev->vram);
>> - arch_io_free_memtype_wc(pci_resource_start(dev->pdev, 0),
>> - pci_resource_len(dev->pdev, 0));
>> - arch_phys_wc_del(mdev->fb_mtrr);
>> - mdev->fb_mtrr = 0;
>> -}
>> --
>> 2.26.2
>>
>
--
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5, 90409 Nürnberg, Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Felix Imendörffer
signature.asc
Description: OpenPGP digital signature
_______________________________________________ dri-devel mailing list dri-devel@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/dri-devel
