Re: [PATCH] drm/ttm: ioremap buffer according to TTM mem caching setting

[Christian König <ckoenig.leichtzumerken@xxxxxxxxx>]

    Hi Oak,
    
    as far as I know some architectures like PowerPC/ARM/MIPS need that.
    And we at least officially support PowerPC and ARM and MIPS is best
    effort and shouldn't break if possible.
    
    Thomas just recently had a whole bunch of DMA-buf patches to also
    fix that up for DMA-bufs vmap as well, previously we have just been
    casting away the __iomem and hoped that it always works.
    
    Regards,
    Christian.

Am 03.03.21 um 21:59 schrieb Zeng, Oak:

[AMD Official Use Only - Internal Distribution Only]

Hi Christian,

 

Can you explain why __iomem annotation is mandatory for amdgpu driver? If this is the case, we can’t switch to memremap. The only fix seems to me is add a #ifdef __x86_64__ to the ioremap_cache codes.

 

Regards,

Oak

 

From: Christian König <ckoenig.leichtzumerken@xxxxxxxxx>
Sent: Wednesday, March 3, 2021 5:46 AM
To: Zeng, Oak <Oak.Zeng@xxxxxxx>; amd-gfx@xxxxxxxxxxxxxxxxxxxxx; dri-devel@xxxxxxxxxxxxxxxxxxxxx; Daniel Vetter <daniel@xxxxxxxx>; Dave Airlie <airlied@xxxxxxxxxx>; Thomas Hellström (Intel) <thomas_os@xxxxxxxxxxxx>; dan.j.williams@xxxxxxxxx
Cc: kbuild-all@xxxxxxxxxxxx; Kuehling, Felix <Felix.Kuehling@xxxxxxx>; Kasiviswanathan, Harish <Harish.Kasiviswanathan@xxxxxxx>; Deucher, Alexander <Alexander.Deucher@xxxxxxx>; Huang, JinHuiEric <JinHuiEric.Huang@xxxxxxx>; Koenig, Christian <Christian.Koenig@xxxxxxx>
Subject: Re: [PATCH] drm/ttm: ioremap buffer according to TTM mem caching setting

 

Hi Oak,

config: parisc-randconfig-r012-20210302 (attached as .config)

It's not the Intel driver build which fails here, but the build bot is just hosted by Intel.

The problem is that the parisc architecture doesn't defines the ioremap_cache() function.

I've looked at using memremap() instead of ioremap_cache(). The problem is that we do support architectures with the TTM as well as amndgpu code where the __iomem annotation is mandatory and correct.

Regards,
Christian.

Am 02.03.21 um 23:45 schrieb Zeng, Oak:

[AMD Official Use Only - Internal Distribution Only]

 

Hi Daniel, Thomas, Dan,

 

Does below message mean the calling ioremap_cache failed intel’s driver build? I can see both ioremap_cache and ioremap_wc are defined in arch/x86/mm/ioremap.c – why ioremap_wc doesn’t break intel driver’s build?

 

Are we supposed to use memremap (offset, size, MEMREMAP_WB) to replace ioremap_cache? When I read here https://lwn.net/Articles/653585/ I felt that ioremap_cache returns an address annotated with _iomem while memremap returns an address without __iomem annotation. In our use case, GPU memory is treated as UEFI SPM (specific purpose memory). I am not very sure whether memremap (thus no __iomem annotation) is the right thing to do. What I am sure is, we have tested ioremap_cache and it works fine on AMD system.

 

I will send out a test patch replacing ioremap_cache with ioremap_wc, to trigger Intel build robot to see whether it fails Intel build. I suppose it will not fail Intel build.

 

Regards,

Oak

 

From: Christian König <ckoenig.leichtzumerken@xxxxxxxxx>
Sent: Tuesday, March 2, 2021 6:31 AM
To: amd-gfx@xxxxxxxxxxxxxxxxxxxxx; dri-devel@xxxxxxxxxxxxxxxxxxxxx; Daniel Vetter <daniel@xxxxxxxx>; Dave Airlie <airlied@xxxxxxxxxx>; Thomas Hellström (Intel) <thomas_os@xxxxxxxxxxxx>
Cc: Zeng, Oak <Oak.Zeng@xxxxxxx>; kbuild-all@xxxxxxxxxxxx; Kuehling, Felix <Felix.Kuehling@xxxxxxx>; Kasiviswanathan, Harish <Harish.Kasiviswanathan@xxxxxxx>; Deucher, Alexander <Alexander.Deucher@xxxxxxx>; Huang, JinHuiEric <JinHuiEric.Huang@xxxxxxx>; Koenig, Christian <Christian.Koenig@xxxxxxx>
Subject: Re: [PATCH] drm/ttm: ioremap buffer according to TTM mem caching setting

 

Hi guys,

adding the usual suspects direct. Does anybody of hand know how to check if an architecture supports ioremap_cache()?

For now we only need this on X86, but I would feel better if we don't use an #ifdef here.

Regards,
Christian.

Am 02.03.21 um 05:12 schrieb kernel test robot:

Hi Oak,

Thank you for the patch! Yet something to improve:

[auto build test ERROR on drm-intel/for-linux-next]

[also build test ERROR on drm-tip/drm-tip linus/master v5.12-rc1 next-20210302]

[cannot apply to tegra-drm/drm/tegra/for-next drm-exynos/exynos-drm-next drm/drm-next]

[If your patch is applied to the wrong git tree, kindly drop us a note.

And when submitting patch, we suggest to use '--base' as documented in

https://git-scm.com/docs/git-format-patch]

url:    https://github.com/0day-ci/linux/commits/Oak-Zeng/drm-ttm-ioremap-buffer-according-to-TTM-mem-caching-setting/20210302-064500

base:   git://anongit.freedesktop.org/drm-intel for-linux-next

config: parisc-randconfig-r012-20210302 (attached as .config)

compiler: hppa-linux-gcc (GCC) 9.3.0

reproduce (this is a W=1 build):

        wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross

        chmod +x ~/bin/make.cross

        # https://github.com/0day-ci/linux/commit/225bb3711439ec559dd72ae5af8e62d34ea60a64

        git remote add linux-review https://github.com/0day-ci/linux

        git fetch --no-tags linux-review Oak-Zeng/drm-ttm-ioremap-buffer-according-to-TTM-mem-caching-setting/20210302-064500

        git checkout 225bb3711439ec559dd72ae5af8e62d34ea60a64

        # save the attached .config to linux build tree

        COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-9.3.0 make.cross ARCH=parisc 

If you fix the issue, kindly add following tag as appropriate

Reported-by: kernel test robot <lkp@xxxxxxxxx>

All errors (new ones prefixed by >>):

   drivers/gpu/drm/ttm/ttm_bo_util.c: In function 'ttm_resource_ioremap':

drivers/gpu/drm/ttm/ttm_bo_util.c:95:11: error: implicit declaration of function 'ioremap_cache'; did you mean 'ioremap_uc'? [-Werror=implicit-function-declaration]

      95 |    addr = ioremap_cache(mem->bus.offset, bus_size);

         |           ^~~~~~~~~~~~~

         |           ioremap_uc

   drivers/gpu/drm/ttm/ttm_bo_util.c:95:9: warning: assignment to 'void *' from 'int' makes pointer from integer without a cast [-Wint-conversion]

      95 |    addr = ioremap_cache(mem->bus.offset, bus_size);

         |         ^

   drivers/gpu/drm/ttm/ttm_bo_util.c: In function 'ttm_bo_ioremap':

   drivers/gpu/drm/ttm/ttm_bo_util.c:379:17: warning: assignment to 'void *' from 'int' makes pointer from integer without a cast [-Wint-conversion]

     379 |    map->virtual = ioremap_cache(bo->mem.bus.offset + offset,

         |                 ^

   drivers/gpu/drm/ttm/ttm_bo_util.c: In function 'ttm_bo_vmap':

   drivers/gpu/drm/ttm/ttm_bo_util.c:500:16: warning: assignment to 'void *' from 'int' makes pointer from integer without a cast [-Wint-conversion]

     500 |    vaddr_iomem = ioremap_cache(mem->bus.offset,

         |                ^

   cc1: some warnings being treated as errors

vim +95 drivers/gpu/drm/ttm/ttm_bo_util.c

    74  

    75  static int ttm_resource_ioremap(struct ttm_bo_device *bdev,

    76                                struct ttm_resource *mem,

    77                                void **virtual)

    78  {

    79          int ret;

    80          void *addr;

    81  

    82          *virtual = NULL;

    83          ret = ttm_mem_io_reserve(bdev, mem);

    84          if (ret || !mem->bus.is_iomem)

    85                  return ret;

    86  

    87          if (mem->bus.addr) {

    88                  addr = mem->bus.addr;

    89          } else {

    90                  size_t bus_size = (size_t)mem->num_pages << PAGE_SHIFT;

    91  

    92                  if (mem->bus.caching == ttm_write_combined)

    93                         addr = ioremap_wc(mem->bus.offset, bus_size);

    94                  else if (mem->bus.caching == ttm_cached)

  > 95                         addr = ioremap_cache(mem->bus.offset, bus_size);

    96                  else

    97                         addr = ioremap(mem->bus.offset, bus_size);

    98                  if (!addr) {

    99                         ttm_mem_io_free(bdev, mem);

   100                         return -ENOMEM;

   101                  }

   102          }

   103          *virtual = addr;

   104          return 0;

   105  }

   106  

---

0-DAY CI Kernel Test Service, Intel Corporation

https://lists.01.org/hyperkitty/list/kbuild-all@xxxxxxxxxxxx

_______________________________________________

amd-gfx mailing list

amd-gfx@xxxxxxxxxxxxxxxxxxxxx

https://lists.freedesktop.org/mailman/listinfo/amd-gfx

 

 

_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/dri-devel