[PATCH v2] dma-buf: allow nested dma_resv_reserve_fences
kernel test robot
lkp at intel.com
Fri Jun 23 21:20:48 UTC 2023
Hi Yunxiang,
kernel test robot noticed the following build errors:
url: https://github.com/intel-lab-lkp/linux/commits/UPDATE-20230624-040209/Yunxiang-Li/drm-amdgpu-fix-missing-fence-reserve-in-amdgpu_vm_sdma_commit/20230622-002915
base: the 2th patch of https://lore.kernel.org/r/20230621162652.10875-3-Yunxiang.Li%40amd.com
patch link: https://lore.kernel.org/r/20230623200113.62051-1-Yunxiang.Li%40amd.com
patch subject: [PATCH v2] dma-buf: allow nested dma_resv_reserve_fences
config: m68k-randconfig-r002-20230622 (https://download.01.org/0day-ci/archive/20230624/202306240538.Mz3kjtT7-lkp@intel.com/config)
compiler: m68k-linux-gcc (GCC) 12.3.0
reproduce: (https://download.01.org/0day-ci/archive/20230624/202306240538.Mz3kjtT7-lkp@intel.com/reproduce)
If you fix the issue in a separate patch/commit (i.e. not just a new version of
the same patch/commit), kindly add following tags
| Reported-by: kernel test robot <lkp at intel.com>
| Closes: https://lore.kernel.org/oe-kbuild-all/202306240538.Mz3kjtT7-lkp@intel.com/
All errors (new ones prefixed by >>):
drivers/dma-buf/dma-resv.c: In function 'dma_resv_add_fence':
>> drivers/dma-buf/dma-resv.c:326: error: unterminated #else
326 | #ifdef CONFIG_DEBUG_MUTEXES
|
drivers/dma-buf/dma-resv.c:327:9: note: '-Wmisleading-indentation' is disabled from this point onwards, since column-tracking was disabled due to the size of the code/headers
327 | else
| ^~~~
drivers/dma-buf/dma-resv.c:327:9: note: adding '-flarge-source-files' will allow for more column-tracking support, at the expense of compilation time and memory
>> drivers/dma-buf/dma-resv.c:328:17: error: expected declaration or statement at end of input
328 | WARN_ON(1); // missing fence slot allocation
| ^~~~~~~
vim +326 drivers/dma-buf/dma-resv.c
274
275 /**
276 * dma_resv_add_fence - Add a fence to the dma_resv obj
277 * @obj: the reservation object
278 * @fence: the fence to add
279 * @usage: how the fence is used, see enum dma_resv_usage
280 *
281 * Add a fence to a slot, @obj must be locked with dma_resv_lock(), and
282 * dma_resv_reserve_fences() has been called.
283 *
284 * See also &dma_resv.fence for a discussion of the semantics.
285 */
286 void dma_resv_add_fence(struct dma_resv *obj, struct dma_fence *fence,
287 enum dma_resv_usage usage)
288 {
289 struct dma_resv_list *fobj;
290 struct dma_fence *old;
291 unsigned int i, count;
292
293 dma_fence_get(fence);
294
295 dma_resv_assert_held(obj);
296
297 /* Drivers should not add containers here, instead add each fence
298 * individually.
299 */
300 WARN_ON(dma_fence_is_container(fence));
301
302 retry:
303 fobj = dma_resv_fences_list(obj);
304 count = fobj->num_fences;
305
306 for (i = 0; i < count; ++i) {
307 enum dma_resv_usage old_usage;
308
309 dma_resv_list_entry(fobj, i, obj, &old, &old_usage);
310 if ((old->context == fence->context && old_usage >= usage &&
311 dma_fence_is_later(fence, old)) ||
312 dma_fence_is_signaled(old)) {
313 dma_resv_list_set(fobj, i, fence, usage);
314 dma_fence_put(old);
315 return;
316 }
317 }
318
319 if (WARN_ON(fobj->num_fences == fobj->max_fences)) {
320 // try our best to avoid memory corruption
321 dma_resv_reserve_fences(obj, 1);
322 goto retry;
323 }
324 if (fobj->reserved_fences)
325 fobj->reserved_fences -= 1;
> 326 #ifdef CONFIG_DEBUG_MUTEXES
327 else
> 328 WARN_ON(1); // missing fence slot allocation
329 #else
330 count++;
331
332 dma_resv_list_set(fobj, i, fence, usage);
333 /* pointer update must be visible before we extend the num_fences */
334 smp_store_mb(fobj->num_fences, count);
335 }
336 EXPORT_SYMBOL(dma_resv_add_fence);
337
--
0-DAY CI Kernel Test Service
https://github.com/intel/lkp-tests/wiki
More information about the dri-devel
mailing list