[kbuild-all] [drm-drm-intel:drm-intel-next-queued 9/12] drivers/gpu/drm/i915/i915_perf.c:386 i915_oa_config_release() warn: passing zero to 'PTR_ERR'

tree: git://anongit.freedesktop.org/drm/drm-intel drm-intel-next-queued head: 5f5c382ecfdd06e17316d1c9f1362522c20cdfef commit: 6a45008ab7bb5e13b543de0c141b94aaa71d8397 [9/12] drm/i915/perf: allow for CS OA configs to be created lazily If you fix the issue, kindly add following tag Reported-by: kbuild test robot <lkp(a)intel.com&gt; Reported-by: Dan Carpenter <dan.carpenter(a)oracle.com&gt; New smatch warnings: drivers/gpu/drm/i915/i915_perf.c:386 i915_oa_config_release() warn: passing zero to 'PTR_ERR' Old smatch warnings: drivers/gpu/drm/i915/i915_perf.c:388 i915_oa_config_release() warn: passing zero to 'PTR_ERR' drivers/gpu/drm/i915/i915_perf.c:390 i915_oa_config_release() warn: passing zero to 'PTR_ERR' drivers/gpu/drm/i915/i915_perf.c:3420 i915_perf_add_config_ioctl() error: double unlock 'mutex:&perf->metrics_lock' drivers/gpu/drm/i915/i915_perf.c:3427 i915_perf_add_config_ioctl() error: double unlock 'mutex:&perf->metrics_lock' drivers/gpu/drm/i915/i915_perf.c:3481 i915_perf_remove_config_ioctl() error: double unlock 'mutex:&perf->metrics_lock' drivers/gpu/drm/i915/i915_perf.c:3490 i915_perf_remove_config_ioctl() error: double unlock 'mutex:&perf->metrics_lock' git remote add drm-drm-intel git://anongit.freedesktop.org/drm/drm-intel git remote update drm-drm-intel git checkout 6a45008ab7bb5e13b543de0c141b94aaa71d8397 vim +/PTR_ERR +386 drivers/gpu/drm/i915/i915_perf.c 6a45008ab7bb5e Lionel Landwerlin 2019-10-12 381 void i915_oa_config_release(struct kref *ref) f89823c212246d Lionel Landwerlin 2017-08-03 382 { 6a45008ab7bb5e Lionel Landwerlin 2019-10-12 383 struct i915_oa_config *oa_config = 6a45008ab7bb5e Lionel Landwerlin 2019-10-12 384 container_of(ref, typeof(*oa_config), ref); 6a45008ab7bb5e Lionel Landwerlin 2019-10-12 385 f89823c212246d Lionel Landwerlin 2017-08-03 @386 if (!PTR_ERR(oa_config->flex_regs)) These should be IS_ERR() instead of PTR_ERR(). I haven't looked at the conext outside of this email but it feels like a design mistake that it's possible for these to be error pointers anyway. f89823c212246d Lionel Landwerlin 2017-08-03 387 kfree(oa_config->flex_regs); f89823c212246d Lionel Landwerlin 2017-08-03 388 if (!PTR_ERR(oa_config->b_counter_regs)) f89823c212246d Lionel Landwerlin 2017-08-03 389 kfree(oa_config->b_counter_regs); f89823c212246d Lionel Landwerlin 2017-08-03 390 if (!PTR_ERR(oa_config->mux_regs)) f89823c212246d Lionel Landwerlin 2017-08-03 391 kfree(oa_config->mux_regs); f89823c212246d Lionel Landwerlin 2017-08-03 392 6a45008ab7bb5e Lionel Landwerlin 2019-10-12 393 kfree_rcu(oa_config, rcu); f89823c212246d Lionel Landwerlin 2017-08-03 394 } --- 0-DAY kernel test infrastructure Open Source Technology Center https://lists.01.org/pipermail/kbuild-all Intel Corporation