drm/msm/gpu: Capture the GPU state on a GPU hang

Capture the GPU state on a GPU hang and store it for later playback
via the devcoredump facility. Only one crash state is stored at a
time on the assumption that the first hang is usually the most
interesting. The existing crash state can be cleared after capturing
it and then a new one will be captured on the next hang.

Signed-off-by: Jordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: Rob Clark <robdclark@gmail.com>

diff --git a/drivers/gpu/drm/msm/Kconfig b/drivers/gpu/drm/msm/Kconfig
index 38cbde971b48cdc24f0f435602459d22dec58656..843a9d40c05e33f5f13d1cbb7b09c751127940f2 100644 (file)
--- a/drivers/gpu/drm/msm/Kconfig
+++ b/drivers/gpu/drm/msm/Kconfig
@@ -12,6 +12,7 @@ config DRM_MSM
        select SHMEM
        select TMPFS
        select QCOM_SCM
+       select WANT_DEV_COREDUMP
        select SND_SOC_HDMI_CODEC if SND_SOC
        select SYNC_FILE
        select PM_OPP

diff --git a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
index 4cffec2b6adcc19381f0e0113b2b493c1d6377bb..fc502e41213289dedfc5d00fbc0c11c18f40e0e4 100644 (file)
--- a/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a3xx_gpu.c
@@ -454,7 +454,7 @@ static const struct adreno_gpu_funcs funcs = {
                .active_ring = adreno_active_ring,
                .irq = a3xx_irq,
                .destroy = a3xx_destroy,
-#ifdef CONFIG_DEBUG_FS
+#if defined(CONFIG_DEBUG_FS) || defined(CONFIG_DEV_COREDUMP)
                .show = adreno_show,
 #endif
                .gpu_state_get = a3xx_gpu_state_get,

diff --git a/drivers/gpu/drm/msm/adreno/a4xx_gpu.c b/drivers/gpu/drm/msm/adreno/a4xx_gpu.c
index 95f08c22e8d7cd136a5a23b6d9d7914e24753731..8129cf037db194c025c76d96b3d95af090352fb2 100644 (file)
--- a/drivers/gpu/drm/msm/adreno/a4xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a4xx_gpu.c
@@ -540,7 +540,7 @@ static const struct adreno_gpu_funcs funcs = {
                .active_ring = adreno_active_ring,
                .irq = a4xx_irq,
                .destroy = a4xx_destroy,
-#ifdef CONFIG_DEBUG_FS
+#if defined(CONFIG_DEBUG_FS) || defined(CONFIG_DEV_COREDUMP)
                .show = adreno_show,
 #endif
                .gpu_state_get = a4xx_gpu_state_get,

diff --git a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
index 5f1aab3c1cb1189597bc8f679acffc7de02f549c..16074fa6bf1ef3878171762b06bcc73454b144cf 100644 (file)
--- a/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/a5xx_gpu.c
@@ -1243,8 +1243,10 @@ static const struct adreno_gpu_funcs funcs = {
                .active_ring = a5xx_active_ring,
                .irq = a5xx_irq,
                .destroy = a5xx_destroy,
-#ifdef CONFIG_DEBUG_FS
+#if defined(CONFIG_DEBUG_FS) || defined(CONFIG_DEV_COREDUMP)
                .show = adreno_show,
+#endif
+#if defined(CONFIG_DEBUG_FS)
                .debugfs_init = a5xx_debugfs_init,
 #endif
                .gpu_busy = a5xx_gpu_busy,

diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.c b/drivers/gpu/drm/msm/adreno/adreno_gpu.c
index 42828253ede5a2ea3e5dc18c70e0f8cb427e469d..c72e3afc43a83ed9a7962377b35d776637af0288 100644 (file)
--- a/drivers/gpu/drm/msm/adreno/adreno_gpu.c
+++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.c
@@ -378,6 +378,8 @@ struct msm_gpu_state *adreno_gpu_state_get(struct msm_gpu *gpu)
        if (!state)
                return ERR_PTR(-ENOMEM);
 
+       kref_init(&state->ref);
+
        do_gettimeofday(&state->time);
 
        for (i = 0; i < gpu->nr_rings; i++) {
@@ -414,18 +416,28 @@ struct msm_gpu_state *adreno_gpu_state_get(struct msm_gpu *gpu)
        return state;
 }
 
-void adreno_gpu_state_put(struct msm_gpu_state *state)
+static void adreno_gpu_state_destroy(struct kref *kref)
 {
-       if (IS_ERR_OR_NULL(state))
-               return;
+       struct msm_gpu_state *state = container_of(kref,
+               struct msm_gpu_state, ref);
 
+       kfree(state->comm);
+       kfree(state->cmd);
        kfree(state->registers);
        kfree(state);
 }
 
-#ifdef CONFIG_DEBUG_FS
+int adreno_gpu_state_put(struct msm_gpu_state *state)
+{
+       if (IS_ERR_OR_NULL(state))
+               return 1;
+
+       return kref_put(&state->ref, adreno_gpu_state_destroy);
+}
+
+#if defined(CONFIG_DEBUG_FS) || defined(CONFIG_DEV_COREDUMP)
 void adreno_show(struct msm_gpu *gpu, struct msm_gpu_state *state,
-               struct seq_file *m)
+               struct drm_printer *p)
 {
        struct adreno_gpu *adreno_gpu = to_adreno_gpu(gpu);
        int i;
@@ -433,23 +445,23 @@ void adreno_show(struct msm_gpu *gpu, struct msm_gpu_state *state,
        if (IS_ERR_OR_NULL(state))
                return;
 
-       seq_printf(m, "status:   %08x\n", state->rbbm_status);
-       seq_printf(m, "revision: %d (%d.%d.%d.%d)\n",
+       drm_printf(p, "status:   %08x\n", state->rbbm_status);
+       drm_printf(p, "revision: %d (%d.%d.%d.%d)\n",
                        adreno_gpu->info->revn, adreno_gpu->rev.core,
                        adreno_gpu->rev.major, adreno_gpu->rev.minor,
                        adreno_gpu->rev.patchid);
 
        for (i = 0; i < gpu->nr_rings; i++) {
-               seq_printf(m, "rb %d: fence:    %d/%d\n", i,
+               drm_printf(p, "rb %d: fence:    %d/%d\n", i,
                        state->ring[i].fence, state->ring[i].seqno);
 
-               seq_printf(m, "      rptr:     %d\n", state->ring[i].rptr);
-               seq_printf(m, "rb wptr:  %d\n", state->ring[i].wptr);
+               drm_printf(p, "      rptr:     %d\n", state->ring[i].rptr);
+               drm_printf(p, "rb wptr:  %d\n", state->ring[i].wptr);
        }
 
-       seq_printf(m, "IO:region %s 00000000 00020000\n", gpu->name);
+       drm_printf(p, "IO:region %s 00000000 00020000\n", gpu->name);
        for (i = 0; i < state->nr_registers; i++) {
-               seq_printf(m, "IO:R %08x %08x\n",
+               drm_printf(p, "IO:R %08x %08x\n",
                        state->registers[i * 2] << 2,
                        state->registers[(i * 2) + 1]);
        }

diff --git a/drivers/gpu/drm/msm/adreno/adreno_gpu.h b/drivers/gpu/drm/msm/adreno/adreno_gpu.h
index 90b6b59252afaf6c27876ae3cfc160940a6863b4..4a868aaf1a705cfe0885d7c512e7f901bc4eb235 100644 (file)
--- a/drivers/gpu/drm/msm/adreno/adreno_gpu.h
+++ b/drivers/gpu/drm/msm/adreno/adreno_gpu.h
@@ -215,9 +215,9 @@ void adreno_submit(struct msm_gpu *gpu, struct msm_gem_submit *submit,
                struct msm_file_private *ctx);
 void adreno_flush(struct msm_gpu *gpu, struct msm_ringbuffer *ring);
 bool adreno_idle(struct msm_gpu *gpu, struct msm_ringbuffer *ring);
-#ifdef CONFIG_DEBUG_FS
+#if defined(CONFIG_DEBUG_FS) || defined(CONFIG_DEV_COREDUMP)
 void adreno_show(struct msm_gpu *gpu, struct msm_gpu_state *state,
-               struct seq_file *m);
+               struct drm_printer *p);
 #endif
 void adreno_dump_info(struct msm_gpu *gpu);
 void adreno_dump(struct msm_gpu *gpu);
@@ -231,7 +231,7 @@ void adreno_gpu_cleanup(struct adreno_gpu *gpu);
 
 
 struct msm_gpu_state *adreno_gpu_state_get(struct msm_gpu *gpu);
-void adreno_gpu_state_put(struct msm_gpu_state *state);
+int adreno_gpu_state_put(struct msm_gpu_state *state);
 
 /* ringbuffer helpers (the parts that are adreno specific) */
 

diff --git a/drivers/gpu/drm/msm/msm_debugfs.c b/drivers/gpu/drm/msm/msm_debugfs.c
index c3da12179888a421c11a209ff17cc56f7b4773c8..f0da0d3c8a80f7cf9ab5082095aed6df3e3c9529 100644 (file)
--- a/drivers/gpu/drm/msm/msm_debugfs.c
+++ b/drivers/gpu/drm/msm/msm_debugfs.c
@@ -29,6 +29,7 @@ struct msm_gpu_show_priv {
 
 static int msm_gpu_show(struct seq_file *m, void *arg)
 {
+       struct drm_printer p = drm_seq_file_printer(m);
        struct msm_gpu_show_priv *show_priv = m->private;
        struct msm_drm_private *priv = show_priv->dev->dev_private;
        struct msm_gpu *gpu = priv->gpu;
@@ -38,8 +39,8 @@ static int msm_gpu_show(struct seq_file *m, void *arg)
        if (ret)
                return ret;
 
-       seq_printf(m, "%s Status:\n", gpu->name);
-       gpu->funcs->show(gpu, show_priv->state, m);
+       drm_printf(&p, "%s Status:\n", gpu->name);
+       gpu->funcs->show(gpu, show_priv->state, &p);
 
        mutex_unlock(&show_priv->dev->struct_mutex);
 

diff --git a/drivers/gpu/drm/msm/msm_gpu.c b/drivers/gpu/drm/msm/msm_gpu.c
index 03ba8872cc99bc1a31b48196b74d46500ddd9a71..5f39549d9a8b9851c7926f44d5b4cae80bd9c5a7 100644 (file)
--- a/drivers/gpu/drm/msm/msm_gpu.c
+++ b/drivers/gpu/drm/msm/msm_gpu.c
@@ -20,10 +20,11 @@
 #include "msm_mmu.h"
 #include "msm_fence.h"
 
+#include <generated/utsrelease.h>
 #include <linux/string_helpers.h>
 #include <linux/pm_opp.h>
 #include <linux/devfreq.h>
-
+#include <linux/devcoredump.h>
 
 /*
  * Power Management:
@@ -273,6 +274,81 @@ int msm_gpu_hw_init(struct msm_gpu *gpu)
        return ret;
 }
 
+#ifdef CONFIG_DEV_COREDUMP
+static ssize_t msm_gpu_devcoredump_read(char *buffer, loff_t offset,
+               size_t count, void *data, size_t datalen)
+{
+       struct msm_gpu *gpu = data;
+       struct drm_print_iterator iter;
+       struct drm_printer p;
+       struct msm_gpu_state *state;
+
+       state = msm_gpu_crashstate_get(gpu);
+       if (!state)
+               return 0;
+
+       iter.data = buffer;
+       iter.offset = 0;
+       iter.start = offset;
+       iter.remain = count;
+
+       p = drm_coredump_printer(&iter);
+
+       drm_printf(&p, "---\n");
+       drm_printf(&p, "kernel: " UTS_RELEASE "\n");
+       drm_printf(&p, "module: " KBUILD_MODNAME "\n");
+       drm_printf(&p, "time: %ld.%ld\n",
+               state->time.tv_sec, state->time.tv_usec);
+       if (state->comm)
+               drm_printf(&p, "comm: %s\n", state->comm);
+       if (state->cmd)
+               drm_printf(&p, "cmdline: %s\n", state->cmd);
+
+       gpu->funcs->show(gpu, state, &p);
+
+       msm_gpu_crashstate_put(gpu);
+
+       return count - iter.remain;
+}
+
+static void msm_gpu_devcoredump_free(void *data)
+{
+       struct msm_gpu *gpu = data;
+
+       msm_gpu_crashstate_put(gpu);
+}
+
+static void msm_gpu_crashstate_capture(struct msm_gpu *gpu, char *comm,
+               char *cmd)
+{
+       struct msm_gpu_state *state;
+
+       /* Only save one crash state at a time */
+       if (gpu->crashstate)
+               return;
+
+       state = gpu->funcs->gpu_state_get(gpu);
+       if (IS_ERR_OR_NULL(state))
+               return;
+
+       /* Fill in the additional crash state information */
+       state->comm = kstrdup(comm, GFP_KERNEL);
+       state->cmd = kstrdup(cmd, GFP_KERNEL);
+
+       /* Set the active crash state to be dumped on failure */
+       gpu->crashstate = state;
+
+       /* FIXME: Release the crashstate if this errors out? */
+       dev_coredumpm(gpu->dev->dev, THIS_MODULE, gpu, 0, GFP_KERNEL,
+               msm_gpu_devcoredump_read, msm_gpu_devcoredump_free);
+}
+#else
+static void msm_gpu_crashstate_capture(struct msm_gpu *gpu, char *comm,
+               char *cmd)
+{
+}
+#endif
+
 /*
  * Hangcheck detection for locked gpu:
  */
@@ -356,6 +432,11 @@ static void recover_worker(struct work_struct *work)
                        msm_rd_dump_submit(priv->hangrd, submit, NULL);
        }
 
+       /* Record the crash state */
+       pm_runtime_get_sync(&gpu->pdev->dev);
+       msm_gpu_crashstate_capture(gpu, comm, cmd);
+       pm_runtime_put_sync(&gpu->pdev->dev);
+
        kfree(cmd);
        kfree(comm);
 

diff --git a/drivers/gpu/drm/msm/msm_gpu.h b/drivers/gpu/drm/msm/msm_gpu.h
index d204eca8518ec8c1db0369944b00fe2366fdb0a0..b1cb44922441db071da6166ca098846129a08831 100644 (file)
--- a/drivers/gpu/drm/msm/msm_gpu.h
+++ b/drivers/gpu/drm/msm/msm_gpu.h
@@ -66,13 +66,13 @@ struct msm_gpu_funcs {
 #ifdef CONFIG_DEBUG_FS
        /* show GPU status in debugfs: */
        void (*show)(struct msm_gpu *gpu, struct msm_gpu_state *state,
-                       struct seq_file *m);
+                       struct drm_printer *p);
        /* for generation specific debugfs: */
        int (*debugfs_init)(struct msm_gpu *gpu, struct drm_minor *minor);
 #endif
        int (*gpu_busy)(struct msm_gpu *gpu, uint64_t *value);
        struct msm_gpu_state *(*gpu_state_get)(struct msm_gpu *gpu);
-       void (*gpu_state_put)(struct msm_gpu_state *state);
+       int (*gpu_state_put)(struct msm_gpu_state *state);
 };
 
 struct msm_gpu {
@@ -133,6 +133,8 @@ struct msm_gpu {
                u64 busy_cycles;
                ktime_t time;
        } devfreq;
+
+       struct msm_gpu_state *crashstate;
 };
 
 /* It turns out that all targets use the same ringbuffer size */
@@ -180,6 +182,7 @@ struct msm_gpu_submitqueue {
 };
 
 struct msm_gpu_state {
+       struct kref ref;
        struct timeval time;
 
        struct {
@@ -194,6 +197,9 @@ struct msm_gpu_state {
        u32 *registers;
 
        u32 rbbm_status;
+
+       char *comm;
+       char *cmd;
 };
 
 static inline void gpu_write(struct msm_gpu *gpu, u32 reg, u32 data)
@@ -275,4 +281,32 @@ static inline void msm_submitqueue_put(struct msm_gpu_submitqueue *queue)
                kref_put(&queue->ref, msm_submitqueue_destroy);
 }
 
+static inline struct msm_gpu_state *msm_gpu_crashstate_get(struct msm_gpu *gpu)
+{
+       struct msm_gpu_state *state = NULL;
+
+       mutex_lock(&gpu->dev->struct_mutex);
+
+       if (gpu->crashstate) {
+               kref_get(&gpu->crashstate->ref);
+               state = gpu->crashstate;
+       }
+
+       mutex_unlock(&gpu->dev->struct_mutex);
+
+       return state;
+}
+
+static inline void msm_gpu_crashstate_put(struct msm_gpu *gpu)
+{
+       mutex_lock(&gpu->dev->struct_mutex);
+
+       if (gpu->crashstate) {
+               if (gpu->funcs->gpu_state_put(gpu->crashstate))
+                       gpu->crashstate = NULL;
+       }
+
+       mutex_unlock(&gpu->dev->struct_mutex);
+}
+
 #endif /* __MSM_GPU_H__ */