Commit 1df4289d authored by Sharat Masetty's avatar Sharat Masetty Committed by Rob Clark

drm/msm: Optimize adreno_show_object()

When the userspace tries to read the crashstate dump, the read side
implementation in the driver currently ascii85 encodes all the binary
buffers and it does this each time the read system call is called.
A userspace tool like cat typically does a page by page read and the
number of read calls depends on the size of the data captured by the
driver. This is certainly not desirable and does not scale well with
large captures.

This patch encodes the buffer only once in the read path. With this there
is an immediate >10X speed improvement in crashstate save time.
Signed-off-by: default avatarSharat Masetty <smasetty@codeaurora.org>
Reviewed-by: default avatarJordan Crouse <jcrouse@codeaurora.org>
Signed-off-by: default avatarRob Clark <robdclark@gmail.com>
parent 84c61275
...@@ -398,7 +398,7 @@ int adreno_gpu_state_get(struct msm_gpu *gpu, struct msm_gpu_state *state) ...@@ -398,7 +398,7 @@ int adreno_gpu_state_get(struct msm_gpu *gpu, struct msm_gpu_state *state)
size = j + 1; size = j + 1;
if (size) { if (size) {
state->ring[i].data = kmalloc(size << 2, GFP_KERNEL); state->ring[i].data = kvmalloc(size << 2, GFP_KERNEL);
if (state->ring[i].data) { if (state->ring[i].data) {
memcpy(state->ring[i].data, gpu->rb[i]->start, size << 2); memcpy(state->ring[i].data, gpu->rb[i]->start, size << 2);
state->ring[i].data_size = size << 2; state->ring[i].data_size = size << 2;
...@@ -441,7 +441,7 @@ void adreno_gpu_state_destroy(struct msm_gpu_state *state) ...@@ -441,7 +441,7 @@ void adreno_gpu_state_destroy(struct msm_gpu_state *state)
int i; int i;
for (i = 0; i < ARRAY_SIZE(state->ring); i++) for (i = 0; i < ARRAY_SIZE(state->ring); i++)
kfree(state->ring[i].data); kvfree(state->ring[i].data);
for (i = 0; state->bos && i < state->nr_bos; i++) for (i = 0; state->bos && i < state->nr_bos; i++)
kvfree(state->bos[i].data); kvfree(state->bos[i].data);
...@@ -471,34 +471,74 @@ int adreno_gpu_state_put(struct msm_gpu_state *state) ...@@ -471,34 +471,74 @@ int adreno_gpu_state_put(struct msm_gpu_state *state)
#if defined(CONFIG_DEBUG_FS) || defined(CONFIG_DEV_COREDUMP) #if defined(CONFIG_DEBUG_FS) || defined(CONFIG_DEV_COREDUMP)
static void adreno_show_object(struct drm_printer *p, u32 *ptr, int len) static char *adreno_gpu_ascii85_encode(u32 *src, size_t len)
{ {
void *buf;
size_t buf_itr = 0, buffer_size;
char out[ASCII85_BUFSZ]; char out[ASCII85_BUFSZ];
long l, datalen, i; long l;
int i;
if (!src || !len)
return NULL;
l = ascii85_encode_len(len);
/*
* Ascii85 outputs either a 5 byte string or a 1 byte string. So we
* account for the worst case of 5 bytes per dword plus the 1 for '\0'
*/
buffer_size = (l * 5) + 1;
buf = kvmalloc(buffer_size, GFP_KERNEL);
if (!buf)
return NULL;
for (i = 0; i < l; i++)
buf_itr += snprintf(buf + buf_itr, buffer_size - buf_itr, "%s",
ascii85_encode(src[i], out));
return buf;
}
if (!ptr || !len) /* len is expected to be in bytes */
static void adreno_show_object(struct drm_printer *p, void **ptr, int len,
bool *encoded)
{
if (!*ptr || !len)
return; return;
if (!*encoded) {
long datalen, i;
u32 *buf = *ptr;
/*
* Only dump the non-zero part of the buffer - rarely will
* any data completely fill the entire allocated size of
* the buffer.
*/
for (datalen = 0, i = 0; i < len >> 2; i++)
if (buf[i])
datalen = ((i + 1) << 2);
/* /*
* Only dump the non-zero part of the buffer - rarely will any data * If we reach here, then the originally captured binary buffer
* completely fill the entire allocated size of the buffer * will be replaced with the ascii85 encoded string
*/ */
for (datalen = 0, i = 0; i < len >> 2; i++) { *ptr = adreno_gpu_ascii85_encode(buf, datalen);
if (ptr[i])
datalen = (i << 2) + 1; kvfree(buf);
*encoded = true;
} }
/* Skip printing the object if it is empty */ if (!*ptr)
if (datalen == 0)
return; return;
l = ascii85_encode_len(datalen);
drm_puts(p, " data: !!ascii85 |\n"); drm_puts(p, " data: !!ascii85 |\n");
drm_puts(p, " "); drm_puts(p, " ");
for (i = 0; i < l; i++) drm_puts(p, *ptr);
drm_puts(p, ascii85_encode(ptr[i], out));
drm_puts(p, "\n"); drm_puts(p, "\n");
} }
...@@ -530,8 +570,8 @@ void adreno_show(struct msm_gpu *gpu, struct msm_gpu_state *state, ...@@ -530,8 +570,8 @@ void adreno_show(struct msm_gpu *gpu, struct msm_gpu_state *state,
drm_printf(p, " wptr: %d\n", state->ring[i].wptr); drm_printf(p, " wptr: %d\n", state->ring[i].wptr);
drm_printf(p, " size: %d\n", MSM_GPU_RINGBUFFER_SZ); drm_printf(p, " size: %d\n", MSM_GPU_RINGBUFFER_SZ);
adreno_show_object(p, state->ring[i].data, adreno_show_object(p, &state->ring[i].data,
state->ring[i].data_size); state->ring[i].data_size, &state->ring[i].encoded);
} }
if (state->bos) { if (state->bos) {
...@@ -542,8 +582,8 @@ void adreno_show(struct msm_gpu *gpu, struct msm_gpu_state *state, ...@@ -542,8 +582,8 @@ void adreno_show(struct msm_gpu *gpu, struct msm_gpu_state *state,
state->bos[i].iova); state->bos[i].iova);
drm_printf(p, " size: %zd\n", state->bos[i].size); drm_printf(p, " size: %zd\n", state->bos[i].size);
adreno_show_object(p, state->bos[i].data, adreno_show_object(p, &state->bos[i].data,
state->bos[i].size); state->bos[i].size, &state->bos[i].encoded);
} }
} }
......
...@@ -187,6 +187,7 @@ struct msm_gpu_state_bo { ...@@ -187,6 +187,7 @@ struct msm_gpu_state_bo {
u64 iova; u64 iova;
size_t size; size_t size;
void *data; void *data;
bool encoded;
}; };
struct msm_gpu_state { struct msm_gpu_state {
...@@ -201,6 +202,7 @@ struct msm_gpu_state { ...@@ -201,6 +202,7 @@ struct msm_gpu_state {
u32 wptr; u32 wptr;
void *data; void *data;
int data_size; int data_size;
bool encoded;
} ring[MSM_GPU_MAX_RINGS]; } ring[MSM_GPU_MAX_RINGS];
int nr_registers; int nr_registers;
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment