media: amphion: Use kmalloc instead of vmalloc

Replace vmalloc/vfree with kmalloc/kfree for allocating small
driver structures (vpu_inst, vdec_t, venc_t, vpu_cmd_t, and
frame objects).

vmalloc() is designed for large memory allocations and incurs
unnecessary overhead for small objects due to virtual memory
mapping. kmalloc() is more appropriate as it allocates physically
contiguous memory with lower overhead.

ftrace measurements of vpu_alloc_cmd() show significant improvement:

  Before (vmalloc):  35-72 us   (avg ~45.7 us)
  After (kmalloc):   11-26 us   (avg ~16.8 us)

This reduces allocation time by approximately 63%.

No functional changes are intended.

Signed-off-by: Ming Qian <ming.qian@oss.nxp.com>
Reviewed-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Hans Verkuil <hverkuil+cisco@kernel.org>

diff --git a/drivers/media/platform/amphion/vdec.c b/drivers/media/platform/amphion/vdec.c
index beeaf75c76b44d05f2ca0db04e3ba37ea70fbaee..adf53b49020e9a327ebd9e9e4368ecc48be867c7 100644 (file)
--- a/drivers/media/platform/amphion/vdec.c
+++ b/drivers/media/platform/amphion/vdec.c
@@ -9,7 +9,6 @@
 #include <linux/list.h>
 #include <linux/kernel.h>
 #include <linux/module.h>
-#include <linux/vmalloc.h>
 #include <linux/videodev2.h>
 #include <media/v4l2-device.h>
 #include <media/v4l2-event.h>
@@ -17,7 +16,6 @@
 #include <media/v4l2-ioctl.h>
 #include <media/videobuf2-v4l2.h>
 #include <media/videobuf2-dma-contig.h>
-#include <media/videobuf2-vmalloc.h>
 #include "vpu.h"
 #include "vpu_defs.h"
 #include "vpu_core.h"
@@ -1662,9 +1660,9 @@ static void vdec_cleanup(struct vpu_inst *inst)
                vdec->slots = NULL;
                vdec->slot_count = 0;
        }
-       vfree(vdec);
+       kfree(vdec);
        inst->priv = NULL;
-       vfree(inst);
+       kfree(inst);
 }
 
 static void vdec_init_params(struct vdec_t *vdec)
@@ -1929,13 +1927,13 @@ static int vdec_open(struct file *file)
        struct vdec_t *vdec;
        int ret;
 
-       inst = vzalloc(sizeof(*inst));
+       inst = kzalloc(sizeof(*inst), GFP_KERNEL);
        if (!inst)
                return -ENOMEM;
 
-       vdec = vzalloc(sizeof(*vdec));
+       vdec = kzalloc(sizeof(*vdec), GFP_KERNEL);
        if (!vdec) {
-               vfree(inst);
+               kfree(inst);
                return -ENOMEM;
        }
 
@@ -1943,8 +1941,8 @@ static int vdec_open(struct file *file)
                                    sizeof(*vdec->slots),
                                    GFP_KERNEL | __GFP_ZERO);
        if (!vdec->slots) {
-               vfree(vdec);
-               vfree(inst);
+               kfree(vdec);
+               kfree(inst);
                return -ENOMEM;
        }
        vdec->slot_count = VDEC_SLOT_CNT_DFT;

diff --git a/drivers/media/platform/amphion/venc.c b/drivers/media/platform/amphion/venc.c
index aced76401b690a900706d79fa360bc1ff369ca29..9e5cbc2b0d3f4a9d173309f1eaaf676aefff0cd2 100644 (file)
--- a/drivers/media/platform/amphion/venc.c
+++ b/drivers/media/platform/amphion/venc.c
@@ -13,14 +13,12 @@
 #include <linux/videodev2.h>
 #include <linux/ktime.h>
 #include <linux/rational.h>
-#include <linux/vmalloc.h>
 #include <media/v4l2-device.h>
 #include <media/v4l2-event.h>
 #include <media/v4l2-mem2mem.h>
 #include <media/v4l2-ioctl.h>
 #include <media/videobuf2-v4l2.h>
 #include <media/videobuf2-dma-contig.h>
-#include <media/videobuf2-vmalloc.h>
 #include "vpu.h"
 #include "vpu_defs.h"
 #include "vpu_core.h"
@@ -844,7 +842,7 @@ static int venc_get_encoded_frames(struct vpu_inst *inst)
                                               v4l2_m2m_dst_buf_remove(inst->fh.m2m_ctx)))
                        break;
                list_del_init(&frame->list);
-               vfree(frame);
+               kfree(frame);
        }
 
        return 0;
@@ -860,7 +858,7 @@ static int venc_frame_encoded(struct vpu_inst *inst, void *arg)
        if (!info)
                return -EINVAL;
        venc = inst->priv;
-       frame = vzalloc(sizeof(*frame));
+       frame = kzalloc(sizeof(*frame), GFP_KERNEL);
        if (!frame)
                return -ENOMEM;
 
@@ -912,9 +910,9 @@ static void venc_cleanup(struct vpu_inst *inst)
                return;
 
        venc = inst->priv;
-       vfree(venc);
+       kfree(venc);
        inst->priv = NULL;
-       vfree(inst);
+       kfree(inst);
 }
 
 static int venc_start_session(struct vpu_inst *inst, u32 type)
@@ -1067,7 +1065,7 @@ static void venc_cleanup_frames(struct venc_t *venc)
 
        list_for_each_entry_safe(frame, tmp, &venc->frames, list) {
                list_del_init(&frame->list);
-               vfree(frame);
+               kfree(frame);
        }
 }
 
@@ -1151,7 +1149,7 @@ static int venc_process_capture(struct vpu_inst *inst, struct vb2_buffer *vb)
                return ret;
 
        list_del_init(&frame->list);
-       vfree(frame);
+       kfree(frame);
        return 0;
 }
 
@@ -1309,13 +1307,13 @@ static int venc_open(struct file *file)
        struct venc_t *venc;
        int ret;
 
-       inst = vzalloc(sizeof(*inst));
+       inst = kzalloc(sizeof(*inst), GFP_KERNEL);
        if (!inst)
                return -ENOMEM;
 
-       venc = vzalloc(sizeof(*venc));
+       venc = kzalloc(sizeof(*venc), GFP_KERNEL);
        if (!venc) {
-               vfree(inst);
+               kfree(inst);
                return -ENOMEM;
        }
 

diff --git a/drivers/media/platform/amphion/vpu_cmds.c b/drivers/media/platform/amphion/vpu_cmds.c
index 5695f5c1cb3e843de9b50b1983fa76c4dbac73ea..ab69412e0aa79e323d4766bff8359f438bb4c09f 100644 (file)
--- a/drivers/media/platform/amphion/vpu_cmds.c
+++ b/drivers/media/platform/amphion/vpu_cmds.c
@@ -13,7 +13,6 @@
 #include <linux/slab.h>
 #include <linux/types.h>
 #include <linux/delay.h>
-#include <linux/vmalloc.h>
 #include "vpu.h"
 #include "vpu_defs.h"
 #include "vpu_cmds.h"
@@ -84,13 +83,13 @@ static struct vpu_cmd_t *vpu_alloc_cmd(struct vpu_inst *inst, u32 id, void *data
        int i;
        int ret;
 
-       cmd = vzalloc(sizeof(*cmd));
+       cmd = kzalloc(sizeof(*cmd), GFP_KERNEL);
        if (!cmd)
                return NULL;
 
-       cmd->pkt = vzalloc(sizeof(*cmd->pkt));
+       cmd->pkt = kzalloc(sizeof(*cmd->pkt), GFP_KERNEL);
        if (!cmd->pkt) {
-               vfree(cmd);
+               kfree(cmd);
                return NULL;
        }
 
@@ -98,8 +97,8 @@ static struct vpu_cmd_t *vpu_alloc_cmd(struct vpu_inst *inst, u32 id, void *data
        ret = vpu_iface_pack_cmd(inst->core, cmd->pkt, inst->id, id, data);
        if (ret) {
                dev_err(inst->dev, "iface pack cmd %s fail\n", vpu_id_name(id));
-               vfree(cmd->pkt);
-               vfree(cmd);
+               kfree(cmd->pkt);
+               kfree(cmd);
                return NULL;
        }
        for (i = 0; i < ARRAY_SIZE(vpu_cmd_requests); i++) {
@@ -118,8 +117,8 @@ static void vpu_free_cmd(struct vpu_cmd_t *cmd)
                return;
        if (cmd->last_response_cmd)
                atomic_long_set(cmd->last_response_cmd, cmd->key);
-       vfree(cmd->pkt);
-       vfree(cmd);
+       kfree(cmd->pkt);
+       kfree(cmd);
 }
 
 static int vpu_session_process_cmd(struct vpu_inst *inst, struct vpu_cmd_t *cmd)

diff --git a/drivers/media/platform/amphion/vpu_core.c b/drivers/media/platform/amphion/vpu_core.c
index 168f0514851e42273e974a015e376f37b7541323..85cc4a14f8ed3c895aff0b7b2f49037c6b056e92 100644 (file)
--- a/drivers/media/platform/amphion/vpu_core.c
+++ b/drivers/media/platform/amphion/vpu_core.c
@@ -17,7 +17,6 @@
 #include <linux/pm_runtime.h>
 #include <linux/pm_domain.h>
 #include <linux/firmware.h>
-#include <linux/vmalloc.h>
 #include "vpu.h"
 #include "vpu_defs.h"
 #include "vpu_core.h"
@@ -265,7 +264,7 @@ static int vpu_core_register(struct device *dev, struct vpu_core *core)
        INIT_WORK(&core->msg_work, vpu_msg_run_work);
        INIT_DELAYED_WORK(&core->msg_delayed_work, vpu_msg_delayed_work);
        buffer_size = roundup_pow_of_two(VPU_MSG_BUFFER_SIZE);
-       core->msg_buffer = vzalloc(buffer_size);
+       core->msg_buffer = kzalloc(buffer_size, GFP_KERNEL);
        if (!core->msg_buffer) {
                dev_err(core->dev, "failed allocate buffer for fifo\n");
                ret = -ENOMEM;
@@ -282,10 +281,8 @@ static int vpu_core_register(struct device *dev, struct vpu_core *core)
 
        return 0;
 error:
-       if (core->msg_buffer) {
-               vfree(core->msg_buffer);
-               core->msg_buffer = NULL;
-       }
+       kfree(core->msg_buffer);
+       core->msg_buffer = NULL;
        if (core->workqueue) {
                destroy_workqueue(core->workqueue);
                core->workqueue = NULL;
@@ -308,7 +305,7 @@ static int vpu_core_unregister(struct device *dev, struct vpu_core *core)
 
        vpu_core_put_vpu(core);
        core->vpu = NULL;
-       vfree(core->msg_buffer);
+       kfree(core->msg_buffer);
        core->msg_buffer = NULL;
 
        if (core->workqueue) {