Re: [PATCH v2, 5/9] media: mtk-jpegenc: Generalize jpeg encode irq interfaces
From: Tzung-Bi Shih <hidden>
Date: 2021-07-06 11:00:59
Also in:
linux-arm-kernel, linux-media, linux-mediatek, lkml
On Wed, Jun 30, 2021 at 3:28 PM kyrie.wu [off-list ref] wrote:
Generalizes jpeg encode irq interfaces to support different hardware.
There are some missing pieces for using the code. I guess the patch needs to be submitted with other patches or needs to further be divided.
quoted hunk
+ * mtk_jpeg_enc_param: General jpeg encoding parameters + * @enc_w: image width + * @enc_h: image height + * @enable_exif: EXIF enable for jpeg encode mode + * @enc_quality: destination image quality in encode mode + * @enc_format: input image format + * @restart_interval: JPEG restart interval for JPEG encoding + * @img_stride: jpeg encoder image stride + * @mem_stride: jpeg encoder memory stride + * @total_encdu: total 8x8 block number
They are not well-aligned.
quoted hunk
+struct mtk_jpeg_enc_param { + u32 enc_w; + u32 enc_h; + u32 enable_exif; + u32 enc_quality; + u32 enc_format; + u32 restart_interval; + u32 img_stride; + u32 mem_stride; + u32 total_encdu; +};
They are not used.
+ u32 bs_size; + int flags;
They are not used.
+ struct mtk_jpeg_enc_param enc_param; + struct mtk_jpeg_ctx *curr_ctx;
They are not used.
quoted hunk
+void mtk_jpeg_put_buf(struct mtk_jpeg_dev *jpeg) +{ + struct mtk_jpeg_ctx *ctx; + struct vb2_v4l2_buffer *dst_buffer; + struct list_head *temp_entry; + struct list_head *pos; + struct mtk_jpeg_src_buf *dst_done_buf, *tmp_dst_done_buf; + unsigned long flags; + + ctx = jpeg->hw_param.curr_ctx; + if (!ctx) { + dev_err(jpeg->dev, "comp_jpeg ctx fail !!!\n"); + return; + } + + dst_buffer = jpeg->hw_param.dst_buffer; + if (!dst_buffer) { + dev_err(jpeg->dev, "comp_jpeg dst_buffer fail !!!\n"); + return; + }
The caller "mtk_jpegenc_hw_irq_handler()" doesn't even check ctx and dst_buffer. Does mtk_jpeg_put_buf() need to validate them?
quoted hunk
+ spin_lock_irqsave(&ctx->done_queue_lock, flags); + list_add_tail(&dst_done_buf->list, &ctx->dst_done_queue); + while (!list_empty(&ctx->dst_done_queue) && + (pos != &ctx->dst_done_queue)) {
Why does it need to compare `pos != &ctx->dst_done_queue`? On a related note, at the first time, pos will be some garbage data from stack.
+irqreturn_t mtk_jpegenc_hw_irq_handler(int irq, void *priv)
No code is using mtk_jpegenc_hw_irq_handler. Have no enough context to review the code.
quoted hunk
+ src_buf = jpeg->hw_param.src_buffer; + dst_buf = jpeg->hw_param.dst_buffer; + ctx = jpeg->hw_param.curr_ctx; + master_jpeg = ctx->jpeg;
Could they be inlined to above where the variables are declared?
quoted hunk
+enum mtk_jpeg_hw_state { + MTK_JPEG_HW_IDLE = 0, + MTK_JPEG_HW_BUSY = 1,
MTK_JPEG_HW_BUSY is not used.
quoted hunk
@@ -124,13 +135,18 @@ struct mtk_jpeg_dev { struct v4l2_m2m_dev *m2m_dev; void *alloc_ctx; struct video_device *vdev; - void __iomem *reg_base; struct device *larb; struct delayed_work job_timeout_work; const struct mtk_jpeg_variant *variant; + void __iomem *reg_base[MTK_JPEGENC_HW_MAX]; + int jpegenc_irq;
jpegenc_irq is not used.
quoted hunk
@@ -189,6 +205,12 @@ struct mtk_jpeg_ctx { u8 enc_quality; u8 restart_interval; struct v4l2_ctrl_handler ctrl_hdl; + + struct list_head dst_done_queue; + spinlock_t done_queue_lock; /* spinlock protecting done queue */ + u32 total_frame_num;
total_frame_num is not used. Need to double confirm: why sometimes the code uses jpeg->reg_base[MTK_JPEGENC_HW0] but sometimes jpeg->reg_base[0]?