Re: [PATCH 2/3] dmaengine: dw-edma: Create a new dw_edma_core_ops structure to abstract controller operation

From: Serge Semin
Date: Fri Feb 10 2023 - 12:06:32 EST

Next message: Vinod Koul: "Re: [PATCH] dt-bindings: phy: amlogic,g12a-usb3-pcie-phy: add missing optional phy-supply property"
Previous message: Nick Alcock: "[PATCH 1/8] kbuild, PCI: generic,versatile: comment out MODULE_LICENSE in non-modules"
In reply to: Serge Semin: "Re: [PATCH 2/3] dmaengine: dw-edma: Create a new dw_edma_core_ops structure to abstract controller operation"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]

On Wed, Sep 21, 2022 at 02:48:51PM +0800, Cai Huoqing wrote:
> From: caihuoqing <caihuoqing@xxxxxxxxx>
>
> The structure dw_edma_core_ops has a set of the pointers
> abstracting out the DW eDMA vX and DW HDMA Native controllers.
> And use dw_edma_v0_core_register to set up operation.
>
> Signed-off-by: caihuoqing <caihuoqing@xxxxxxxxx>
> ---
> drivers/dma/dw-edma/dw-edma-core.c | 61 ++++++------------
> drivers/dma/dw-edma/dw-edma-core.h | 19 ++++++
> drivers/dma/dw-edma/dw-edma-v0-core.c | 90 ++++++++++++++++++++++++---
> drivers/dma/dw-edma/dw-edma-v0-core.h | 14 +----
> 4 files changed, 120 insertions(+), 64 deletions(-)
>
> diff --git a/drivers/dma/dw-edma/dw-edma-core.c b/drivers/dma/dw-edma/dw-edma-core.c
> index 5736a537f4c8..a66e4216a5d3 100644
> --- a/drivers/dma/dw-edma/dw-edma-core.c
> +++ b/drivers/dma/dw-edma/dw-edma-core.c
> @@ -187,6 +187,7 @@ static void dw_edma_start_transfer(struct dw_edma_chan *chan)
> struct dw_edma_chunk *child;
> struct dw_edma_desc *desc;
> struct virt_dma_desc *vd;

> + struct dw_edma *dw = chan->dw;

Either use the reverse xmas tree order or alternatively you could
directly pass "chan->dw" to the corresponding function (see my
suggestion below).

>
> vd = vchan_next_desc(&chan->vc);
> if (!vd)
> @@ -201,7 +202,7 @@ static void dw_edma_start_transfer(struct dw_edma_chan *chan)
> if (!child)
> return;
>

> - dw_edma_v0_core_start(child, !desc->xfer_sz);
> + dw->core->dw_xdma_core_start(child, !desc->xfer_sz);

Using the naked function pointers isn't that much readable. What about
introducing a set of static inline methods defined in the
dw-edma-core.h file for each op-pointer? For instance in this case the
wrapper would look like this:

static inline
dw_edma_v0_core_start(struct dw_edma *dw, struct dw_edma_chunk *chunk, bool first)
{
dw->core->start(child, first);
}

> desc->xfer_sz += child->ll_region.sz;
> dw_edma_free_burst(child);
> list_del(&child->list);
> @@ -277,6 +278,7 @@ static int dw_edma_device_resume(struct dma_chan *dchan)
> static int dw_edma_device_terminate_all(struct dma_chan *dchan)
> {
> struct dw_edma_chan *chan = dchan2dw_edma_chan(dchan);
> + struct dw_edma *dw = chan->dw;
> int err = 0;
>
> if (!chan->configured) {
> @@ -286,7 +288,7 @@ static int dw_edma_device_terminate_all(struct dma_chan *dchan)
> chan->configured = false;
> } else if (chan->status == EDMA_ST_IDLE) {
> chan->configured = false;
> - } else if (dw_edma_v0_core_ch_status(chan) == DMA_COMPLETE) {
> + } else if (dw->core->dw_xdma_core_ch_status(chan) == DMA_COMPLETE) {
> /*
> * The channel is in a false BUSY state, probably didn't
> * receive or lost an interrupt
> @@ -593,9 +595,10 @@ static void dw_edma_done_interrupt(struct dw_edma_chan *chan)
> {

> struct dw_edma_desc *desc;
> struct virt_dma_desc *vd;
> + struct dw_edma *dw = chan->dw;

ditto

> unsigned long flags;
>
> - dw_edma_v0_core_clear_done_int(chan);
> + dw->core->dw_xdma_core_clear_done_int(chan);
>
> spin_lock_irqsave(&chan->vc.lock, flags);
> vd = vchan_next_desc(&chan->vc);
> @@ -635,9 +638,10 @@ static void dw_edma_done_interrupt(struct dw_edma_chan *chan)
> static void dw_edma_abort_interrupt(struct dw_edma_chan *chan)
> {

> struct virt_dma_desc *vd;
> + struct dw_edma *dw = chan->dw;

ditto

> unsigned long flags;
>
> - dw_edma_v0_core_clear_abort_int(chan);
> + dw->core->dw_xdma_core_clear_abort_int(chan);
>
> spin_lock_irqsave(&chan->vc.lock, flags);
> vd = vchan_next_desc(&chan->vc);
> @@ -654,39 +658,10 @@ static irqreturn_t dw_edma_interrupt(int irq, void *data, bool write)
> {
> struct dw_edma_irq *dw_irq = data;
> struct dw_edma *dw = dw_irq->dw;
> - unsigned long total, pos, val;
> - unsigned long off;
> - u32 mask;
> -
> - if (write) {
> - total = dw->wr_ch_cnt;
> - off = 0;
> - mask = dw_irq->wr_mask;
> - } else {
> - total = dw->rd_ch_cnt;
> - off = dw->wr_ch_cnt;
> - mask = dw_irq->rd_mask;
> - }
> -
> - val = dw_edma_v0_core_status_done_int(dw, write ?
> - EDMA_DIR_WRITE :
> - EDMA_DIR_READ);
> - val &= mask;
> - for_each_set_bit(pos, &val, total) {
> - struct dw_edma_chan *chan = &dw->chan[pos + off];
> -
> - dw_edma_done_interrupt(chan);
> - }
> -
> - val = dw_edma_v0_core_status_abort_int(dw, write ?
> - EDMA_DIR_WRITE :
> - EDMA_DIR_READ);
> - val &= mask;
> - for_each_set_bit(pos, &val, total) {
> - struct dw_edma_chan *chan = &dw->chan[pos + off];

> + enum dw_edma_dir dir = write ? EDMA_DIR_WRITE : EDMA_DIR_READ;

ditto. Mind using reverse xmas tree vars definition order.

>
> - dw_edma_abort_interrupt(chan);
> - }
> + dw->core->dw_xdma_core_done_handle(dw_irq, dw_edma_done_interrupt, dir);

> + dw->core->dw_xdma_core_abort_handle(dw_irq, dw_edma_abort_interrupt, dir);

In addition to the status_done_int() and status_abort_int() global
methods you can get rid from two callbacks in this case:
dw_xdma_core_clear_done_int()
dw_xdma_core_clear_abort_int()
Just call the corresponding functions within the core-specific
core_done_handle() and core_abort_handle() methods before calling the
callbacks: dw_edma_done_interrupt() and dw_edma_abort_interrupt().

That way the IRQs handling procedure will be simplified a bit.

>
> return IRQ_HANDLED;
> }
> @@ -832,7 +807,7 @@ static int dw_edma_channel_setup(struct dw_edma *dw, u32 wr_alloc, u32 rd_alloc)
>
> vchan_init(&chan->vc, dma);
>
> - dw_edma_v0_core_device_config(chan);
> + dw->core->dw_xdma_core_ch_config(chan);
> }
>
> /* Set DMA channel capabilities */
> @@ -977,14 +952,16 @@ int dw_edma_probe(struct dw_edma_chip *chip)
>
> dw->chip = chip;
>
> + dw_edma_v0_core_register(dw);
> +
> raw_spin_lock_init(&dw->lock);
>
> dw->wr_ch_cnt = min_t(u16, chip->ll_wr_cnt,
> - dw_edma_v0_core_ch_count(dw, EDMA_DIR_WRITE));
> + dw->core->dw_xdma_core_ch_count(dw, EDMA_DIR_WRITE));
> dw->wr_ch_cnt = min_t(u16, dw->wr_ch_cnt, EDMA_MAX_WR_CH);
>
> dw->rd_ch_cnt = min_t(u16, chip->ll_rd_cnt,
> - dw_edma_v0_core_ch_count(dw, EDMA_DIR_READ));
> + dw->core->dw_xdma_core_ch_count(dw, EDMA_DIR_READ));
> dw->rd_ch_cnt = min_t(u16, dw->rd_ch_cnt, EDMA_MAX_RD_CH);
>
> if (!dw->wr_ch_cnt && !dw->rd_ch_cnt)
> @@ -1003,7 +980,7 @@ int dw_edma_probe(struct dw_edma_chip *chip)
> dev_name(chip->dev));
>
> /* Disable eDMA, only to establish the ideal initial conditions */
> - dw_edma_v0_core_off(dw);
> + dw->core->dw_xdma_core_off(dw);
>
> /* Request IRQs */
> err = dw_edma_irq_request(dw, &wr_alloc, &rd_alloc);
> @@ -1019,7 +996,7 @@ int dw_edma_probe(struct dw_edma_chip *chip)
> pm_runtime_enable(dev);
>
> /* Turn debugfs on */
> - dw_edma_v0_core_debugfs_on(dw);
> + dw->core->dw_xdma_core_debugfs_on(dw);
>
> chip->dw = dw;
>
> @@ -1045,7 +1022,7 @@ int dw_edma_remove(struct dw_edma_chip *chip)
> return -ENODEV;
>
> /* Disable eDMA */
> - dw_edma_v0_core_off(dw);
> + dw->core->dw_xdma_core_off(dw);
>
> /* Free irqs */
> for (i = (dw->nr_irqs - 1); i >= 0; i--)
> diff --git a/drivers/dma/dw-edma/dw-edma-core.h b/drivers/dma/dw-edma/dw-edma-core.h
> index 0ab2b6dba880..9b472d3c1596 100644
> --- a/drivers/dma/dw-edma/dw-edma-core.h
> +++ b/drivers/dma/dw-edma/dw-edma-core.h
> @@ -95,6 +95,23 @@ struct dw_edma_irq {
> struct dw_edma *dw;
> };
>

> +struct dw_edma_core_ops {
> + void (*dw_xdma_core_off)(struct dw_edma *dw);
> + u16 (*dw_xdma_core_ch_count)(struct dw_edma *dw, enum dw_edma_dir dir);
> + enum dma_status (*dw_xdma_core_ch_status)(struct dw_edma_chan *chan);

> + void (*dw_xdma_core_clear_done_int)(struct dw_edma_chan *chan);
> + void (*dw_xdma_core_clear_abort_int)(struct dw_edma_chan *chan);

As I said above just call these methods in the IRQ-handle loops within
the core-specific module. Thus you'll be able to drop these pointers.

> + void (*dw_xdma_core_done_handle)(struct dw_edma_irq *dw_irq,
> + void (*done_cb)(struct dw_edma_chan *chan),
> + enum dw_edma_dir dir);
> + void (*dw_xdma_core_abort_handle)(struct dw_edma_irq *dw_irq,
> + void (*abort_cb)(struct dw_edma_chan *chan),
> + enum dw_edma_dir dir);

1. Please fix the indentation so the arguments would be vertically
aligned.
2. "_cb" suffix seems redundant. It's obvious that the function
pointer is a callback.

> + void (*dw_xdma_core_start)(struct dw_edma_chunk *chunk, bool first);
> + int (*dw_xdma_core_ch_config)(struct dw_edma_chan *chan);
> + void (*dw_xdma_core_debugfs_on)(struct dw_edma *dw);
> +};

You can completely drop the "dw_xdma_core_" prefix from the field names.
It's obvious that they refer to the IP-core specific operation.

> +
> struct dw_edma {
> char name[32];
>
> @@ -111,6 +128,8 @@ struct dw_edma {
> raw_spinlock_t lock; /* Only for legacy */
>
> struct dw_edma_chip *chip;

> +
> + const struct dw_edma_core_ops *core;
^
+-------------------------------+
Use tab in here?

> };
>
> struct dw_edma_sg {
> diff --git a/drivers/dma/dw-edma/dw-edma-v0-core.c b/drivers/dma/dw-edma/dw-edma-v0-core.c
> index 51a34b43434c..0c1a74245ecc 100644
> --- a/drivers/dma/dw-edma/dw-edma-v0-core.c
> +++ b/drivers/dma/dw-edma/dw-edma-v0-core.c
> @@ -223,7 +223,7 @@ static inline u64 readq_ch(struct dw_edma *dw, enum dw_edma_dir dir, u16 ch,
> writeq(value, ll)
>
> /* eDMA management callbacks */
> -void dw_edma_v0_core_off(struct dw_edma *dw)
> +static void dw_edma_v0_core_off(struct dw_edma *dw)
> {
> SET_BOTH_32(dw, int_mask,
> EDMA_V0_DONE_INT_MASK | EDMA_V0_ABORT_INT_MASK);
> @@ -232,7 +232,7 @@ void dw_edma_v0_core_off(struct dw_edma *dw)
> SET_BOTH_32(dw, engine_en, 0);
> }
>
> -u16 dw_edma_v0_core_ch_count(struct dw_edma *dw, enum dw_edma_dir dir)
> +static u16 dw_edma_v0_core_ch_count(struct dw_edma *dw, enum dw_edma_dir dir)
> {
> u32 num_ch;
>
> @@ -249,7 +249,7 @@ u16 dw_edma_v0_core_ch_count(struct dw_edma *dw, enum dw_edma_dir dir)
> return (u16)num_ch;
> }
>
> -enum dma_status dw_edma_v0_core_ch_status(struct dw_edma_chan *chan)
> +static enum dma_status dw_edma_v0_core_ch_status(struct dw_edma_chan *chan)
> {
> struct dw_edma *dw = chan->dw;
> u32 tmp;
> @@ -265,7 +265,7 @@ enum dma_status dw_edma_v0_core_ch_status(struct dw_edma_chan *chan)
> return DMA_ERROR;
> }
>
> -void dw_edma_v0_core_clear_done_int(struct dw_edma_chan *chan)
> +static void dw_edma_v0_core_clear_done_int(struct dw_edma_chan *chan)
> {
> struct dw_edma *dw = chan->dw;
>
> @@ -273,7 +273,7 @@ void dw_edma_v0_core_clear_done_int(struct dw_edma_chan *chan)
> FIELD_PREP(EDMA_V0_DONE_INT_MASK, BIT(chan->id)));
> }
>
> -void dw_edma_v0_core_clear_abort_int(struct dw_edma_chan *chan)
> +static void dw_edma_v0_core_clear_abort_int(struct dw_edma_chan *chan)
> {
> struct dw_edma *dw = chan->dw;
>
> @@ -281,18 +281,70 @@ void dw_edma_v0_core_clear_abort_int(struct dw_edma_chan *chan)
> FIELD_PREP(EDMA_V0_ABORT_INT_MASK, BIT(chan->id)));
> }
>
> -u32 dw_edma_v0_core_status_done_int(struct dw_edma *dw, enum dw_edma_dir dir)
> +static u32 dw_edma_v0_core_status_done_int(struct dw_edma *dw, enum dw_edma_dir dir)
> {
> return FIELD_GET(EDMA_V0_DONE_INT_MASK,
> GET_RW_32(dw, dir, int_status));
> }
>
> -u32 dw_edma_v0_core_status_abort_int(struct dw_edma *dw, enum dw_edma_dir dir)
> +static u32 dw_edma_v0_core_status_abort_int(struct dw_edma *dw, enum dw_edma_dir dir)
> {
> return FIELD_GET(EDMA_V0_ABORT_INT_MASK,
> GET_RW_32(dw, dir, int_status));
> }
>
> +static
> +void dw_edma_v0_core_done_handle(struct dw_edma_irq *dw_irq,
> + void (*done_cb)(struct dw_edma_chan *chan),
> + enum dw_edma_dir dir)
> +{
> + struct dw_edma *dw = dw_irq->dw;
> + unsigned long total, pos, val;
> + unsigned long off;
> + u32 mask;
> +
> + if (dir == EDMA_DIR_WRITE) {
> + total = dw->wr_ch_cnt;
> + off = 0;
> + mask = dw_irq->wr_mask;
> + } else {
> + total = dw->rd_ch_cnt;
> + off = dw->wr_ch_cnt;
> + mask = dw_irq->rd_mask;
> + }
> +
> + val = dw_edma_v0_core_status_done_int(dw, dir);
> + val &= mask;
> + for_each_set_bit(pos, &val, total)
<-+
> + done_cb(&dw->chan[pos + off]); |
|
dw_edma_v0_core_clear_done_int()----------------+
?

> +}
> +
> +static
> +void dw_edma_v0_core_abort_handle(struct dw_edma_irq *dw_irq,
> + void (*abort_cb)(struct dw_edma_chan *chan),
> + enum dw_edma_dir dir)
> +{
> + struct dw_edma *dw = dw_irq->dw;
> + unsigned long total, pos, val;
> + unsigned long off;
> + u32 mask;
> +
> + if (dir == EDMA_DIR_WRITE) {
> + total = dw->wr_ch_cnt;
> + off = 0;
> + mask = dw_irq->wr_mask;
> + } else {
> + total = dw->rd_ch_cnt;
> + off = dw->wr_ch_cnt;
> + mask = dw_irq->rd_mask;
> + }
> +
> + val = dw_edma_v0_core_status_abort_int(dw, dir);
> + val &= mask;
> + for_each_set_bit(pos, &val, total)
<-+
> + abort_cb(&dw->chan[pos + off]); |
|
dw_edma_v0_core_clear_abort_int()---------------+
?

> +}
> +
> static void dw_edma_v0_core_write_chunk(struct dw_edma_chunk *chunk)
> {
> struct dw_edma_burst *child;
> @@ -338,7 +390,7 @@ static void dw_edma_v0_core_write_chunk(struct dw_edma_chunk *chunk)
> SET_LL_64(&llp->llp.reg, chunk->ll_region.paddr);
> }
>
> -void dw_edma_v0_core_start(struct dw_edma_chunk *chunk, bool first)
> +static void dw_edma_v0_core_start(struct dw_edma_chunk *chunk, bool first)
> {
> struct dw_edma_chan *chan = chunk->chan;
> struct dw_edma *dw = chan->dw;
> @@ -409,7 +461,7 @@ void dw_edma_v0_core_start(struct dw_edma_chunk *chunk, bool first)
> FIELD_PREP(EDMA_V0_DOORBELL_CH_MASK, chan->id));
> }
>
> -int dw_edma_v0_core_device_config(struct dw_edma_chan *chan)
> +static int dw_edma_v0_core_ch_config(struct dw_edma_chan *chan)
> {
> struct dw_edma *dw = chan->dw;
> u32 tmp = 0;
> @@ -481,7 +533,25 @@ int dw_edma_v0_core_device_config(struct dw_edma_chan *chan)
> }
>
> /* eDMA debugfs callbacks */
> -void dw_edma_v0_core_debugfs_on(struct dw_edma *dw)
> +static void dw_edma_v0_core_debugfs_on(struct dw_edma *dw)
> {
> dw_edma_v0_debugfs_on(dw);
> }
> +

> +static const struct dw_edma_core_ops edma_core = {

Please use dw_edma_core name for the structure instance in order to
maintain the driver naming convention.

> + .dw_xdma_core_off = dw_edma_v0_core_off,
> + .dw_xdma_core_ch_count = dw_edma_v0_core_ch_count,
> + .dw_xdma_core_ch_status = dw_edma_v0_core_ch_status,
> + .dw_xdma_core_clear_done_int = dw_edma_v0_core_clear_done_int,
> + .dw_xdma_core_clear_abort_int = dw_edma_v0_core_clear_abort_int,
> + .dw_xdma_core_done_handle = dw_edma_v0_core_done_handle,
> + .dw_xdma_core_abort_handle = dw_edma_v0_core_abort_handle,
> + .dw_xdma_core_start = dw_edma_v0_core_start,
> + .dw_xdma_core_ch_config = dw_edma_v0_core_ch_config,
> + .dw_xdma_core_debugfs_on = dw_edma_v0_core_debugfs_on,
> +};
> +
> +void dw_edma_v0_core_register(struct dw_edma *dw)
> +{
> + dw->core = &edma_core;
> +}
> diff --git a/drivers/dma/dw-edma/dw-edma-v0-core.h b/drivers/dma/dw-edma/dw-edma-v0-core.h
> index ab96a1f48080..972994ecc96f 100644
> --- a/drivers/dma/dw-edma/dw-edma-v0-core.h
> +++ b/drivers/dma/dw-edma/dw-edma-v0-core.h
> @@ -11,17 +11,7 @@
>
> #include <linux/dma/edma.h>
>
> -/* eDMA management callbacks */
> -void dw_edma_v0_core_off(struct dw_edma *chan);
> -u16 dw_edma_v0_core_ch_count(struct dw_edma *chan, enum dw_edma_dir dir);
> -enum dma_status dw_edma_v0_core_ch_status(struct dw_edma_chan *chan);
> -void dw_edma_v0_core_clear_done_int(struct dw_edma_chan *chan);
> -void dw_edma_v0_core_clear_abort_int(struct dw_edma_chan *chan);
> -u32 dw_edma_v0_core_status_done_int(struct dw_edma *chan, enum dw_edma_dir dir);
> -u32 dw_edma_v0_core_status_abort_int(struct dw_edma *chan, enum dw_edma_dir dir);
> -void dw_edma_v0_core_start(struct dw_edma_chunk *chunk, bool first);
> -int dw_edma_v0_core_device_config(struct dw_edma_chan *chan);
> -/* eDMA debug fs callbacks */
> -void dw_edma_v0_core_debugfs_on(struct dw_edma *dw);

> +/* eDMA core operation register*/

"eDMA core register"?

-Serge(y)

> +void dw_edma_v0_core_register(struct dw_edma *dw);
>
> #endif /* _DW_EDMA_V0_CORE_H */
> --
> 2.25.1
>

Next message: Vinod Koul: "Re: [PATCH] dt-bindings: phy: amlogic,g12a-usb3-pcie-phy: add missing optional phy-supply property"
Previous message: Nick Alcock: "[PATCH 1/8] kbuild, PCI: generic,versatile: comment out MODULE_LICENSE in non-modules"
In reply to: Serge Semin: "Re: [PATCH 2/3] dmaengine: dw-edma: Create a new dw_edma_core_ops structure to abstract controller operation"
Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]