2022-09-21 03:08:47

by Bhupesh Sharma

[permalink] [raw]
Subject: [PATCH v7 0/1] dma: qcom: bam_dma: Add support to initialize interconnect path

Changes since v6:
=================
- v6 can be seen here: https://lore.kernel.org/linux-arm-msm/[email protected]/
- As per Krzysztof's suggestion on v6, clubbed the crypto driver and
dt-bindings changes together. Now the overall v5 patchset into 3
separate patchsets, one each for the following areas to allow easier
review and handling from the maintainer:
arm-msm, crypto and dma
- This patch targets the dma tree specifically.

Changes since v5:
=================
- v5 can be seen here: https://lore.kernel.org/lkml/[email protected]/
- As per Bjorn's suggestion on irc, broke down the patchset into 4
separate patchsets, one each for the following areas to allow easier
review and handling from the maintainer:
arm-msm, crypto, dma and devicetree
- Addressed Rob's, Vladimir's and Bjorn's review comments received on
v5.
- Added Tested-by from Jordan received on the v5.

Changes since v4:
=================
- v4 for sm8250 can be seen here: https://lore.kernel.org/linux-arm-msm/[email protected]/
- v1 for sm8150 qce enablement can be seen here: https://lore.kernel.org/linux-arm-msm/[email protected]/
- Merged the sm8150 and sm8250 enablement patches in the same patchset,
as per suggestions from Bjorn.
- Dropped a couple of patches from v4, as these have been picked by
Bjorn already via his tree.
- Addressed review comments from Vladimir, Thara and Rob.
- Collect Reviewed-by from Rob and Thara on some of the patches from the
v4 patchset.

Changes since v3:
=================
- v3 can be seen here: https://lore.kernel.org/linux-arm-msm/[email protected]/
- Dropped a couple of patches from v3, on basis of the review comments:
~ [PATCH 13/17] crypto: qce: core: Make clocks optional
~ [PATCH 15/17] crypto: qce: Convert the device found dev_dbg() to dev_info()
- Addressed review comments from Thara, Rob and Stephan Gerhold.
- Collect Reviewed-by from Rob and Thara on some of the patches from the
v3 patchset.

Changes since v2:
=================
- v2 can be seen here: https://lore.kernel.org/dmaengine/[email protected]/
- Drop a couple of patches from v1, which tried to address the defered
probing of qce driver in case bam dma driver is not yet probed.
Replace it instead with a single (simpler) patch [PATCH 16/17].
- Convert bam dma and qce crypto dt-bindings to YAML.
- Addressed review comments from Thara, Bjorn, Vinod and Rob.

Changes since v1:
=================
- v1 can be seen here: https://lore.kernel.org/linux-arm-msm/[email protected]/
- v1 did not work well as reported earlier by Dmitry, so v2 contains the following
changes/fixes:
~ Enable the interconnect path b/w BAM DMA and main memory first
before trying to access the BAM DMA registers.
~ Enable the interconnect path b/w qce crytpo and main memory first
before trying to access the qce crypto registers.
~ Make sure to document the required and optional properties for both
BAM DMA and qce crypto drivers.
~ Add a few debug related print messages in case the qce crypto driver
passes or fails to probe.
~ Convert the qce crypto driver probe to a defered one in case the BAM DMA
or the interconnect driver(s) (needed on specific Qualcomm parts) are not
yet probed.

BAM dma engine associated with certain Qualcomm Snapdragon hardware blocks requires
relevant interconnect pieces be initialized prior to the dma engine initialization.
For e.g. crypto bam dma engine on sm8250.

Such requirement is passed on to the bam dma driver from dt via the "interconnects"
property. Add support in bam_dma driver to check whether the interconnect
path is accessible/enabled prior to attempting driver intializations.

If interconnects are not yet setup, defer the BAM DMA driver probe().

Tested on sm8150-mtp, sa8155p-adp, sm8250-mtp and RB5 boards.

Cc: [email protected]
Cc: [email protected]
Cc: [email protected]
Tested-by: Jordan Crouse <[email protected]>

Thara Gopinath (1):
dma: qcom: bam_dma: Add support to initialize interconnect path

drivers/dma/qcom/bam_dma.c | 10 ++++++++++
1 file changed, 10 insertions(+)

--
2.37.1


2022-09-21 03:15:13

by Bhupesh Sharma

[permalink] [raw]
Subject: [PATCH v7 1/1] dma: qcom: bam_dma: Add support to initialize interconnect path

From: Thara Gopinath <[email protected]>

BAM dma engine associated with certain hardware blocks could require
relevant interconnect pieces be initialized prior to the dma engine
initialization. For e.g. crypto bam dma engine on sm8250. Such requirement
is passed on to the bam dma driver from dt via the "interconnects"
property. Add support in bam_dma driver to check whether the interconnect
path is accessible/enabled prior to attempting driver intializations.

If interconnects are not yet setup, defer the BAM DMA driver probe().

Cc: Bjorn Andersson <[email protected]>
Cc: Rob Herring <[email protected]>
Signed-off-by: Thara Gopinath <[email protected]>
Signed-off-by: Bhupesh Sharma <[email protected]>
[Bhupesh: Make header file inclusion alphabetical and use 'devm_of_icc_get()']
---
drivers/dma/qcom/bam_dma.c | 10 ++++++++++
1 file changed, 10 insertions(+)

diff --git a/drivers/dma/qcom/bam_dma.c b/drivers/dma/qcom/bam_dma.c
index 2ff787df513e..a5b0cf28ffb7 100644
--- a/drivers/dma/qcom/bam_dma.c
+++ b/drivers/dma/qcom/bam_dma.c
@@ -26,6 +26,7 @@
#include <linux/kernel.h>
#include <linux/io.h>
#include <linux/init.h>
+#include <linux/interconnect.h>
#include <linux/slab.h>
#include <linux/module.h>
#include <linux/interrupt.h>
@@ -394,6 +395,7 @@ struct bam_device {
const struct reg_offset_data *layout;

struct clk *bamclk;
+ struct icc_path *mem_path;
int irq;

/* dma start transaction tasklet */
@@ -1294,6 +1296,14 @@ static int bam_dma_probe(struct platform_device *pdev)
if (IS_ERR(bdev->bamclk))
return PTR_ERR(bdev->bamclk);

+ /* Ensure that interconnects are initialized */
+ bdev->mem_path = devm_of_icc_get(bdev->dev, "memory");
+ if (IS_ERR(bdev->mem_path)) {
+ ret = dev_err_probe(bdev->dev, PTR_ERR(bdev->mem_path),
+ "failed to acquire icc path\n");
+ return ret;
+ }
+
ret = clk_prepare_enable(bdev->bamclk);
if (ret) {
dev_err(bdev->dev, "failed to prepare/enable clock\n");
--
2.37.1

2022-09-22 02:45:09

by Vinod Koul

[permalink] [raw]
Subject: Re: [PATCH v7 1/1] dma: qcom: bam_dma: Add support to initialize interconnect path

On 21-09-22, 08:36, Bhupesh Sharma wrote:
> From: Thara Gopinath <[email protected]>
>
> BAM dma engine associated with certain hardware blocks could require
> relevant interconnect pieces be initialized prior to the dma engine
> initialization. For e.g. crypto bam dma engine on sm8250. Such requirement
> is passed on to the bam dma driver from dt via the "interconnects"
> property. Add support in bam_dma driver to check whether the interconnect
> path is accessible/enabled prior to attempting driver intializations.
>
> If interconnects are not yet setup, defer the BAM DMA driver probe().
>
> Cc: Bjorn Andersson <[email protected]>
> Cc: Rob Herring <[email protected]>
> Signed-off-by: Thara Gopinath <[email protected]>
> Signed-off-by: Bhupesh Sharma <[email protected]>
> [Bhupesh: Make header file inclusion alphabetical and use 'devm_of_icc_get()']
> ---
> drivers/dma/qcom/bam_dma.c | 10 ++++++++++
> 1 file changed, 10 insertions(+)
>
> diff --git a/drivers/dma/qcom/bam_dma.c b/drivers/dma/qcom/bam_dma.c
> index 2ff787df513e..a5b0cf28ffb7 100644
> --- a/drivers/dma/qcom/bam_dma.c
> +++ b/drivers/dma/qcom/bam_dma.c
> @@ -26,6 +26,7 @@
> #include <linux/kernel.h>
> #include <linux/io.h>
> #include <linux/init.h>
> +#include <linux/interconnect.h>
> #include <linux/slab.h>
> #include <linux/module.h>
> #include <linux/interrupt.h>
> @@ -394,6 +395,7 @@ struct bam_device {
> const struct reg_offset_data *layout;
>
> struct clk *bamclk;
> + struct icc_path *mem_path;
> int irq;
>
> /* dma start transaction tasklet */
> @@ -1294,6 +1296,14 @@ static int bam_dma_probe(struct platform_device *pdev)
> if (IS_ERR(bdev->bamclk))
> return PTR_ERR(bdev->bamclk);
>
> + /* Ensure that interconnects are initialized */
> + bdev->mem_path = devm_of_icc_get(bdev->dev, "memory");
> + if (IS_ERR(bdev->mem_path)) {
> + ret = dev_err_probe(bdev->dev, PTR_ERR(bdev->mem_path),
> + "failed to acquire icc path\n");
> + return ret;
> + }

So this makes us fail on older DT where icc path may not be present.
Should this not be an optional thing?

> +
> ret = clk_prepare_enable(bdev->bamclk);
> if (ret) {
> dev_err(bdev->dev, "failed to prepare/enable clock\n");
> --
> 2.37.1

--
~Vinod

2022-09-23 05:27:22

by Bhupesh Sharma

[permalink] [raw]
Subject: Re: [PATCH v7 1/1] dma: qcom: bam_dma: Add support to initialize interconnect path

On Thu, 22 Sept 2022 at 08:08, Vinod Koul <[email protected]> wrote:
>
> On 21-09-22, 08:36, Bhupesh Sharma wrote:
> > From: Thara Gopinath <[email protected]>
> >
> > BAM dma engine associated with certain hardware blocks could require
> > relevant interconnect pieces be initialized prior to the dma engine
> > initialization. For e.g. crypto bam dma engine on sm8250. Such requirement
> > is passed on to the bam dma driver from dt via the "interconnects"
> > property. Add support in bam_dma driver to check whether the interconnect
> > path is accessible/enabled prior to attempting driver intializations.
> >
> > If interconnects are not yet setup, defer the BAM DMA driver probe().
> >
> > Cc: Bjorn Andersson <[email protected]>
> > Cc: Rob Herring <[email protected]>
> > Signed-off-by: Thara Gopinath <[email protected]>
> > Signed-off-by: Bhupesh Sharma <[email protected]>
> > [Bhupesh: Make header file inclusion alphabetical and use 'devm_of_icc_get()']
> > ---
> > drivers/dma/qcom/bam_dma.c | 10 ++++++++++
> > 1 file changed, 10 insertions(+)
> >
> > diff --git a/drivers/dma/qcom/bam_dma.c b/drivers/dma/qcom/bam_dma.c
> > index 2ff787df513e..a5b0cf28ffb7 100644
> > --- a/drivers/dma/qcom/bam_dma.c
> > +++ b/drivers/dma/qcom/bam_dma.c
> > @@ -26,6 +26,7 @@
> > #include <linux/kernel.h>
> > #include <linux/io.h>
> > #include <linux/init.h>
> > +#include <linux/interconnect.h>
> > #include <linux/slab.h>
> > #include <linux/module.h>
> > #include <linux/interrupt.h>
> > @@ -394,6 +395,7 @@ struct bam_device {
> > const struct reg_offset_data *layout;
> >
> > struct clk *bamclk;
> > + struct icc_path *mem_path;
> > int irq;
> >
> > /* dma start transaction tasklet */
> > @@ -1294,6 +1296,14 @@ static int bam_dma_probe(struct platform_device *pdev)
> > if (IS_ERR(bdev->bamclk))
> > return PTR_ERR(bdev->bamclk);
> >
> > + /* Ensure that interconnects are initialized */
> > + bdev->mem_path = devm_of_icc_get(bdev->dev, "memory");
> > + if (IS_ERR(bdev->mem_path)) {
> > + ret = dev_err_probe(bdev->dev, PTR_ERR(bdev->mem_path),
> > + "failed to acquire icc path\n");
> > + return ret;
> > + }
>
> So this makes us fail on older DT where icc path may not be present.
> Should this not be an optional thing?

That's a good point. I am not sure if Thara tried this on platforms
which don't have the icc path available.

I will fix this in v8.

Thanks,
Bhupesh

> > +
> > ret = clk_prepare_enable(bdev->bamclk);
> > if (ret) {
> > dev_err(bdev->dev, "failed to prepare/enable clock\n");
> > --
> > 2.37.1
>
> --
> ~Vinod