* [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional
@ 2021-09-30 18:21 Bjorn Andersson
2021-09-30 18:21 ` [PATCH v2 1/4] dt-bindings: sram: Document qcom,rpm-msg-ram Bjorn Andersson
` (5 more replies)
0 siblings, 6 replies; 13+ messages in thread
From: Bjorn Andersson @ 2021-09-30 18:21 UTC (permalink / raw)
To: Andy Gross, Bjorn Andersson, Rob Herring, Frank Rowand
Cc: linux-arm-msm, devicetree, linux-kernel, Chris Lew, Deepak Kumar Singh
In the modern Qualcomm platform there's no reason for having smem as a
separate node, so let's change this.
Bjorn Andersson (4):
dt-bindings: sram: Document qcom,rpm-msg-ram
dt-bindings: soc: smem: Make indirection optional
soc: qcom: smem: Support reserved-memory description
arm64: dts: qcom: sdm845: Drop standalone smem node
.../bindings/soc/qcom/qcom,smem.yaml | 34 +++++++++--
.../devicetree/bindings/sram/sram.yaml | 5 +-
arch/arm64/boot/dts/qcom/sdm845.dtsi | 10 +---
drivers/of/platform.c | 1 +
drivers/soc/qcom/smem.c | 57 +++++++++++++------
5 files changed, 77 insertions(+), 30 deletions(-)
--
2.29.2
^ permalink raw reply [flat|nested] 13+ messages in thread
* [PATCH v2 1/4] dt-bindings: sram: Document qcom,rpm-msg-ram
2021-09-30 18:21 [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
@ 2021-09-30 18:21 ` Bjorn Andersson
2021-10-06 21:18 ` Rob Herring
2021-09-30 18:21 ` [PATCH v2 2/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
` (4 subsequent siblings)
5 siblings, 1 reply; 13+ messages in thread
From: Bjorn Andersson @ 2021-09-30 18:21 UTC (permalink / raw)
To: Andy Gross, Bjorn Andersson, Rob Herring, Frank Rowand
Cc: linux-arm-msm, devicetree, linux-kernel, Chris Lew, Deepak Kumar Singh
The Qualcomm SMEM binding always depended on a reference to a SRAM node
of compatible "qcom,rpm-msg-ram", document this as part of the SRAM
binding.
The SRAM is consumed as a whole and not split up using subnodes, so
properties related to this are not required.
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
---
Changes since v1:
- New patch, to resolve issue with the existing qcom,smem example
Documentation/devicetree/bindings/sram/sram.yaml | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/Documentation/devicetree/bindings/sram/sram.yaml b/Documentation/devicetree/bindings/sram/sram.yaml
index 3eda5049d183..e3311b54db33 100644
--- a/Documentation/devicetree/bindings/sram/sram.yaml
+++ b/Documentation/devicetree/bindings/sram/sram.yaml
@@ -31,6 +31,7 @@ properties:
- amlogic,meson-gxbb-sram
- arm,juno-sram-ns
- atmel,sama5d2-securam
+ - qcom,rpm-msg-ram
- rockchip,rk3288-pmu-sram
reg:
@@ -135,7 +136,9 @@ if:
properties:
compatible:
contains:
- const: rockchip,rk3288-pmu-sram
+ enum:
+ - qcom,rpm-msg-ram
+ - rockchip,rk3288-pmu-sram
else:
required:
--
2.29.2
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v2 2/4] dt-bindings: soc: smem: Make indirection optional
2021-09-30 18:21 [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
2021-09-30 18:21 ` [PATCH v2 1/4] dt-bindings: sram: Document qcom,rpm-msg-ram Bjorn Andersson
@ 2021-09-30 18:21 ` Bjorn Andersson
2021-10-06 21:19 ` Rob Herring
2021-09-30 18:21 ` [PATCH v2 3/4] soc: qcom: smem: Support reserved-memory description Bjorn Andersson
` (3 subsequent siblings)
5 siblings, 1 reply; 13+ messages in thread
From: Bjorn Andersson @ 2021-09-30 18:21 UTC (permalink / raw)
To: Andy Gross, Bjorn Andersson, Rob Herring, Frank Rowand
Cc: linux-arm-msm, devicetree, linux-kernel, Chris Lew, Deepak Kumar Singh
In the olden days the Qualcomm shared memory (SMEM) region consisted of
multiple chunks of memory, so SMEM was described as a standalone node
with references to its various memory regions.
But practically all modern Qualcomm platforms has a single reserved memory
region used for SMEM. So rather than having to use two nodes to describe
the one SMEM region, update the binding to allow the reserved-memory
region alone to describe SMEM.
The olden format is preserved as valid, as this is widely used already.
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
---
Changes since v1:
- None
.../bindings/soc/qcom/qcom,smem.yaml | 34 ++++++++++++++++---
1 file changed, 30 insertions(+), 4 deletions(-)
diff --git a/Documentation/devicetree/bindings/soc/qcom/qcom,smem.yaml b/Documentation/devicetree/bindings/soc/qcom/qcom,smem.yaml
index f7e17713b3d8..4149cf2b66be 100644
--- a/Documentation/devicetree/bindings/soc/qcom/qcom,smem.yaml
+++ b/Documentation/devicetree/bindings/soc/qcom/qcom,smem.yaml
@@ -10,14 +10,18 @@ maintainers:
- Andy Gross <agross@kernel.org>
- Bjorn Andersson <bjorn.andersson@linaro.org>
-description: |
- This binding describes the Qualcomm Shared Memory Manager, used to share data
- between various subsystems and OSes in Qualcomm platforms.
+description:
+ This binding describes the Qualcomm Shared Memory Manager, a region of
+ reserved-memory used to share data between various subsystems and OSes in
+ Qualcomm platforms.
properties:
compatible:
const: qcom,smem
+ reg:
+ maxItems: 1
+
memory-region:
maxItems: 1
description: handle to memory reservation for main SMEM memory region.
@@ -29,11 +33,19 @@ properties:
$ref: /schemas/types.yaml#/definitions/phandle
description: handle to RPM message memory resource
+ no-map: true
+
required:
- compatible
- - memory-region
- hwlocks
+oneOf:
+ - required:
+ - reg
+ - no-map
+ - required:
+ - memory-region
+
additionalProperties: false
examples:
@@ -43,6 +55,20 @@ examples:
#size-cells = <1>;
ranges;
+ smem@fa00000 {
+ compatible = "qcom,smem";
+ reg = <0xfa00000 0x200000>;
+ no-map;
+
+ hwlocks = <&tcsr_mutex 3>;
+ };
+ };
+ - |
+ reserved-memory {
+ #address-cells = <1>;
+ #size-cells = <1>;
+ ranges;
+
smem_region: smem@fa00000 {
reg = <0xfa00000 0x200000>;
no-map;
--
2.29.2
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v2 3/4] soc: qcom: smem: Support reserved-memory description
2021-09-30 18:21 [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
2021-09-30 18:21 ` [PATCH v2 1/4] dt-bindings: sram: Document qcom,rpm-msg-ram Bjorn Andersson
2021-09-30 18:21 ` [PATCH v2 2/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
@ 2021-09-30 18:21 ` Bjorn Andersson
2021-10-01 7:08 ` Vladimir Zapolskiy
2021-10-06 21:20 ` Rob Herring
2021-09-30 18:21 ` [PATCH v2 4/4] arm64: dts: qcom: sdm845: Drop standalone smem node Bjorn Andersson
` (2 subsequent siblings)
5 siblings, 2 replies; 13+ messages in thread
From: Bjorn Andersson @ 2021-09-30 18:21 UTC (permalink / raw)
To: Andy Gross, Bjorn Andersson, Rob Herring, Frank Rowand
Cc: linux-arm-msm, devicetree, linux-kernel, Chris Lew, Deepak Kumar Singh
Practically all modern Qualcomm platforms has a single reserved-memory
region for SMEM. So rather than having to describe SMEM in the form of a
node with a reference to a reserved-memory node, allow the SMEM device
to be instantiated directly from the reserved-memory node.
The current means of falling back to dereferencing the "memory-region"
is kept as a fallback, if it's determined that the SMEM node is a
reserved-memory node.
The "qcom,smem" compatible is added to the reserved_mem_matches list, to
allow the reserved-memory device to be probed.
In order to retain the readability of the code, the resolution of
resources is split from the actual ioremapping.
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
---
Changes since v1:
- None
drivers/of/platform.c | 1 +
drivers/soc/qcom/smem.c | 57 ++++++++++++++++++++++++++++-------------
2 files changed, 40 insertions(+), 18 deletions(-)
diff --git a/drivers/of/platform.c b/drivers/of/platform.c
index 32d5ff8df747..07813fb1ef37 100644
--- a/drivers/of/platform.c
+++ b/drivers/of/platform.c
@@ -505,6 +505,7 @@ EXPORT_SYMBOL_GPL(of_platform_default_populate);
static const struct of_device_id reserved_mem_matches[] = {
{ .compatible = "qcom,rmtfs-mem" },
{ .compatible = "qcom,cmd-db" },
+ { .compatible = "qcom,smem" },
{ .compatible = "ramoops" },
{ .compatible = "nvmem-rmem" },
{}
diff --git a/drivers/soc/qcom/smem.c b/drivers/soc/qcom/smem.c
index 4fb5aeeb0843..c7e519bfdc8a 100644
--- a/drivers/soc/qcom/smem.c
+++ b/drivers/soc/qcom/smem.c
@@ -9,6 +9,7 @@
#include <linux/module.h>
#include <linux/of.h>
#include <linux/of_address.h>
+#include <linux/of_reserved_mem.h>
#include <linux/platform_device.h>
#include <linux/sizes.h>
#include <linux/slab.h>
@@ -240,7 +241,7 @@ static const u8 SMEM_INFO_MAGIC[] = { 0x53, 0x49, 0x49, 0x49 }; /* SIII */
* @size: size of the memory region
*/
struct smem_region {
- u32 aux_base;
+ phys_addr_t aux_base;
void __iomem *virt_base;
size_t size;
};
@@ -499,7 +500,7 @@ static void *qcom_smem_get_global(struct qcom_smem *smem,
for (i = 0; i < smem->num_regions; i++) {
region = &smem->regions[i];
- if (region->aux_base == aux_base || !aux_base) {
+ if ((u32)region->aux_base == aux_base || !aux_base) {
if (size != NULL)
*size = le32_to_cpu(entry->size);
return region->virt_base + le32_to_cpu(entry->offset);
@@ -664,7 +665,7 @@ phys_addr_t qcom_smem_virt_to_phys(void *p)
if (p < region->virt_base + region->size) {
u64 offset = p - region->virt_base;
- return (phys_addr_t)region->aux_base + offset;
+ return region->aux_base + offset;
}
}
@@ -863,12 +864,12 @@ qcom_smem_enumerate_partitions(struct qcom_smem *smem, u16 local_host)
return 0;
}
-static int qcom_smem_map_memory(struct qcom_smem *smem, struct device *dev,
- const char *name, int i)
+static int qcom_smem_resolve_mem(struct qcom_smem *smem, const char *name,
+ struct smem_region *region)
{
+ struct device *dev = smem->dev;
struct device_node *np;
struct resource r;
- resource_size_t size;
int ret;
np = of_parse_phandle(dev->of_node, name, 0);
@@ -881,13 +882,9 @@ static int qcom_smem_map_memory(struct qcom_smem *smem, struct device *dev,
of_node_put(np);
if (ret)
return ret;
- size = resource_size(&r);
- smem->regions[i].virt_base = devm_ioremap_wc(dev, r.start, size);
- if (!smem->regions[i].virt_base)
- return -ENOMEM;
- smem->regions[i].aux_base = (u32)r.start;
- smem->regions[i].size = size;
+ region->aux_base = r.start;
+ region->size = resource_size(&r);
return 0;
}
@@ -895,12 +892,14 @@ static int qcom_smem_map_memory(struct qcom_smem *smem, struct device *dev,
static int qcom_smem_probe(struct platform_device *pdev)
{
struct smem_header *header;
+ struct reserved_mem *rmem;
struct qcom_smem *smem;
size_t array_size;
int num_regions;
int hwlock_id;
u32 version;
int ret;
+ int i;
num_regions = 1;
if (of_find_property(pdev->dev.of_node, "qcom,rpm-msg-ram", NULL))
@@ -914,13 +913,35 @@ static int qcom_smem_probe(struct platform_device *pdev)
smem->dev = &pdev->dev;
smem->num_regions = num_regions;
- ret = qcom_smem_map_memory(smem, &pdev->dev, "memory-region", 0);
- if (ret)
- return ret;
+ rmem = of_reserved_mem_lookup(pdev->dev.of_node);
+ if (rmem) {
+ smem->regions[0].aux_base = rmem->base;
+ smem->regions[0].size = rmem->size;
+ } else {
+ /*
+ * Fall back to the memory-region reference, if we're not a
+ * reserved-memory node.
+ */
+ ret = qcom_smem_resolve_mem(smem, "memory-region", &smem->regions[0]);
+ if (ret)
+ return ret;
+ }
- if (num_regions > 1 && (ret = qcom_smem_map_memory(smem, &pdev->dev,
- "qcom,rpm-msg-ram", 1)))
- return ret;
+ if (num_regions > 1) {
+ ret = qcom_smem_resolve_mem(smem, "qcom,rpm-msg-ram", &smem->regions[1]);
+ if (ret)
+ return ret;
+ }
+
+ for (i = 0; i < num_regions; i++) {
+ smem->regions[i].virt_base = devm_ioremap_wc(&pdev->dev,
+ smem->regions[i].aux_base,
+ smem->regions[i].size);
+ if (!smem->regions[i].virt_base) {
+ dev_err(&pdev->dev, "failed to remap %pa\n", &smem->regions[i].aux_base);
+ return -ENOMEM;
+ }
+ }
header = smem->regions[0].virt_base;
if (le32_to_cpu(header->initialized) != 1 ||
--
2.29.2
^ permalink raw reply related [flat|nested] 13+ messages in thread
* [PATCH v2 4/4] arm64: dts: qcom: sdm845: Drop standalone smem node
2021-09-30 18:21 [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
` (2 preceding siblings ...)
2021-09-30 18:21 ` [PATCH v2 3/4] soc: qcom: smem: Support reserved-memory description Bjorn Andersson
@ 2021-09-30 18:21 ` Bjorn Andersson
2021-10-17 15:31 ` (subset) " Bjorn Andersson
2021-10-01 7:11 ` [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional Vladimir Zapolskiy
2021-10-17 15:31 ` (subset) " Bjorn Andersson
5 siblings, 1 reply; 13+ messages in thread
From: Bjorn Andersson @ 2021-09-30 18:21 UTC (permalink / raw)
To: Andy Gross, Bjorn Andersson, Rob Herring, Frank Rowand
Cc: linux-arm-msm, devicetree, linux-kernel, Chris Lew, Deepak Kumar Singh
Now that the SMEM binding and driver allows the SMEM node to be
described in the reserved-memory region directly, move the compatible
and hwlock properties to the reserved-memory node and drop the
standadlone node.
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
---
Changes since v1:
- Node name is changed to "smem", as "memory" is reserved for memory nodes.
arch/arm64/boot/dts/qcom/sdm845.dtsi | 10 +++-------
1 file changed, 3 insertions(+), 7 deletions(-)
diff --git a/arch/arm64/boot/dts/qcom/sdm845.dtsi b/arch/arm64/boot/dts/qcom/sdm845.dtsi
index beee57087d05..64119d2efdcf 100644
--- a/arch/arm64/boot/dts/qcom/sdm845.dtsi
+++ b/arch/arm64/boot/dts/qcom/sdm845.dtsi
@@ -99,9 +99,11 @@ aop_cmd_db_mem: memory@85fe0000 {
no-map;
};
- smem_mem: memory@86000000 {
+ smem@86000000 {
+ compatible = "qcom,smem";
reg = <0x0 0x86000000 0 0x200000>;
no-map;
+ hwlocks = <&tcsr_mutex 3>;
};
tz_mem: memory@86200000 {
@@ -941,12 +943,6 @@ tcsr_mutex: hwlock {
#hwlock-cells = <1>;
};
- smem {
- compatible = "qcom,smem";
- memory-region = <&smem_mem>;
- hwlocks = <&tcsr_mutex 3>;
- };
-
smp2p-cdsp {
compatible = "qcom,smp2p";
qcom,smem = <94>, <432>;
--
2.29.2
^ permalink raw reply related [flat|nested] 13+ messages in thread
* Re: [PATCH v2 3/4] soc: qcom: smem: Support reserved-memory description
2021-09-30 18:21 ` [PATCH v2 3/4] soc: qcom: smem: Support reserved-memory description Bjorn Andersson
@ 2021-10-01 7:08 ` Vladimir Zapolskiy
2021-10-01 14:23 ` Bjorn Andersson
2021-10-06 21:20 ` Rob Herring
1 sibling, 1 reply; 13+ messages in thread
From: Vladimir Zapolskiy @ 2021-10-01 7:08 UTC (permalink / raw)
To: Bjorn Andersson, Andy Gross, Rob Herring, Frank Rowand
Cc: linux-arm-msm, devicetree, linux-kernel, Chris Lew, Deepak Kumar Singh
Hi Bjorn,
On 9/30/21 9:21 PM, Bjorn Andersson wrote:
> Practically all modern Qualcomm platforms has a single reserved-memory
> region for SMEM. So rather than having to describe SMEM in the form of a
> node with a reference to a reserved-memory node, allow the SMEM device
> to be instantiated directly from the reserved-memory node.
>
> The current means of falling back to dereferencing the "memory-region"
> is kept as a fallback, if it's determined that the SMEM node is a
> reserved-memory node.
>
> The "qcom,smem" compatible is added to the reserved_mem_matches list, to
> allow the reserved-memory device to be probed.
>
> In order to retain the readability of the code, the resolution of
> resources is split from the actual ioremapping.
>
> Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
> ---
>
> Changes since v1:
> - None
>
> drivers/of/platform.c | 1 +
> drivers/soc/qcom/smem.c | 57 ++++++++++++++++++++++++++++-------------
> 2 files changed, 40 insertions(+), 18 deletions(-)
>
> diff --git a/drivers/of/platform.c b/drivers/of/platform.c
> index 32d5ff8df747..07813fb1ef37 100644
> --- a/drivers/of/platform.c
> +++ b/drivers/of/platform.c
> @@ -505,6 +505,7 @@ EXPORT_SYMBOL_GPL(of_platform_default_populate);
> static const struct of_device_id reserved_mem_matches[] = {
> { .compatible = "qcom,rmtfs-mem" },
> { .compatible = "qcom,cmd-db" },
> + { .compatible = "qcom,smem" },
> { .compatible = "ramoops" },
> { .compatible = "nvmem-rmem" },
> {}
> diff --git a/drivers/soc/qcom/smem.c b/drivers/soc/qcom/smem.c
> index 4fb5aeeb0843..c7e519bfdc8a 100644
> --- a/drivers/soc/qcom/smem.c
> +++ b/drivers/soc/qcom/smem.c
> @@ -9,6 +9,7 @@
> #include <linux/module.h>
> #include <linux/of.h>
> #include <linux/of_address.h>
> +#include <linux/of_reserved_mem.h>
> #include <linux/platform_device.h>
> #include <linux/sizes.h>
> #include <linux/slab.h>
> @@ -240,7 +241,7 @@ static const u8 SMEM_INFO_MAGIC[] = { 0x53, 0x49, 0x49, 0x49 }; /* SIII */
> * @size: size of the memory region
> */
> struct smem_region {
> - u32 aux_base;
> + phys_addr_t aux_base;
> void __iomem *virt_base;
> size_t size;
> };
> @@ -499,7 +500,7 @@ static void *qcom_smem_get_global(struct qcom_smem *smem,
> for (i = 0; i < smem->num_regions; i++) {
> region = &smem->regions[i];
>
> - if (region->aux_base == aux_base || !aux_base) {
> + if ((u32)region->aux_base == aux_base || !aux_base) {
> if (size != NULL)
> *size = le32_to_cpu(entry->size);
> return region->virt_base + le32_to_cpu(entry->offset);
> @@ -664,7 +665,7 @@ phys_addr_t qcom_smem_virt_to_phys(void *p)
> if (p < region->virt_base + region->size) {
> u64 offset = p - region->virt_base;
>
> - return (phys_addr_t)region->aux_base + offset;
> + return region->aux_base + offset;
> }
> }
>
> @@ -863,12 +864,12 @@ qcom_smem_enumerate_partitions(struct qcom_smem *smem, u16 local_host)
> return 0;
> }
>
> -static int qcom_smem_map_memory(struct qcom_smem *smem, struct device *dev,
> - const char *name, int i)
> +static int qcom_smem_resolve_mem(struct qcom_smem *smem, const char *name,
> + struct smem_region *region)
> {
> + struct device *dev = smem->dev;
> struct device_node *np;
> struct resource r;
> - resource_size_t size;
> int ret;
>
> np = of_parse_phandle(dev->of_node, name, 0);
> @@ -881,13 +882,9 @@ static int qcom_smem_map_memory(struct qcom_smem *smem, struct device *dev,
> of_node_put(np);
> if (ret)
> return ret;
> - size = resource_size(&r);
>
> - smem->regions[i].virt_base = devm_ioremap_wc(dev, r.start, size);
> - if (!smem->regions[i].virt_base)
> - return -ENOMEM;
> - smem->regions[i].aux_base = (u32)r.start;
> - smem->regions[i].size = size;
> + region->aux_base = r.start;
> + region->size = resource_size(&r);
>
> return 0;
> }
> @@ -895,12 +892,14 @@ static int qcom_smem_map_memory(struct qcom_smem *smem, struct device *dev,
> static int qcom_smem_probe(struct platform_device *pdev)
> {
> struct smem_header *header;
> + struct reserved_mem *rmem;
> struct qcom_smem *smem;
> size_t array_size;
> int num_regions;
> int hwlock_id;
> u32 version;
> int ret;
> + int i;
Just a nitpicking, the index can be unsigned.
>
> num_regions = 1;
> if (of_find_property(pdev->dev.of_node, "qcom,rpm-msg-ram", NULL))
> @@ -914,13 +913,35 @@ static int qcom_smem_probe(struct platform_device *pdev)
> smem->dev = &pdev->dev;
> smem->num_regions = num_regions;
>
> - ret = qcom_smem_map_memory(smem, &pdev->dev, "memory-region", 0);
> - if (ret)
> - return ret;
> + rmem = of_reserved_mem_lookup(pdev->dev.of_node);
> + if (rmem) {
> + smem->regions[0].aux_base = rmem->base;
> + smem->regions[0].size = rmem->size;
> + } else {
> + /*
> + * Fall back to the memory-region reference, if we're not a
> + * reserved-memory node.
> + */
> + ret = qcom_smem_resolve_mem(smem, "memory-region", &smem->regions[0]);
> + if (ret)
> + return ret;
> + }
>
> - if (num_regions > 1 && (ret = qcom_smem_map_memory(smem, &pdev->dev,
> - "qcom,rpm-msg-ram", 1)))
> - return ret;
> + if (num_regions > 1) {
> + ret = qcom_smem_resolve_mem(smem, "qcom,rpm-msg-ram", &smem->regions[1]);
> + if (ret)
> + return ret;
> + }
> +
> + for (i = 0; i < num_regions; i++) {
> + smem->regions[i].virt_base = devm_ioremap_wc(&pdev->dev,
> + smem->regions[i].aux_base,
> + smem->regions[i].size);
> + if (!smem->regions[i].virt_base) {
> + dev_err(&pdev->dev, "failed to remap %pa\n", &smem->regions[i].aux_base);
> + return -ENOMEM;
> + }
> + }
>
> header = smem->regions[0].virt_base;
> if (le32_to_cpu(header->initialized) != 1 ||
>
--
Best wishes,
Vladimir
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional
2021-09-30 18:21 [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
` (3 preceding siblings ...)
2021-09-30 18:21 ` [PATCH v2 4/4] arm64: dts: qcom: sdm845: Drop standalone smem node Bjorn Andersson
@ 2021-10-01 7:11 ` Vladimir Zapolskiy
2021-10-17 15:31 ` (subset) " Bjorn Andersson
5 siblings, 0 replies; 13+ messages in thread
From: Vladimir Zapolskiy @ 2021-10-01 7:11 UTC (permalink / raw)
To: Bjorn Andersson, Andy Gross, Rob Herring, Frank Rowand
Cc: linux-arm-msm, devicetree, linux-kernel, Chris Lew, Deepak Kumar Singh
Hi Bjorn,
On 9/30/21 9:21 PM, Bjorn Andersson wrote:
> In the modern Qualcomm platform there's no reason for having smem as a
> separate node, so let's change this.
>
> Bjorn Andersson (4):
> dt-bindings: sram: Document qcom,rpm-msg-ram
> dt-bindings: soc: smem: Make indirection optional
> soc: qcom: smem: Support reserved-memory description
> arm64: dts: qcom: sdm845: Drop standalone smem node
>
> .../bindings/soc/qcom/qcom,smem.yaml | 34 +++++++++--
> .../devicetree/bindings/sram/sram.yaml | 5 +-
> arch/arm64/boot/dts/qcom/sdm845.dtsi | 10 +---
> drivers/of/platform.c | 1 +
> drivers/soc/qcom/smem.c | 57 +++++++++++++------
> 5 files changed, 77 insertions(+), 30 deletions(-)
>
the patch series looks good.
Reviewed-by: Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>
--
Best wishes,
Vladimir
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH v2 3/4] soc: qcom: smem: Support reserved-memory description
2021-10-01 7:08 ` Vladimir Zapolskiy
@ 2021-10-01 14:23 ` Bjorn Andersson
0 siblings, 0 replies; 13+ messages in thread
From: Bjorn Andersson @ 2021-10-01 14:23 UTC (permalink / raw)
To: Vladimir Zapolskiy
Cc: Andy Gross, Rob Herring, Frank Rowand, linux-arm-msm, devicetree,
linux-kernel, Chris Lew, Deepak Kumar Singh
On Fri 01 Oct 00:08 PDT 2021, Vladimir Zapolskiy wrote:
> On 9/30/21 9:21 PM, Bjorn Andersson wrote:
[..]
> > diff --git a/drivers/soc/qcom/smem.c b/drivers/soc/qcom/smem.c
[..]
> > @@ -895,12 +892,14 @@ static int qcom_smem_map_memory(struct qcom_smem *smem, struct device *dev,
> > static int qcom_smem_probe(struct platform_device *pdev)
> > {
> > struct smem_header *header;
> > + struct reserved_mem *rmem;
> > struct qcom_smem *smem;
> > size_t array_size;
> > int num_regions;
> > int hwlock_id;
> > u32 version;
> > int ret;
> > + int i;
>
> Just a nitpicking, the index can be unsigned.
>
It's compared against the "num_regions", which is also signed. So I
think it should be signed.
[..]
> > + for (i = 0; i < num_regions; i++) {
Thanks for the review!
Regards,
Bjorn
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH v2 1/4] dt-bindings: sram: Document qcom,rpm-msg-ram
2021-09-30 18:21 ` [PATCH v2 1/4] dt-bindings: sram: Document qcom,rpm-msg-ram Bjorn Andersson
@ 2021-10-06 21:18 ` Rob Herring
0 siblings, 0 replies; 13+ messages in thread
From: Rob Herring @ 2021-10-06 21:18 UTC (permalink / raw)
To: Bjorn Andersson
Cc: Andy Gross, Rob Herring, Frank Rowand, Chris Lew, linux-kernel,
linux-arm-msm, devicetree, Deepak Kumar Singh
On Thu, 30 Sep 2021 11:21:08 -0700, Bjorn Andersson wrote:
> The Qualcomm SMEM binding always depended on a reference to a SRAM node
> of compatible "qcom,rpm-msg-ram", document this as part of the SRAM
> binding.
>
> The SRAM is consumed as a whole and not split up using subnodes, so
> properties related to this are not required.
>
> Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
> ---
>
> Changes since v1:
> - New patch, to resolve issue with the existing qcom,smem example
>
> Documentation/devicetree/bindings/sram/sram.yaml | 5 ++++-
> 1 file changed, 4 insertions(+), 1 deletion(-)
>
Acked-by: Rob Herring <robh@kernel.org>
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH v2 2/4] dt-bindings: soc: smem: Make indirection optional
2021-09-30 18:21 ` [PATCH v2 2/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
@ 2021-10-06 21:19 ` Rob Herring
0 siblings, 0 replies; 13+ messages in thread
From: Rob Herring @ 2021-10-06 21:19 UTC (permalink / raw)
To: Bjorn Andersson
Cc: linux-kernel, linux-arm-msm, Deepak Kumar Singh, Andy Gross,
devicetree, Chris Lew, Rob Herring, Frank Rowand
On Thu, 30 Sep 2021 11:21:09 -0700, Bjorn Andersson wrote:
> In the olden days the Qualcomm shared memory (SMEM) region consisted of
> multiple chunks of memory, so SMEM was described as a standalone node
> with references to its various memory regions.
>
> But practically all modern Qualcomm platforms has a single reserved memory
> region used for SMEM. So rather than having to use two nodes to describe
> the one SMEM region, update the binding to allow the reserved-memory
> region alone to describe SMEM.
>
> The olden format is preserved as valid, as this is widely used already.
>
> Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
> ---
>
> Changes since v1:
> - None
>
> .../bindings/soc/qcom/qcom,smem.yaml | 34 ++++++++++++++++---
> 1 file changed, 30 insertions(+), 4 deletions(-)
>
Reviewed-by: Rob Herring <robh@kernel.org>
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: [PATCH v2 3/4] soc: qcom: smem: Support reserved-memory description
2021-09-30 18:21 ` [PATCH v2 3/4] soc: qcom: smem: Support reserved-memory description Bjorn Andersson
2021-10-01 7:08 ` Vladimir Zapolskiy
@ 2021-10-06 21:20 ` Rob Herring
1 sibling, 0 replies; 13+ messages in thread
From: Rob Herring @ 2021-10-06 21:20 UTC (permalink / raw)
To: Bjorn Andersson
Cc: Andy Gross, Frank Rowand, linux-arm-msm, devicetree,
linux-kernel, Chris Lew, Deepak Kumar Singh
On Thu, Sep 30, 2021 at 11:21:10AM -0700, Bjorn Andersson wrote:
> Practically all modern Qualcomm platforms has a single reserved-memory
> region for SMEM. So rather than having to describe SMEM in the form of a
> node with a reference to a reserved-memory node, allow the SMEM device
> to be instantiated directly from the reserved-memory node.
>
> The current means of falling back to dereferencing the "memory-region"
> is kept as a fallback, if it's determined that the SMEM node is a
> reserved-memory node.
>
> The "qcom,smem" compatible is added to the reserved_mem_matches list, to
> allow the reserved-memory device to be probed.
>
> In order to retain the readability of the code, the resolution of
> resources is split from the actual ioremapping.
>
> Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
> ---
>
> Changes since v1:
> - None
>
> drivers/of/platform.c | 1 +
Acked-by: Rob Herring <robh@kernel.org>
> drivers/soc/qcom/smem.c | 57 ++++++++++++++++++++++++++++-------------
> 2 files changed, 40 insertions(+), 18 deletions(-)
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: (subset) [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional
2021-09-30 18:21 [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
` (4 preceding siblings ...)
2021-10-01 7:11 ` [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional Vladimir Zapolskiy
@ 2021-10-17 15:31 ` Bjorn Andersson
5 siblings, 0 replies; 13+ messages in thread
From: Bjorn Andersson @ 2021-10-17 15:31 UTC (permalink / raw)
To: Andy Gross, Rob Herring, Bjorn Andersson, Frank Rowand
Cc: linux-arm-msm, Deepak Kumar Singh, Chris Lew, linux-kernel, devicetree
On Thu, 30 Sep 2021 11:21:07 -0700, Bjorn Andersson wrote:
> In the modern Qualcomm platform there's no reason for having smem as a
> separate node, so let's change this.
>
> Bjorn Andersson (4):
> dt-bindings: sram: Document qcom,rpm-msg-ram
> dt-bindings: soc: smem: Make indirection optional
> soc: qcom: smem: Support reserved-memory description
> arm64: dts: qcom: sdm845: Drop standalone smem node
>
> [...]
Applied, thanks!
[4/4] arm64: dts: qcom: sdm845: Drop standalone smem node
commit: 622adb84b3e7c48a888c3df26fbf28679ded660b
Best regards,
--
Bjorn Andersson <bjorn.andersson@linaro.org>
^ permalink raw reply [flat|nested] 13+ messages in thread
* Re: (subset) [PATCH v2 4/4] arm64: dts: qcom: sdm845: Drop standalone smem node
2021-09-30 18:21 ` [PATCH v2 4/4] arm64: dts: qcom: sdm845: Drop standalone smem node Bjorn Andersson
@ 2021-10-17 15:31 ` Bjorn Andersson
0 siblings, 0 replies; 13+ messages in thread
From: Bjorn Andersson @ 2021-10-17 15:31 UTC (permalink / raw)
To: Andy Gross, Rob Herring, Bjorn Andersson, Frank Rowand
Cc: linux-arm-msm, Deepak Kumar Singh, Chris Lew, linux-kernel, devicetree
On Thu, 30 Sep 2021 11:21:11 -0700, Bjorn Andersson wrote:
> Now that the SMEM binding and driver allows the SMEM node to be
> described in the reserved-memory region directly, move the compatible
> and hwlock properties to the reserved-memory node and drop the
> standadlone node.
>
>
Applied, thanks!
[4/4] arm64: dts: qcom: sdm845: Drop standalone smem node
commit: 622adb84b3e7c48a888c3df26fbf28679ded660b
Best regards,
--
Bjorn Andersson <bjorn.andersson@linaro.org>
^ permalink raw reply [flat|nested] 13+ messages in thread
end of thread, other threads:[~2021-10-17 15:31 UTC | newest]
Thread overview: 13+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-09-30 18:21 [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
2021-09-30 18:21 ` [PATCH v2 1/4] dt-bindings: sram: Document qcom,rpm-msg-ram Bjorn Andersson
2021-10-06 21:18 ` Rob Herring
2021-09-30 18:21 ` [PATCH v2 2/4] dt-bindings: soc: smem: Make indirection optional Bjorn Andersson
2021-10-06 21:19 ` Rob Herring
2021-09-30 18:21 ` [PATCH v2 3/4] soc: qcom: smem: Support reserved-memory description Bjorn Andersson
2021-10-01 7:08 ` Vladimir Zapolskiy
2021-10-01 14:23 ` Bjorn Andersson
2021-10-06 21:20 ` Rob Herring
2021-09-30 18:21 ` [PATCH v2 4/4] arm64: dts: qcom: sdm845: Drop standalone smem node Bjorn Andersson
2021-10-17 15:31 ` (subset) " Bjorn Andersson
2021-10-01 7:11 ` [PATCH v2 0/4] dt-bindings: soc: smem: Make indirection optional Vladimir Zapolskiy
2021-10-17 15:31 ` (subset) " Bjorn Andersson
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).