Add binding documentation for the Qualcomm Hardware Mutex.
Signed-off-by: Bjorn Andersson <[email protected]>
---
I think the conclusion on the dt binding discussion for hwspinlocks was that
we're down to having the #hwlock-cells intact. So this version includes that,
but non of the other previously discussed properties.
Changes since v5:
- Extracted the dt binding documentation into a separate patch
- Moved the driver to consume a syscon
- Dropped previously suggested generic hwlock dt bindings
.../devicetree/bindings/hwlock/qcom-hwspinlock.txt | 39 ++++++++++++++++++++++
1 file changed, 39 insertions(+)
create mode 100644 Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
diff --git a/Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt b/Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
new file mode 100644
index 0000000..28ade7d
--- /dev/null
+++ b/Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
@@ -0,0 +1,39 @@
+Qualcomm Hardware Mutex Block:
+
+The hardware block provides mutexes utilized between different processors on
+the SoC as part of the communication protocol used by these processors.
+
+- compatible:
+ Usage: required
+ Value type: <string>
+ Definition: must be one of:
+ "qcom,sfpb-mutex",
+ "qcom,tcsr-mutex"
+
+- syscon:
+ Usage: required
+ Value type: <prop-encoded-array>
+ Definition: one cell containing:
+ syscon phandle
+ offset of the hwmutex block within the syscon
+ stride of the hwmutex registers
+
+- #hwlock-cells:
+ Usage: required
+ Value type: <u32>
+ Definition: must be 1, the specified cell represent the lock id
+ (hwlock standard property, see hwlock.txt)
+
+Example:
+
+ tcsr: syscon@1a400000 {
+ compatible = "qcom,tcsr-msm8974", "syscon";
+ reg = <0xfd484000 0x2000>;
+ };
+
+ hwlock@fd484000 {
+ compatible = "qcom,tcsr-mutex";
+ syscon = <&tcsr 0 0x80>;
+
+ #hwlock-cells = <1>;
+ };
--
1.8.2.2
Add driver for Qualcomm Hardware Mutex block found in many Qualcomm
SoCs.
Based on initial effort by Kumar Gala <[email protected]>
Signed-off-by: Bjorn Andersson <[email protected]>
---
As Andy Gross introduced the tcsr syscon we can no longer just ioremap the
memory directly, so rework the driver to run ontop of syscon.
Changes since v5:
- Dropped all but hwspinlock specific dt bindings
- Hardcoded the number of locks (there is 32)
- Rework to sit ontop of syscon
Changes since v4:
- Aligned with devicetree support in hwlock framework and hence depends on [1]
Changes since v3:
- Reverted back to getting stride from of_match, per Kumars request
Changes since v2:
- MODULE_DEVICE_TABLE
- Changed prefix to qcom
- Cleaned up includes
- Rely on reg and num-locks to figure out stride, instead of of_match data
Changes since v1:
- Added the pm_runtime calls needed to be able to boot a kernel with
pm_runtime and this driver, patch from Courtney.
- Added sfpb-mutex compatible, for re-use of the driver in family A platforms.
- Updated formatting of DT binding documentation, while adding the extra
compatible.
- Dropped Stephen Boyds Reviewed-by due to these changes.
drivers/hwspinlock/Kconfig | 11 +++
drivers/hwspinlock/Makefile | 1 +
drivers/hwspinlock/qcom_hwspinlock.c | 181 +++++++++++++++++++++++++++++++++++
3 files changed, 193 insertions(+)
create mode 100644 drivers/hwspinlock/qcom_hwspinlock.c
diff --git a/drivers/hwspinlock/Kconfig b/drivers/hwspinlock/Kconfig
index 3612cb5..762216d 100644
--- a/drivers/hwspinlock/Kconfig
+++ b/drivers/hwspinlock/Kconfig
@@ -18,6 +18,17 @@ config HWSPINLOCK_OMAP
If unsure, say N.
+config HWSPINLOCK_QCOM
+ tristate "Qualcomm Hardware Spinlock device"
+ depends on ARCH_QCOM
+ select HWSPINLOCK
+ help
+ Say y here to support the Qualcomm Hardware Mutex functionality, which
+ provides a synchronisation mechanism for the various processors on
+ the SoC.
+
+ If unsure, say N.
+
config HSEM_U8500
tristate "STE Hardware Semaphore functionality"
depends on ARCH_U8500
diff --git a/drivers/hwspinlock/Makefile b/drivers/hwspinlock/Makefile
index 93eb64b..68f95d9 100644
--- a/drivers/hwspinlock/Makefile
+++ b/drivers/hwspinlock/Makefile
@@ -4,4 +4,5 @@
obj-$(CONFIG_HWSPINLOCK) += hwspinlock_core.o
obj-$(CONFIG_HWSPINLOCK_OMAP) += omap_hwspinlock.o
+obj-$(CONFIG_HWSPINLOCK_QCOM) += qcom_hwspinlock.o
obj-$(CONFIG_HSEM_U8500) += u8500_hsem.o
diff --git a/drivers/hwspinlock/qcom_hwspinlock.c b/drivers/hwspinlock/qcom_hwspinlock.c
new file mode 100644
index 0000000..93b62e0
--- /dev/null
+++ b/drivers/hwspinlock/qcom_hwspinlock.c
@@ -0,0 +1,181 @@
+/*
+ * Copyright (c) 2013, The Linux Foundation. All rights reserved.
+ * Copyright (c) 2015, Sony Mobile Communications AB
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/hwspinlock.h>
+#include <linux/io.h>
+#include <linux/kernel.h>
+#include <linux/mfd/syscon.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/of_device.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+#include <linux/regmap.h>
+
+#include "hwspinlock_internal.h"
+
+#define QCOM_MUTEX_APPS_PROC_ID 1
+#define QCOM_MUTEX_NUM_LOCKS 32
+
+static int qcom_hwspinlock_trylock(struct hwspinlock *lock)
+{
+ struct regmap_field *field = lock->priv;
+ u32 lock_owner;
+ int ret;
+
+ ret = regmap_field_write(field, QCOM_MUTEX_APPS_PROC_ID);
+ if (ret)
+ return ret;
+
+ ret = regmap_field_read(field, &lock_owner);
+ if (ret)
+ return ret;
+
+ return lock_owner == QCOM_MUTEX_APPS_PROC_ID;
+}
+
+static void qcom_hwspinlock_unlock(struct hwspinlock *lock)
+{
+ struct regmap_field *field = lock->priv;
+ u32 lock_owner;
+ int ret;
+
+ ret = regmap_field_read(field, &lock_owner);
+ if (ret) {
+ pr_err("%s: unable to query spinlock owner\n", __func__);
+ return;
+ }
+
+ if (lock_owner != QCOM_MUTEX_APPS_PROC_ID) {
+ pr_err("%s: spinlock not owned by us (actual owner is %d)\n",
+ __func__, lock_owner);
+ }
+
+ ret = regmap_field_write(field, 0);
+ if (ret)
+ pr_err("%s: failed to unlock spinlock\n", __func__);
+}
+
+static const struct hwspinlock_ops qcom_hwspinlock_ops = {
+ .trylock = qcom_hwspinlock_trylock,
+ .unlock = qcom_hwspinlock_unlock,
+};
+
+static const struct of_device_id qcom_hwspinlock_of_match[] = {
+ { .compatible = "qcom,sfpb-mutex" },
+ { .compatible = "qcom,tcsr-mutex" },
+ { }
+};
+MODULE_DEVICE_TABLE(of, qcom_hwspinlock_of_match);
+
+static int qcom_hwspinlock_probe(struct platform_device *pdev)
+{
+ struct hwspinlock_device *bank;
+ struct device_node *syscon;
+ struct reg_field field;
+ struct regmap *regmap;
+ size_t array_size;
+ u32 stride;
+ u32 base;
+ int ret;
+ int i;
+
+ syscon = of_parse_phandle(pdev->dev.of_node, "syscon", 0);
+ if (!syscon) {
+ dev_err(&pdev->dev, "no syscon property\n");
+ return -ENODEV;
+ }
+
+ regmap = syscon_node_to_regmap(syscon);
+ if (IS_ERR(regmap))
+ return PTR_ERR(regmap);
+
+ ret = of_property_read_u32_index(pdev->dev.of_node, "syscon", 1, &base);
+ if (ret < 0) {
+ dev_err(&pdev->dev, "no offset in syscon\n");
+ return -EINVAL;
+ }
+
+ ret = of_property_read_u32_index(pdev->dev.of_node, "syscon", 2, &stride);
+ if (ret < 0) {
+ dev_err(&pdev->dev, "no stride syscon\n");
+ return -EINVAL;
+ }
+
+ array_size = QCOM_MUTEX_NUM_LOCKS * sizeof(struct hwspinlock);
+ bank = devm_kzalloc(&pdev->dev, sizeof(*bank) + array_size, GFP_KERNEL);
+ if (!bank)
+ return -ENOMEM;
+
+ platform_set_drvdata(pdev, bank);
+
+ for (i = 0; i < QCOM_MUTEX_NUM_LOCKS; i++) {
+ field.reg = base + i * stride;
+ field.lsb = 0;
+ field.msb = 32;
+
+ bank->lock[i].priv = devm_regmap_field_alloc(&pdev->dev,
+ regmap, field);
+ }
+
+ pm_runtime_enable(&pdev->dev);
+
+ ret = hwspin_lock_register(bank, &pdev->dev, &qcom_hwspinlock_ops,
+ 0, QCOM_MUTEX_NUM_LOCKS);
+ if (ret)
+ pm_runtime_disable(&pdev->dev);
+
+ return ret;
+}
+
+static int qcom_hwspinlock_remove(struct platform_device *pdev)
+{
+ struct hwspinlock_device *bank = platform_get_drvdata(pdev);
+ int ret;
+
+ ret = hwspin_lock_unregister(bank);
+ if (ret) {
+ dev_err(&pdev->dev, "%s failed: %d\n", __func__, ret);
+ return ret;
+ }
+
+ pm_runtime_disable(&pdev->dev);
+
+ return 0;
+}
+
+static struct platform_driver qcom_hwspinlock_driver = {
+ .probe = qcom_hwspinlock_probe,
+ .remove = qcom_hwspinlock_remove,
+ .driver = {
+ .name = "qcom_hwspinlock",
+ .of_match_table = qcom_hwspinlock_of_match,
+ },
+};
+
+static int __init qcom_hwspinlock_init(void)
+{
+ return platform_driver_register(&qcom_hwspinlock_driver);
+}
+/* board init code might need to reserve hwspinlocks for predefined purposes */
+postcore_initcall(qcom_hwspinlock_init);
+
+static void __exit qcom_hwspinlock_exit(void)
+{
+ platform_driver_unregister(&qcom_hwspinlock_driver);
+}
+module_exit(qcom_hwspinlock_exit);
+
+MODULE_LICENSE("GPL v2");
+MODULE_DESCRIPTION("Hardware spinlock driver for Qualcomm SoCs");
--
1.8.2.2
On Thu, Mar 12, 2015 at 2:29 AM, Ohad Ben-Cohen <[email protected]> wrote:
> Hi Mark, Rob,
>
> On Sat, Feb 28, 2015 at 12:30 AM, Bjorn Andersson
> <[email protected]> wrote:
>> Add binding documentation for the Qualcomm Hardware Mutex.
>>
>> Signed-off-by: Bjorn Andersson <[email protected]>
>> ---
>>
>> I think the conclusion on the dt binding discussion for hwspinlocks was that
>> we're down to having the #hwlock-cells intact. So this version includes that,
>> but non of the other previously discussed properties.
>>
>> Changes since v5:
>> - Extracted the dt binding documentation into a separate patch
>> - Moved the driver to consume a syscon
>> - Dropped previously suggested generic hwlock dt bindings
>>
>> .../devicetree/bindings/hwlock/qcom-hwspinlock.txt | 39 ++++++++++++++++++++++
>> 1 file changed, 39 insertions(+)
>> create mode 100644 Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
>
> Could you please ack this one patch from Bjorn?
>
> I'll need your Ack before I can this forward.
I didn't see an Ack from Mark or Rob. But I did see a question from
Mark and response from Bjorn.
Ohad - did you take this or are you still waiting for something?
Who should I pester about this? :-)
-- Tim
On Thu, Apr 2, 2015 at 12:32 AM, Tim Bird <[email protected]> wrote:
> I didn't see an Ack from Mark or Rob. But I did see a question from
> Mark and response from Bjorn.
>
> Ohad - did you take this or are you still waiting for something?
>
> Who should I pester about this? :-)
Sorry, I can't take this without a DT ack.
Ohad.
On Wed, Apr 1, 2015 at 9:40 PM, Ohad Ben-Cohen <[email protected]> wrote:
> On Thu, Apr 2, 2015 at 12:32 AM, Tim Bird <[email protected]> wrote:
>> I didn't see an Ack from Mark or Rob. But I did see a question from
>> Mark and response from Bjorn.
>>
>> Ohad - did you take this or are you still waiting for something?
>>
>> Who should I pester about this? :-)
>
> Sorry, I can't take this without a DT ack.
Hmmm.
The policy seems to be:
"For driver (not subsystem) bindings: If you are comfortable with the
binding, and it hasn't received an Acked-by from the devicetree
maintainers after a few weeks, go ahead and take it."
The syscon property is only relative to the qcom hwspinlock driver,
(unless I'm missing something) and both Qualcomm and Sony devs are
OK with it. So while an ACK from the DT side would be nice, I don't
think it's required. This is exactly the type of delay that is really
holding up a lot of out-of-tree code.
But, in case that doesn't convince you...
Mark or Rob (or someone, anyone, over on the DT side)...
Can you please ACK or NAK the following binding? This particular driver
is at the bottom of a very large dependency tree for out-of-tree code
for Qualcomm.
Thanks,
-- Tim
---
Add binding documentation for the Qualcomm Hardware Mutex.
Signed-off-by: Bjorn Andersson <[email protected]>
---
I think the conclusion on the dt binding discussion for hwspinlocks was that
we're down to having the #hwlock-cells intact. So this version includes that,
but none of the other previously discussed properties.
Changes since v5:
- Extracted the dt binding documentation into a separate patch
- Moved the driver to consume a syscon
- Dropped previously suggested generic hwlock dt bindings
.../devicetree/bindings/hwlock/qcom-hwspinlock.txt | 39 ++++++++++++++++++++++
1 file changed, 39 insertions(+)
create mode 100644 Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
diff --git a/Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
b/Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
new file mode 100644
index 0000000..28ade7d
--- /dev/null
+++ b/Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
@@ -0,0 +1,39 @@
+Qualcomm Hardware Mutex Block:
+
+The hardware block provides mutexes utilized between different processors on
+the SoC as part of the communication protocol used by these processors.
+
+- compatible:
+ Usage: required
+ Value type: <string>
+ Definition: must be one of:
+ "qcom,sfpb-mutex",
+ "qcom,tcsr-mutex"
+
+- syscon:
+ Usage: required
+ Value type: <prop-encoded-array>
+ Definition: one cell containing:
+ syscon phandle
+ offset of the hwmutex block within the syscon
+ stride of the hwmutex registers
+
+- #hwlock-cells:
+ Usage: required
+ Value type: <u32>
+ Definition: must be 1, the specified cell represent the lock id
+ (hwlock standard property, see hwlock.txt)
+
+Example:
+
+ tcsr: syscon@1a400000 {
+ compatible = "qcom,tcsr-msm8974", "syscon";
+ reg = <0xfd484000 0x2000>;
+ };
+
+ hwlock@fd484000 {
+ compatible = "qcom,tcsr-mutex";
+ syscon = <&tcsr 0 0x80>;
+
+ #hwlock-cells = <1>;
+ };
--
On Thu, Apr 2, 2015 at 9:11 PM, Tim Bird <[email protected]> wrote:
> On Wed, Apr 1, 2015 at 9:40 PM, Ohad Ben-Cohen <[email protected]> wrote:
>> Sorry, I can't take this without a DT ack.
>
> Hmmm.
>
> The policy seems to be:
> "For driver (not subsystem) bindings: If you are comfortable with the
> binding, and it hasn't received an Acked-by from the devicetree
> maintainers after a few weeks, go ahead and take it."
>
> The syscon property is only relative to the qcom hwspinlock driver,
> (unless I'm missing something) and both Qualcomm and Sony devs are
> OK with it. So while an ACK from the DT side would be nice, I don't
> think it's required. This is exactly the type of delay that is really
> holding up a lot of out-of-tree code.
Sorry, I do prefer to make sure Mark is OK with this devicetree patch,
especially since it wasn't clear whether Mark is entirely comfortable
with it in his last response.
Thanks,
Ohad.
On Fri, Apr 3, 2015 at 6:55 AM, Ohad Ben-Cohen <[email protected]> wrote:
> On Thu, Apr 2, 2015 at 9:11 PM, Tim Bird <[email protected]> wrote:
>> On Wed, Apr 1, 2015 at 9:40 PM, Ohad Ben-Cohen <[email protected]> wrote:
>>> Sorry, I can't take this without a DT ack.
>>
>> Hmmm.
>>
>> The policy seems to be:
>> "For driver (not subsystem) bindings: If you are comfortable with the
>> binding, and it hasn't received an Acked-by from the devicetree
>> maintainers after a few weeks, go ahead and take it."
>>
>> The syscon property is only relative to the qcom hwspinlock driver,
>> (unless I'm missing something) and both Qualcomm and Sony devs are
>> OK with it. So while an ACK from the DT side would be nice, I don't
>> think it's required. This is exactly the type of delay that is really
>> holding up a lot of out-of-tree code.
>
> Sorry, I do prefer to make sure Mark is OK with this devicetree patch,
> especially since it wasn't clear whether Mark is entirely comfortable
> with it in his last response.
Just to be clear - do you personally have any objections to the patch?
Are we now stuck in limbo until such time as the device tree maintainers
get around to us?
I'm worried about this status because my understanding is that the
DT maintainers are hugely backlogged and let a lot of stuff drop on the
floor. In particular, see slide 21 of the following presentation:
http://www.elinux.org/images/0/0a/The_Device_Tree_as_a_Stable_ABI-_A_Fairy_Tale%3F.pdf
That graph shows that the DT maintainers are falling way behind in
their reviews and ACKs.
The policy of not waiting for an ACK from device tree maintainers
was specifically created to help the situation we are now in, and yet
you seem to be unwilling to follow it. This is extremely frustrating.
One idea we discussed at a recent meeting on mainlining was to submit
SoC-blocking items to drivers/staging. Then, stuff is at least in the tree
and can be tested by others pending some approval. Do you have any
opinion on that?
Is there any way to move ahead? Or are we doomed to just sit around
and wait indefinitely? For the record, after what amount of time without
a DT ACK would you consider accepting this (if ever)?
Another idea I'm considering is to write our own hwspinlock layer
and become the maintainer of that, so we're not blocked by you.
At this point, the value of using your hwspinlock framework
is vastly outweighed by the negative effect is has on our mainlining
effort.
Regards,
-- Tim Bird
Senior Software Engineer, Sony Mobile
Architecture Group Chair, CE Workgroup, Linux Foundation
On Mon, Apr 6, 2015 at 7:22 PM, Tim Bird <[email protected]> wrote:
> On Fri, Apr 3, 2015 at 6:55 AM, Ohad Ben-Cohen <[email protected]> wrote:
>> On Thu, Apr 2, 2015 at 9:11 PM, Tim Bird <[email protected]> wrote:
>>> On Wed, Apr 1, 2015 at 9:40 PM, Ohad Ben-Cohen <[email protected]> wrote:
>>>> Sorry, I can't take this without a DT ack.
>>>
>>> Hmmm.
>>>
>>> The policy seems to be:
>>> "For driver (not subsystem) bindings: If you are comfortable with the
>>> binding, and it hasn't received an Acked-by from the devicetree
>>> maintainers after a few weeks, go ahead and take it."
>>>
>>> The syscon property is only relative to the qcom hwspinlock driver,
>>> (unless I'm missing something) and both Qualcomm and Sony devs are
>>> OK with it. So while an ACK from the DT side would be nice, I don't
>>> think it's required. This is exactly the type of delay that is really
>>> holding up a lot of out-of-tree code.
>>
>> Sorry, I do prefer to make sure Mark is OK with this devicetree patch,
>> especially since it wasn't clear whether Mark is entirely comfortable
>> with it in his last response.
>
> Just to be clear - do you personally have any objections to the patch?
No, but this patch is for a folder I don't maintain so I prefer
someone who does to take a look.
Mark did take a look, and said he's confused by this patch (see this thread).
Do you want me to ignore him and just send it to Linus anyway?
On Mon, Apr 6, 2015 at 9:31 AM, Ohad Ben-Cohen <[email protected]> wrote:
> On Mon, Apr 6, 2015 at 7:22 PM, Tim Bird <[email protected]> wrote:
>> On Fri, Apr 3, 2015 at 6:55 AM, Ohad Ben-Cohen <[email protected]> wrote:
>>> On Thu, Apr 2, 2015 at 9:11 PM, Tim Bird <[email protected]> wrote:
>>>> On Wed, Apr 1, 2015 at 9:40 PM, Ohad Ben-Cohen <[email protected]> wrote:
>>>>> Sorry, I can't take this without a DT ack.
>>>>
>>>> Hmmm.
>>>>
>>>> The policy seems to be:
>>>> "For driver (not subsystem) bindings: If you are comfortable with the
>>>> binding, and it hasn't received an Acked-by from the devicetree
>>>> maintainers after a few weeks, go ahead and take it."
>>>>
>>>> The syscon property is only relative to the qcom hwspinlock driver,
>>>> (unless I'm missing something) and both Qualcomm and Sony devs are
>>>> OK with it. So while an ACK from the DT side would be nice, I don't
>>>> think it's required. This is exactly the type of delay that is really
>>>> holding up a lot of out-of-tree code.
>>>
>>> Sorry, I do prefer to make sure Mark is OK with this devicetree patch,
>>> especially since it wasn't clear whether Mark is entirely comfortable
>>> with it in his last response.
>>
>> Just to be clear - do you personally have any objections to the patch?
>
> No, but this patch is for a folder I don't maintain so I prefer
> someone who does to take a look.
>
> Mark did take a look, and said he's confused by this patch (see this thread).
>
> Do you want me to ignore him and just send it to Linus anyway?
Ohad, Tim,
For this patch to be useful to us we also need Suman's DT patch, so we
should try to get them both in asap.
Based on the long discussion we had on one of the previous iterations
of Suman's DT binding, with the DT maintainers I believe that it would
be fine to move along and sent Suman's patches to Linus - without an
explicit Ack from the DT guys (or did I just miss one?).
Regarding this patch I agree with Ohad that it would be good if we
verify that Mark's question is answered before moving on. So I will
reach out to him to see if he has any remaining concerns.
Regards,
Bjorn
On Mon, Apr 6, 2015 at 7:48 PM, Bjorn Andersson <[email protected]> wrote:
> Based on the long discussion we had on one of the previous iterations
> of Suman's DT binding, with the DT maintainers I believe that it would
> be fine to move along and sent Suman's patches to Linus - without an
> explicit Ack from the DT guys (or did I just miss one?).
Thanks Bjorn.
Mark, are you OK with the latest iteration from Suman? it would be
nice to get your +1 just to make sure we don't merge stuff you're
uncomfortable with.
Thanks!
Ohad.
On Mon, Apr 6, 2015 at 10:04 PM, Ohad Ben-Cohen <[email protected]> wrote:
> Mark, are you OK with the latest iteration from Suman? it would be
> nice to get your +1 just to make sure we don't merge stuff you're
> uncomfortable with.
Quick update:
As Tim pointed out, we can move forward with the driver binding patch
according to the process described under II.2 of [1]. Both Bjorn and
myself would still prefer to make sure Mark is satisfied with the
response Bjorn sent to Mark's question, but we understand if Mark is
swamped and we eventually would proceed according to the DT's
submitting-patches guidance below. Tim, thanks for pointing that out
as I wasn't aware of this.
What we probably do need a DT ack on is the hwspinlock subsystem
binding submitted by Suman, again according to the process described
under II.2 of [1]: "Subsystem bindings (anything affecting more than a
single device): then getting a devicetree maintainer to review it is
required".
Mark and Rob: thanks so much for all your help so far as you have
substantially helped shaping the hwspinlock binding. Please let us
know if you are satisfied with Suman's latest iteration, still prefer
to take another look at it, or are too swamped. If the latter, then
maybe we can ask Kumar to take a look, as this seems to be blocking
Qualcomm's upstream roadmap.
Thanks,
Ohad.
[1] Documentation/devicetree/bindings/submitting-patches.txt
> On Feb 27, 2015, at 4:30 PM, Bjorn Andersson <[email protected]> wrote:
>
> Add binding documentation for the Qualcomm Hardware Mutex.
>
> Signed-off-by: Bjorn Andersson <[email protected]>
> —
>
Acked-by: Kumar Gala <[email protected]>
- k
--
Qualcomm Innovation Center, Inc.
The Qualcomm Innovation Center, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project
On Tue, Apr 14, 2015 at 10:18 PM, Kumar Gala <[email protected]> wrote:
>> On Feb 27, 2015, at 4:30 PM, Bjorn Andersson <[email protected]> wrote:
>>
>> Add binding documentation for the Qualcomm Hardware Mutex.
>>
>> Signed-off-by: Bjorn Andersson <[email protected]>
>> —
>>
>
> Acked-by: Kumar Gala <[email protected]>
Perfect, thanks a lot, Kumar.
On Mon, Apr 13, 2015 at 5:23 AM, Ohad Ben-Cohen <[email protected]> wrote:
> On Mon, Apr 6, 2015 at 10:04 PM, Ohad Ben-Cohen <[email protected]> wrote:
>> Mark, are you OK with the latest iteration from Suman? it would be
>> nice to get your +1 just to make sure we don't merge stuff you're
>> uncomfortable with.
>
> Quick update:
>
> As Tim pointed out, we can move forward with the driver binding patch
> according to the process described under II.2 of [1]. Both Bjorn and
> myself would still prefer to make sure Mark is satisfied with the
> response Bjorn sent to Mark's question, but we understand if Mark is
> swamped and we eventually would proceed according to the DT's
> submitting-patches guidance below. Tim, thanks for pointing that out
> as I wasn't aware of this.
>
> What we probably do need a DT ack on is the hwspinlock subsystem
> binding submitted by Suman, again according to the process described
> under II.2 of [1]: "Subsystem bindings (anything affecting more than a
> single device): then getting a devicetree maintainer to review it is
> required".
Right, you can't really merge this driver binding until the binding is
done and acked. Your binding would be referring to a binding doc that
doesn't exist in the tree if you merge this.
> Mark and Rob: thanks so much for all your help so far as you have
> substantially helped shaping the hwspinlock binding. Please let us
> know if you are satisfied with Suman's latest iteration, still prefer
> to take another look at it, or are too swamped. If the latter, then
> maybe we can ask Kumar to take a look, as this seems to be blocking
> Qualcomm's upstream roadmap.
While I think the binding looks fine, I'm not going to go around Mark
either. He's invested the most thought in this (of DT maintainers) and
should be the one to ack it. The part I'm not clear what the purpose
of "pool-hwlock" was.
Rob
On Fri, Feb 27, 2015 at 10:30:16PM +0000, Bjorn Andersson wrote:
> Add binding documentation for the Qualcomm Hardware Mutex.
>
> Signed-off-by: Bjorn Andersson <[email protected]>
> ---
>
> I think the conclusion on the dt binding discussion for hwspinlocks was that
> we're down to having the #hwlock-cells intact. So this version includes that,
> but non of the other previously discussed properties.
>
> Changes since v5:
> - Extracted the dt binding documentation into a separate patch
> - Moved the driver to consume a syscon
> - Dropped previously suggested generic hwlock dt bindings
>
> .../devicetree/bindings/hwlock/qcom-hwspinlock.txt | 39 ++++++++++++++++++++++
> 1 file changed, 39 insertions(+)
> create mode 100644 Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
>
> diff --git a/Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt b/Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
> new file mode 100644
> index 0000000..28ade7d
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/hwlock/qcom-hwspinlock.txt
> @@ -0,0 +1,39 @@
> +Qualcomm Hardware Mutex Block:
> +
> +The hardware block provides mutexes utilized between different processors on
> +the SoC as part of the communication protocol used by these processors.
> +
> +- compatible:
> + Usage: required
> + Value type: <string>
> + Definition: must be one of:
> + "qcom,sfpb-mutex",
> + "qcom,tcsr-mutex"
> +
> +- syscon:
> + Usage: required
> + Value type: <prop-encoded-array>
> + Definition: one cell containing:
> + syscon phandle
> + offset of the hwmutex block within the syscon
> + stride of the hwmutex registers
> +
> +- #hwlock-cells:
> + Usage: required
> + Value type: <u32>
> + Definition: must be 1, the specified cell represent the lock id
> + (hwlock standard property, see hwlock.txt)
I guess that this is a linear ID, matching some ID in a data sheet? It
would be nice to have something a little more explicit as to what the
valid values of the cell are.
Otherwise this looks fine, and seems to pair fine with the generic
bindings. The nit above can be corrected with a followup.
Acked-by: Mark Rutland <[email protected]>
Mark.
> +
> +Example:
> +
> + tcsr: syscon@1a400000 {
> + compatible = "qcom,tcsr-msm8974", "syscon";
> + reg = <0xfd484000 0x2000>;
> + };
> +
> + hwlock@fd484000 {
> + compatible = "qcom,tcsr-mutex";
> + syscon = <&tcsr 0 0x80>;
> +
> + #hwlock-cells = <1>;
> + };
> --
> 1.8.2.2
>
>
> The part I'm not clear what the purpose of "pool-hwlock" was.
One use-case was a set of hwlocks having no fixed purpose, and being
available for dynamic allocation between the OS and other entities (e.g.
some RTOS on another core).
The set of locks forming a reusable pool, and any information associated
with them (e.g. the logical IDs used by the other entity) are a property
of that interface rather than the hwlock provider.
So you'd describe those pools of locks in some interface-specific
manner, consuming hwlocks from a set of hwlock providers.
Hopefully that doesn't maek things less clear...
Mark.