add support of hardware random generator on MediaTek MT7622

From: Sean Wang <[email protected]>

Changes since v1:
- update the bindings with the specific "mediatek,mt7622-rng" instead of the generic one as "mediatek,generic-rng"

The series add support of hardware RNG on MediaTek MT7622 and , runtime PM support and add me as the maintainer for the existing and following chipset.

Sean Wang (3):
dt-bindings: rng: add MediaTek MT7622 Hardware Random Generator
bindings
hwrng: mtk - add runtime PM support
MAINTAINERS: add entry for MediaTek Random Number Generator

 Documentation/devicetree/bindings/rng/mtk-rng.txt |  4 ++-
 MAINTAINERS                                       |  5 +++
 drivers/char/hw_random/mtk-rng.c                  | 42 +++++++++++++++++++++++
 3 files changed, 50 insertions(+), 1 deletion(-)

1.9.1

From: Sean Wang <[email protected]>

Document the bindings used by MediaTek MT7622 SoC hardware random number generator.

---
 Documentation/devicetree/bindings/rng/mtk-rng.txt | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index a6d62a2..366b99b 100644
@@ -2,7 +2,9 @@ Device-Tree bindings for Mediatek random number generator
 found in Mediatek SoC family
 
 Required properties:
-- compatible	    : Should be "mediatek,mt7623-rng"
+- compatible	    : Should be
+			"mediatek,mt7622-rng", 	"mediatek,mt7623-rng" : for MT7622
+			"mediatek,mt7623-rng" : for MT7623
 - clocks	    : list of clock specifiers, corresponding to
 		      entries in clock-names property;
 - clock-names	    : Should contain "rng" entries;

1.9.1

From: Sean Wang <[email protected]>

Add runtime PM support. There will be the benefit on SoCs where the clock to the RNG used can be shutdown.

---
 drivers/char/hw_random/mtk-rng.c | 42 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)

index df8eb54..8da7bcf 100644
@@ -25,6 +25,10 @@
 #include <linux/module.h>
 #include <linux/of.h>
 #include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+
+/* Runtime PM autosuspend timeout: */
+#define RNG_AUTOSUSPEND_TIMEOUT		100
 
 #define USEC_POLL			2
 #define TIMEOUT_POLL			20

@@ -90,6 +94,8 @@ static int mtk_rng_read(struct hwrng *rng, void *buf, size_t max, bool wait)
 	struct mtk_rng *priv = to_mtk_rng(rng);
 	int retval = 0;
 
+	pm_runtime_get_sync((struct device *)priv->rng.priv);
+
 	while (max >= sizeof(u32)) {
 		if (!mtk_rng_wait_ready(rng, wait))
 			break;

@@ -100,6 +106,9 @@ static int mtk_rng_read(struct hwrng *rng, void *buf, size_t max, bool wait)
 		max -= sizeof(u32);
 	}
 
+	pm_runtime_mark_last_busy((struct device *)priv->rng.priv);
+	pm_runtime_put_sync_autosuspend((struct device *)priv->rng.priv);
+
 	return retval || !wait ? retval : -EIO;
 }
 

@@ -120,9 +129,12 @@ static int mtk_rng_probe(struct platform_device *pdev)
 		return -ENOMEM;
 
 	priv->rng.name = pdev->name;
+#ifndef CONFIG_PM
 	priv->rng.init = mtk_rng_init;
 	priv->rng.cleanup = mtk_rng_cleanup;
+#endif
 	priv->rng.read = mtk_rng_read;
+	priv->rng.priv = (unsigned long)&pdev->dev;
 
 	priv->clk = devm_clk_get(&pdev->dev, "rng");
 	if (IS_ERR(priv->clk)) {

@@ -142,11 +154,40 @@ static int mtk_rng_probe(struct platform_device *pdev)
 		return ret;
 	}
 
+	dev_set_drvdata(&pdev->dev, priv);
+	pm_runtime_set_autosuspend_delay(&pdev->dev, RNG_AUTOSUSPEND_TIMEOUT);
+	pm_runtime_use_autosuspend(&pdev->dev);
+	pm_runtime_enable(&pdev->dev);
+
 	dev_info(&pdev->dev, "registered RNG driver\n");
 
 	return 0;
 }
 
+#ifdef CONFIG_PM
+static int mtk_rng_runtime_suspend(struct device *dev)
+{
+	struct mtk_rng *priv = dev_get_drvdata(dev);
+
+	mtk_rng_cleanup(&priv->rng);
+
+	return 0;
+}
+
+static int mtk_rng_runtime_resume(struct device *dev)
+{
+	struct mtk_rng *priv = dev_get_drvdata(dev);
+
+	return mtk_rng_init(&priv->rng);
+}
+
+static UNIVERSAL_DEV_PM_OPS(mtk_rng_pm_ops, mtk_rng_runtime_suspend,
+			    mtk_rng_runtime_resume, NULL);
+#define MTK_RNG_PM_OPS (&mtk_rng_pm_ops)
+#else	/* CONFIG_PM */
+#define MTK_RNG_PM_OPS NULL
+#endif	/* CONFIG_PM */
+
 static const struct of_device_id mtk_rng_match[] = {
 	{ .compatible = "mediatek,mt7623-rng" },
 	{},

@@ -157,6 +198,7 @@ static int mtk_rng_probe(struct platform_device *pdev)
 	.probe          = mtk_rng_probe,
 	.driver = {
 		.name = MTK_RNG_DEV,
+		.pm = MTK_RNG_PM_OPS,
 		.of_match_table = mtk_rng_match,
 	},
 };

1.9.1

From: Sean Wang <[email protected]>

I work for MediaTek on maintaining the MediaTek RNG driver for the existing SoCs and adding support for the following SoCs.

---
 MAINTAINERS | 5 +++++
 1 file changed, 5 insertions(+)

index cf86519..da473d0 100644
@@ -8262,6 +8262,11 @@ L:	linux-wireless at vger.kernel.org
 S:	Maintained
 F:	drivers/net/wireless/mediatek/mt7601u/
 
+MEDIATEK RANDOM NUMBER GENERATOR SUPPORT
+M:      Sean Wang <[email protected]>
+S:      Maintained
+F:      drivers/char/hw_random/mtk-rng.c
+
 MEGACHIPS STDPXXXX-GE-B850V3-FW LVDS/DP++ BRIDGES
 M:	Peter Senna Tschudin <[email protected]>
 M:	Martin Donnelly <[email protected]>

1.9.1

All applied. Thanks.

Email: Herbert Xu <[email protected]>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt

Hi Herbert,

thanks for effort reviewing on those patches.

By the way, also loop in Torsten

Could you kindly guide me how to determine appropriate rng->ops.quality value used by the driver?

There is less clues since the value is not being set in most drivers. But good value decided would allow feeding fresh entropy data with the hwrng so i would like add it in the next patch.

I have tested with rngtest on mtk-cir and the result is got as the below log shown. If the rngtest always gives the result for success rate over 99.8%, can I set the rng->ops.quality 998?

rngtest: starting FIPS tests...
rngtest: bits received from input: 20000032
rngtest: FIPS 140-2 successes: 998
rngtest: FIPS 140-2 failures: 2
rngtest: FIPS 140-2(2001-10-10) Monobit: 1
rngtest: FIPS 140-2(2001-10-10) Poker: 0
rngtest: FIPS 140-2(2001-10-10) Runs: 1
rngtest: FIPS 140-2(2001-10-10) Long run: 0
rngtest: FIPS 140-2(2001-10-10) Continuous run: 0
rngtest: input channel speed: (min=715.902; avg=3846.234; max=3906250.000)Kibits/s
rngtest: FIPS tests speed: (min=35.453; avg=37.862; max=38.300)Mibits/s
rngtest: Program run time: 5591758 microseconds

rngtest: starting FIPS tests...
rngtest: bits received from input: 20000032
rngtest: FIPS 140-2 successes: 1000
rngtest: FIPS 140-2 failures: 0
rngtest: FIPS 140-2(2001-10-10) Monobit: 0
rngtest: FIPS 140-2(2001-10-10) Poker: 0
rngtest: FIPS 140-2(2001-10-10) Runs: 0
rngtest: FIPS 140-2(2001-10-10) Long run: 0
rngtest: FIPS 140-2(2001-10-10) Continuous run: 0
rngtest: input channel speed: (min=715.588; avg=4424.787; max=4882812.500)Kibits/s
rngtest: FIPS tests speed: (min=35.785; avg=37.859; max=38.223)Mibits/s
rngtest: Program run time: 4947950 microseconds

No! You'd have to determine the failure threshold of the test and apply some math to find a lower boundary of your RNG's entropy.

What the quality is for: your RNG produces bits, but not all of them are completely independent of each other i.e. not completely random. So you simply lower the quality rating to express the net entropy contained in the data stream.

Torsten

Hi Torsten,

Understood, appreciate your quick and clear explanation.

For the math, is there existing any well-known or recommended open source software assisting identify the lower boundary of RNG entropy? I think the logic should be common for all RNGs.

thanks again

Sean

Sean