linux-crypto.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Sverdlin, Alexander (Nokia - DE/Ulm)" <alexander.sverdlin@nokia.com>
To: "linux-crypto@vger.kernel.org" <linux-crypto@vger.kernel.org>
Cc: "Sverdlin,
	Alexander (Nokia - DE/Ulm)" <alexander.sverdlin@nokia.com>,
	Matt Mackall <mpm@selenic.com>,
	Herbert Xu <herbert@gondor.apana.org.au>,
	Vitaly Andrianov <vitalya@ti.com>
Subject: [PATCH] hwrng: ks-sa: Add minimum sleep time before ready-polling
Date: Wed, 6 Nov 2019 09:30:49 +0000	[thread overview]
Message-ID: <20191106093019.117233-1-alexander.sverdlin@nokia.com> (raw)

From: Alexander Sverdlin <alexander.sverdlin@nokia.com>

Current polling timeout is 25 us. The hardware is currently configured to
harvest the entropy for 81920 us. This leads to timeouts even during
blocking read (wait=1).

Log snippet:
[    5.727589] [<c040ffcc>] (ks_sa_rng_probe) from [<c04181e8>] (platform_drv_probe+0x58/0xb4)
...
[    5.727805] hwrng: no data available
...
[   13.157016] random: systemd: uninitialized urandom read (16 bytes read)
[   13.157033] systemd[1]: Initializing machine ID from random generator.
...
[   15.848770] random: fast init done
...
[   15.848807] random: crng init done

After the patch:
[    6.223534] random: systemd: uninitialized urandom read (16 bytes read)
[    6.223551] systemd[1]: Initializing machine ID from random generator.
...
[    6.876075] random: fast init done
...
[    6.954200] random: systemd: uninitialized urandom read (16 bytes read)
[    6.955244] random: systemd: uninitialized urandom read (16 bytes read)
...
[    7.121948] random: crng init done

Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com>
---
 drivers/char/hw_random/ks-sa-rng.c | 38 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 38 insertions(+)

diff --git a/drivers/char/hw_random/ks-sa-rng.c b/drivers/char/hw_random/ks-sa-rng.c
index a674300..4b223cb 100644
--- a/drivers/char/hw_random/ks-sa-rng.c
+++ b/drivers/char/hw_random/ks-sa-rng.c
@@ -21,6 +21,7 @@
 #include <linux/of.h>
 #include <linux/of_address.h>
 #include <linux/delay.h>
+#include <linux/timekeeping.h>
 
 #define SA_CMD_STATUS_OFS			0x8
 
@@ -85,13 +86,36 @@ struct ks_sa_rng {
 	struct clk	*clk;
 	struct regmap	*regmap_cfg;
 	struct trng_regs *reg_rng;
+	u64 ready_ts;
+	unsigned int refill_delay_ns;
 };
 
+static unsigned int cycles_to_ns(unsigned long clk_rate, unsigned int cycles)
+{
+	return DIV_ROUND_UP_ULL((TRNG_DEF_CLK_DIV_CYCLES + 1) * 1000000000ull *
+				cycles, clk_rate);
+}
+
+static unsigned int startup_delay_ns(unsigned long clk_rate)
+{
+	if (!TRNG_DEF_STARTUP_CYCLES)
+		return cycles_to_ns(clk_rate, BIT(24));
+	return cycles_to_ns(clk_rate, 256 * TRNG_DEF_STARTUP_CYCLES);
+}
+
+static unsigned int refill_delay_ns(unsigned long clk_rate)
+{
+	if (!TRNG_DEF_MAX_REFILL_CYCLES)
+		return cycles_to_ns(clk_rate, BIT(24));
+	return cycles_to_ns(clk_rate, 256 * TRNG_DEF_MAX_REFILL_CYCLES);
+}
+
 static int ks_sa_rng_init(struct hwrng *rng)
 {
 	u32 value;
 	struct device *dev = (struct device *)rng->priv;
 	struct ks_sa_rng *ks_sa_rng = dev_get_drvdata(dev);
+	unsigned long clk_rate = clk_get_rate(ks_sa_rng->clk);
 
 	/* Enable RNG module */
 	regmap_write_bits(ks_sa_rng->regmap_cfg, SA_CMD_STATUS_OFS,
@@ -120,6 +144,10 @@ static int ks_sa_rng_init(struct hwrng *rng)
 	value |= TRNG_CNTL_REG_TRNG_ENABLE;
 	writel(value, &ks_sa_rng->reg_rng->control);
 
+	ks_sa_rng->refill_delay_ns = refill_delay_ns(clk_rate);
+	ks_sa_rng->ready_ts = ktime_get_ns() +
+			      startup_delay_ns(clk_rate);
+
 	return 0;
 }
 
@@ -144,6 +172,7 @@ static int ks_sa_rng_data_read(struct hwrng *rng, u32 *data)
 	data[1] = readl(&ks_sa_rng->reg_rng->output_h);
 
 	writel(TRNG_INTACK_REG_READY, &ks_sa_rng->reg_rng->intack);
+	ks_sa_rng->ready_ts = ktime_get_ns() + ks_sa_rng->refill_delay_ns;
 
 	return sizeof(u32) * 2;
 }
@@ -152,10 +181,19 @@ static int ks_sa_rng_data_present(struct hwrng *rng, int wait)
 {
 	struct device *dev = (struct device *)rng->priv;
 	struct ks_sa_rng *ks_sa_rng = dev_get_drvdata(dev);
+	u64 now = ktime_get_ns();
 
 	u32	ready;
 	int	j;
 
+	if (wait && now < ks_sa_rng->ready_ts) {
+		/* Max delay expected here is 81920000 ns */
+		unsigned long min_delay =
+			DIV_ROUND_UP((u32)(ks_sa_rng->ready_ts - now), 1000);
+
+		usleep_range(min_delay, min_delay + SA_RNG_DATA_RETRY_DELAY);
+	}
+
 	for (j = 0; j < SA_MAX_RNG_DATA_RETRIES; j++) {
 		ready = readl(&ks_sa_rng->reg_rng->status);
 		ready &= TRNG_STATUS_REG_READY;
-- 
2.4.6


             reply	other threads:[~2019-11-06  9:30 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-11-06  9:30 Sverdlin, Alexander (Nokia - DE/Ulm) [this message]
2019-11-15  6:06 ` [PATCH] hwrng: ks-sa: Add minimum sleep time before ready-polling Herbert Xu
2019-11-16  7:32   ` Eric Biggers
2019-11-17  0:42     ` Herbert Xu
2019-11-17  0:43       ` [PATCH] hwrng: ks-sa - Enable COMPILE_TEST Herbert Xu
2019-11-18 12:05         ` Sverdlin, Alexander (Nokia - DE/Ulm)

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20191106093019.117233-1-alexander.sverdlin@nokia.com \
    --to=alexander.sverdlin@nokia.com \
    --cc=herbert@gondor.apana.org.au \
    --cc=linux-crypto@vger.kernel.org \
    --cc=mpm@selenic.com \
    --cc=vitalya@ti.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
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).