All of lore.kernel.org
 help / color / mirror / Atom feed
From: Chen-Yu Tsai <wens@kernel.org>
To: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>,
	Maxime Ripard <maxime.ripard@bootlin.com>
Cc: devicetree@vger.kernel.org, linux-sunxi@googlegroups.com,
	Chen-Yu Tsai <wens@csie.org>,
	linux-arm-kernel@lists.infradead.org,
	linux-kernel@vger.kernel.org
Subject: [PATCH 2/6] nvmem: sunxi_sid: Optimize register read-out method
Date: Mon, 18 Mar 2019 15:33:50 +0800	[thread overview]
Message-ID: <20190318073354.12151-3-wens@kernel.org> (raw)
In-Reply-To: <20190318073354.12151-1-wens@kernel.org>

From: Chen-Yu Tsai <wens@csie.org>

SID cells are 32-bit aligned, and a multiple of 32 bits in length. The
only outlier is the thermal sensor calibration data, which is 16 bits
per sensor. However a whole 64 bits is allocated for this purpose, so
we could consider it conforming to the rule above.

Also, the register read-out method assumes native endian, unlike the
direct MMIO method, which assumes big endian. Thus no endian conversion
is involved.

Under these assumptions, the register read-out method can be slightly
optimized. Instead of reading one word then discarding 3 bytes, read
the whole word directly into the buffer. However, for reads under 4
bytes or trailing bytes, we still use a scratch buffer to extract the
requested bytes.

We could go one step further if .word_size was 4, but changing that
would affect the sysfs interface's behavior.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
---
 drivers/nvmem/sunxi_sid.c | 38 ++++++++++++++++++--------------------
 1 file changed, 18 insertions(+), 20 deletions(-)

diff --git a/drivers/nvmem/sunxi_sid.c b/drivers/nvmem/sunxi_sid.c
index 704c35edf796..15fbfab62595 100644
--- a/drivers/nvmem/sunxi_sid.c
+++ b/drivers/nvmem/sunxi_sid.c
@@ -115,36 +115,34 @@ static int sun8i_sid_register_readout(const struct sunxi_sid *sid,
  * to be not reliable at all.
  * Read by the registers instead.
  */
-static int sun8i_sid_read_byte_by_reg(const struct sunxi_sid *sid,
-				      const unsigned int offset,
-				      u8 *out)
-{
-	u32 word;
-	int ret;
-
-	ret = sun8i_sid_register_readout(sid, offset & ~0x03, &word);
-
-	if (ret)
-		return ret;
-
-	*out = (word >> ((offset & 0x3) * 8)) & 0xff;
-
-	return 0;
-}
-
 static int sun8i_sid_read_by_reg(void *context, unsigned int offset,
 				 void *val, size_t bytes)
 {
 	struct sunxi_sid *sid = context;
-	u8 *buf = val;
+	u32 word;
 	int ret;
 
-	while (bytes--) {
-		ret = sun8i_sid_read_byte_by_reg(sid, offset++, buf++);
+	/* .stride = 4 so offset is guaranteed to be aligned */
+	while (bytes >= 4) {
+		ret = sun8i_sid_register_readout(sid, offset, val);
 		if (ret)
 			return ret;
+
+		val += 4;
+		offset += 4;
+		bytes -= 4;
 	}
 
+	if (!bytes)
+		return 0;
+
+	/* Handle any trailing bytes */
+	ret = sun8i_sid_register_readout(sid, offset, &word);
+	if (ret)
+		return ret;
+
+	memcpy(val, &word, bytes);
+
 	return 0;
 }
 
-- 
2.20.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

WARNING: multiple messages have this Message-ID (diff)
From: Chen-Yu Tsai <wens-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>
To: Srinivas Kandagatla
	<srinivas.kandagatla-QSEj5FYQhm4dnm+yROfE0A@public.gmane.org>,
	Maxime Ripard
	<maxime.ripard-LDxbnhwyfcJBDgjK7y7TUQ@public.gmane.org>
Cc: linux-sunxi-/JYPxA39Uh5TLH3MbocFFw@public.gmane.org,
	Chen-Yu Tsai <wens-jdAy2FN1RRM@public.gmane.org>,
	linux-arm-kernel-IAPFreCvJWM7uuMidbF8XUB+6BGkLq7r@public.gmane.org,
	devicetree-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Subject: [PATCH 2/6] nvmem: sunxi_sid: Optimize register read-out method
Date: Mon, 18 Mar 2019 15:33:50 +0800	[thread overview]
Message-ID: <20190318073354.12151-3-wens@kernel.org> (raw)
In-Reply-To: <20190318073354.12151-1-wens-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>

From: Chen-Yu Tsai <wens-jdAy2FN1RRM@public.gmane.org>

SID cells are 32-bit aligned, and a multiple of 32 bits in length. The
only outlier is the thermal sensor calibration data, which is 16 bits
per sensor. However a whole 64 bits is allocated for this purpose, so
we could consider it conforming to the rule above.

Also, the register read-out method assumes native endian, unlike the
direct MMIO method, which assumes big endian. Thus no endian conversion
is involved.

Under these assumptions, the register read-out method can be slightly
optimized. Instead of reading one word then discarding 3 bytes, read
the whole word directly into the buffer. However, for reads under 4
bytes or trailing bytes, we still use a scratch buffer to extract the
requested bytes.

We could go one step further if .word_size was 4, but changing that
would affect the sysfs interface's behavior.

Signed-off-by: Chen-Yu Tsai <wens-jdAy2FN1RRM@public.gmane.org>
---
 drivers/nvmem/sunxi_sid.c | 38 ++++++++++++++++++--------------------
 1 file changed, 18 insertions(+), 20 deletions(-)

diff --git a/drivers/nvmem/sunxi_sid.c b/drivers/nvmem/sunxi_sid.c
index 704c35edf796..15fbfab62595 100644
--- a/drivers/nvmem/sunxi_sid.c
+++ b/drivers/nvmem/sunxi_sid.c
@@ -115,36 +115,34 @@ static int sun8i_sid_register_readout(const struct sunxi_sid *sid,
  * to be not reliable at all.
  * Read by the registers instead.
  */
-static int sun8i_sid_read_byte_by_reg(const struct sunxi_sid *sid,
-				      const unsigned int offset,
-				      u8 *out)
-{
-	u32 word;
-	int ret;
-
-	ret = sun8i_sid_register_readout(sid, offset & ~0x03, &word);
-
-	if (ret)
-		return ret;
-
-	*out = (word >> ((offset & 0x3) * 8)) & 0xff;
-
-	return 0;
-}
-
 static int sun8i_sid_read_by_reg(void *context, unsigned int offset,
 				 void *val, size_t bytes)
 {
 	struct sunxi_sid *sid = context;
-	u8 *buf = val;
+	u32 word;
 	int ret;
 
-	while (bytes--) {
-		ret = sun8i_sid_read_byte_by_reg(sid, offset++, buf++);
+	/* .stride = 4 so offset is guaranteed to be aligned */
+	while (bytes >= 4) {
+		ret = sun8i_sid_register_readout(sid, offset, val);
 		if (ret)
 			return ret;
+
+		val += 4;
+		offset += 4;
+		bytes -= 4;
 	}
 
+	if (!bytes)
+		return 0;
+
+	/* Handle any trailing bytes */
+	ret = sun8i_sid_register_readout(sid, offset, &word);
+	if (ret)
+		return ret;
+
+	memcpy(val, &word, bytes);
+
 	return 0;
 }
 
-- 
2.20.1

WARNING: multiple messages have this Message-ID (diff)
From: Chen-Yu Tsai <wens@kernel.org>
To: Srinivas Kandagatla <srinivas.kandagatla@linaro.org>,
	Maxime Ripard <maxime.ripard@bootlin.com>
Cc: linux-sunxi@googlegroups.com, Chen-Yu Tsai <wens@csie.org>,
	linux-arm-kernel@lists.infradead.org, devicetree@vger.kernel.org,
	linux-kernel@vger.kernel.org
Subject: [PATCH 2/6] nvmem: sunxi_sid: Optimize register read-out method
Date: Mon, 18 Mar 2019 15:33:50 +0800	[thread overview]
Message-ID: <20190318073354.12151-3-wens@kernel.org> (raw)
In-Reply-To: <20190318073354.12151-1-wens@kernel.org>

From: Chen-Yu Tsai <wens@csie.org>

SID cells are 32-bit aligned, and a multiple of 32 bits in length. The
only outlier is the thermal sensor calibration data, which is 16 bits
per sensor. However a whole 64 bits is allocated for this purpose, so
we could consider it conforming to the rule above.

Also, the register read-out method assumes native endian, unlike the
direct MMIO method, which assumes big endian. Thus no endian conversion
is involved.

Under these assumptions, the register read-out method can be slightly
optimized. Instead of reading one word then discarding 3 bytes, read
the whole word directly into the buffer. However, for reads under 4
bytes or trailing bytes, we still use a scratch buffer to extract the
requested bytes.

We could go one step further if .word_size was 4, but changing that
would affect the sysfs interface's behavior.

Signed-off-by: Chen-Yu Tsai <wens@csie.org>
---
 drivers/nvmem/sunxi_sid.c | 38 ++++++++++++++++++--------------------
 1 file changed, 18 insertions(+), 20 deletions(-)

diff --git a/drivers/nvmem/sunxi_sid.c b/drivers/nvmem/sunxi_sid.c
index 704c35edf796..15fbfab62595 100644
--- a/drivers/nvmem/sunxi_sid.c
+++ b/drivers/nvmem/sunxi_sid.c
@@ -115,36 +115,34 @@ static int sun8i_sid_register_readout(const struct sunxi_sid *sid,
  * to be not reliable at all.
  * Read by the registers instead.
  */
-static int sun8i_sid_read_byte_by_reg(const struct sunxi_sid *sid,
-				      const unsigned int offset,
-				      u8 *out)
-{
-	u32 word;
-	int ret;
-
-	ret = sun8i_sid_register_readout(sid, offset & ~0x03, &word);
-
-	if (ret)
-		return ret;
-
-	*out = (word >> ((offset & 0x3) * 8)) & 0xff;
-
-	return 0;
-}
-
 static int sun8i_sid_read_by_reg(void *context, unsigned int offset,
 				 void *val, size_t bytes)
 {
 	struct sunxi_sid *sid = context;
-	u8 *buf = val;
+	u32 word;
 	int ret;
 
-	while (bytes--) {
-		ret = sun8i_sid_read_byte_by_reg(sid, offset++, buf++);
+	/* .stride = 4 so offset is guaranteed to be aligned */
+	while (bytes >= 4) {
+		ret = sun8i_sid_register_readout(sid, offset, val);
 		if (ret)
 			return ret;
+
+		val += 4;
+		offset += 4;
+		bytes -= 4;
 	}
 
+	if (!bytes)
+		return 0;
+
+	/* Handle any trailing bytes */
+	ret = sun8i_sid_register_readout(sid, offset, &word);
+	if (ret)
+		return ret;
+
+	memcpy(val, &word, bytes);
+
 	return 0;
 }
 
-- 
2.20.1


  parent reply	other threads:[~2019-03-18  7:34 UTC|newest]

Thread overview: 43+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-03-18  7:33 [PATCH 0/6] nvmem: sunxi_sid: native format and A83T/H5 support Chen-Yu Tsai
2019-03-18  7:33 ` Chen-Yu Tsai
2019-03-18  7:33 ` [PATCH 1/6] nvmem: sunxi_sid: Read out SID for randomness without looping Chen-Yu Tsai
2019-03-18  7:33   ` Chen-Yu Tsai
2019-03-18  7:33   ` Chen-Yu Tsai
2019-03-18  7:33 ` Chen-Yu Tsai [this message]
2019-03-18  7:33   ` [PATCH 2/6] nvmem: sunxi_sid: Optimize register read-out method Chen-Yu Tsai
2019-03-18  7:33   ` Chen-Yu Tsai
2019-03-18  7:33 ` [PATCH 3/6] nvmem: sunxi_sid: Dynamically allocate nvmem_config structure Chen-Yu Tsai
2019-03-18  7:33   ` Chen-Yu Tsai
2019-03-18  7:33   ` Chen-Yu Tsai
2019-03-18  7:33 ` [PATCH 4/6] nvmem: sunxi_sid: Read out data in native format Chen-Yu Tsai
2019-03-18  7:33   ` Chen-Yu Tsai
2019-03-18  7:33   ` Chen-Yu Tsai
2019-03-18  8:42   ` Maxime Ripard
2019-03-18  8:42     ` Maxime Ripard
2019-03-18  8:42     ` Maxime Ripard
2019-03-18  8:45     ` Chen-Yu Tsai
2019-03-18  8:45       ` Chen-Yu Tsai
2019-03-18  8:45       ` Chen-Yu Tsai
2019-03-18  8:57       ` Maxime Ripard
2019-03-18  8:57         ` Maxime Ripard
2019-03-18  8:57         ` Maxime Ripard
2019-03-18  9:09         ` Chen-Yu Tsai
2019-03-18  9:09           ` Chen-Yu Tsai
2019-03-18  9:25           ` Maxime Ripard
2019-03-18  9:25             ` Maxime Ripard
2019-03-18  9:25             ` Maxime Ripard
2019-03-18  7:33 ` [PATCH 5/6] nvmem: sunxi_sid: Support SID on A83T and H5 Chen-Yu Tsai
2019-03-18  7:33   ` Chen-Yu Tsai
2019-03-18  7:33 ` [PATCH 6/6] ARM: dts: sunxi: h3/h5: Add device node for SID Chen-Yu Tsai
2019-03-18  7:33   ` Chen-Yu Tsai
2019-03-18  7:33   ` Chen-Yu Tsai
2019-03-19  1:55   ` [linux-sunxi] " Samuel Holland
2019-03-19  1:55     ` Samuel Holland
2019-03-19  1:55     ` Samuel Holland
2019-03-19  3:17     ` Chen-Yu Tsai
2019-03-19  3:17       ` Chen-Yu Tsai
2019-03-19  3:17       ` Chen-Yu Tsai
2019-03-20 14:25 ` [PATCH 0/6] nvmem: sunxi_sid: native format and A83T/H5 support Srinivas Kandagatla
2019-03-20 14:25   ` Srinivas Kandagatla
2019-03-21  9:06   ` Chen-Yu Tsai
2019-03-21  9:06     ` Chen-Yu Tsai

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=20190318073354.12151-3-wens@kernel.org \
    --to=wens@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=linux-arm-kernel@lists.infradead.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-sunxi@googlegroups.com \
    --cc=maxime.ripard@bootlin.com \
    --cc=srinivas.kandagatla@linaro.org \
    --cc=wens@csie.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.