lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20240223-spi-omap2-mcspi-multi-mode-v2-3-afe94476b9c3@bootlin.com>
Date: Fri, 23 Feb 2024 10:32:13 +0100
From: Louis Chauvet <louis.chauvet@...tlin.com>
To: Mark Brown <broonie@...nel.org>
Cc: linux-spi@...r.kernel.org, linux-kernel@...r.kernel.org, 
 thomas.petazzoni@...tlin.com, Miquel Raynal <miquel.raynal@...tlin.com>, 
 yen-mei.goh@...sight.com, koon-kee.lie@...sight.com, 
 jeremie.dautheribes@...tlin.com, Louis Chauvet <louis.chauvet@...tlin.com>
Subject: [PATCH v2 3/3] spi: omap2-mcpsi: Enable MULTI-mode in more
 situations

Enable multimode when a transfer of multiple small words can be
transformed in a transfer with a single bigger word. This is allowed as
long as the result on the cable is the same, so word_delay must be zero.

Signed-off-by: Louis Chauvet <louis.chauvet@...tlin.com>

---

I am not sure if this is the best place to put such "optimization". I
think this improvment should be in the core, as it is not depending on the
driver itself, but I think Mark suggested something like this so if that
fits to what was expected, I am happy to share this small improvement.
---
 drivers/spi/spi-omap2-mcspi.c | 16 ++++++++++++++--
 1 file changed, 14 insertions(+), 2 deletions(-)

diff --git a/drivers/spi/spi-omap2-mcspi.c b/drivers/spi/spi-omap2-mcspi.c
index 36075c4416d5..f64cc8cc7587 100644
--- a/drivers/spi/spi-omap2-mcspi.c
+++ b/drivers/spi/spi-omap2-mcspi.c
@@ -1397,14 +1397,26 @@ static int omap2_mcspi_prepare_message(struct spi_controller *ctlr,
 		else
 			bits_per_word = tr->bits_per_word;
 
-	/* Only a single channel can have the FORCE bit enabled
 		/*
 		 * Check if this transfer contains only one word;
+		 * OR contains 1 to 4 words, with bits_per_word == 8 and no delay between each word
+		 * OR contains 1 to 2 words, with bits_per_word == 16 and no delay between each word
+		 *
+		 * If one of the two last case is true, this also change the bits_per_word of this
+		 * transfer to make it a bit faster.
+		 * It's not an issue to change the bits_per_word here even if the multi-mode is not
+		 * applicable for this message, the signal on the wire will be the same.
 		 */
 		if (bits_per_word < 8 && tr->len == 1) {
 			/* multi-mode is applicable, only one word (1..7 bits) */
+		} else if (tr->word_delay.value == 0 && bits_per_word == 8 && tr->len <= 4) {
+			/* multi-mode is applicable, only one "bigger" word (8,16,24,32 bits) */
+			tr->bits_per_word = tr->len * bits_per_word;
+		} else if (tr->word_delay.value == 0 && bits_per_word == 16 && tr->len <= 2) {
+			/* multi-mode is applicable, only one "bigger" word (16,32 bits) */
+			tr->bits_per_word = tr->len * bits_per_word / 2;
 		} else if (bits_per_word >= 8 && tr->len == bits_per_word / 8) {
-			/* multi-mode is applicable, only one word (8..32 bits) */
+			/* multi-mode is applicable, only one word (9..15,17..32 bits) */
 		} else {
 			/* multi-mode is not applicable: more than one word in the transfer */
 			mcspi->use_multi_mode = false;

-- 
2.43.0


Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ