linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: "Jonathan Neuschäfer" <j.neuschaefer@gmx.net>
To: linux-spi@vger.kernel.org, openbmc@lists.ozlabs.org
Cc: "Rob Herring" <robh+dt@kernel.org>,
	"Krzysztof Kozlowski" <krzysztof.kozlowski+dt@linaro.org>,
	"Jonathan Neuschäfer" <j.neuschaefer@gmx.net>,
	devicetree@vger.kernel.org, linux-kernel@vger.kernel.org,
	"Mark Brown" <broonie@kernel.org>
Subject: [PATCH v2 3/3] spi: wpcm-fiu: Add direct map support
Date: Thu, 24 Nov 2022 20:14:00 +0100	[thread overview]
Message-ID: <20221124191400.287918-4-j.neuschaefer@gmx.net> (raw)
In-Reply-To: <20221124191400.287918-1-j.neuschaefer@gmx.net>

Besides software controlled SPI transfers (UMA, "user mode access"), FIU
also supports a 16 MiB mapping window per attached flash chip.

This patch implements direct mapped read access, to speed up flash reads.


Without direct mapping:

	# time dd if=/dev/mtd0ro of=dump bs=1M
	16+0 records in
	16+0 records out
	real    1m 47.74s
	user    0m 0.00s
	sys     1m 47.75s


With direct mapping:

	# time dd if=/dev/mtd0ro of=dump bs=1M
	16+0 records in
	16+0 records out
	real    0m 30.81s
	user    0m 0.00s
	sys     0m 30.81s

Signed-off-by: Jonathan Neuschäfer <j.neuschaefer@gmx.net>
---

v2:
- Fix a few nits from the kernel test robot
- Fix check for fiu->memory mapping error

v1:
- https://lore.kernel.org/lkml/20221105185911.1547847-9-j.neuschaefer@gmx.net/
---
 drivers/spi/spi-wpcm-fiu.c | 64 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 64 insertions(+)

diff --git a/drivers/spi/spi-wpcm-fiu.c b/drivers/spi/spi-wpcm-fiu.c
index e525fe074f883..ab33710d50ac8 100644
--- a/drivers/spi/spi-wpcm-fiu.c
+++ b/drivers/spi/spi-wpcm-fiu.c
@@ -51,10 +51,16 @@
  */
 #define UMA_WAIT_ITERATIONS 100

+/* The memory-mapped view of flash is 16 MiB long */
+#define MAX_MEMORY_SIZE_PER_CS	(16 << 20)
+#define MAX_MEMORY_SIZE_TOTAL	(4 * MAX_MEMORY_SIZE_PER_CS)
+
 struct wpcm_fiu_spi {
 	struct device *dev;
 	struct clk *clk;
 	void __iomem *regs;
+	void __iomem *memory;
+	size_t memory_size;
 	struct regmap *shm_regmap;
 };

@@ -367,14 +373,64 @@ static int wpcm_fiu_adjust_op_size(struct spi_mem *mem, struct spi_mem_op *op)
 	return 0;
 }

+static int wpcm_fiu_dirmap_create(struct spi_mem_dirmap_desc *desc)
+{
+	struct wpcm_fiu_spi *fiu = spi_controller_get_devdata(desc->mem->spi->controller);
+	int cs = desc->mem->spi->chip_select;
+
+	if (desc->info.op_tmpl.data.dir != SPI_MEM_DATA_IN)
+		return -ENOTSUPP;
+
+	/*
+	 * Unfortunately, FIU only supports a 16 MiB direct mapping window (per
+	 * attached flash chip), but the SPI MEM core doesn't support partial
+	 * direct mappings. This means that we can't support direct mapping on
+	 * flashes that are bigger than 16 MiB.
+	 */
+	if (desc->info.offset + desc->info.length > MAX_MEMORY_SIZE_PER_CS)
+		return -ENOTSUPP;
+
+	/* Don't read past the memory window */
+	if (cs * MAX_MEMORY_SIZE_PER_CS + desc->info.offset + desc->info.length > fiu->memory_size)
+		return -ENOTSUPP;
+
+	return 0;
+}
+
+static ssize_t wpcm_fiu_direct_read(struct spi_mem_dirmap_desc *desc, u64 offs, size_t len, void *buf)
+{
+	struct wpcm_fiu_spi *fiu = spi_controller_get_devdata(desc->mem->spi->controller);
+	int cs = desc->mem->spi->chip_select;
+
+	if (offs >= MAX_MEMORY_SIZE_PER_CS)
+		return -ENOTSUPP;
+
+	offs += cs * MAX_MEMORY_SIZE_PER_CS;
+
+	if (!fiu->memory || offs >= fiu->memory_size)
+		return -ENOTSUPP;
+
+	len = min_t(size_t, len, fiu->memory_size - offs);
+	memcpy_fromio(buf, fiu->memory + offs, len);
+
+	return len;
+}
+
 static const struct spi_controller_mem_ops wpcm_fiu_mem_ops = {
 	.adjust_op_size = wpcm_fiu_adjust_op_size,
 	.supports_op = wpcm_fiu_supports_op,
 	.exec_op = wpcm_fiu_exec_op,
+	.dirmap_create = wpcm_fiu_dirmap_create,
+	.dirmap_read = wpcm_fiu_direct_read,
 };

 static void wpcm_fiu_hw_init(struct wpcm_fiu_spi *fiu)
 {
+	/* Configure memory-mapped flash access */
+	writeb(FIU_BURST_CFG_R16, fiu->regs + FIU_BURST_BFG);
+	writeb(MAX_MEMORY_SIZE_TOTAL / (512 << 10), fiu->regs + FIU_CFG);
+	writeb(MAX_MEMORY_SIZE_PER_CS / (512 << 10) | BIT(6), fiu->regs + FIU_SPI_FL_CFG);
+
 	/* Deassert all manually asserted chip selects */
 	writeb(0x0f, fiu->regs + FIU_UMA_ECTS);
 }
@@ -404,6 +460,14 @@ static int wpcm_fiu_probe(struct platform_device *pdev)
 	if (IS_ERR(fiu->clk))
 		return PTR_ERR(fiu->clk);

+	res = platform_get_resource_byname(pdev, IORESOURCE_MEM, "memory");
+	fiu->memory = devm_ioremap_resource(dev, res);
+	fiu->memory_size = min_t(size_t, resource_size(res), MAX_MEMORY_SIZE_TOTAL);
+	if (IS_ERR(fiu->memory)) {
+		dev_err(dev, "Failed to map flash memory window\n");
+		return PTR_ERR(fiu->memory);
+	}
+
 	fiu->shm_regmap = syscon_regmap_lookup_by_phandle_optional(dev->of_node, "nuvoton,shm");

 	wpcm_fiu_hw_init(fiu);
--
2.35.1


      parent reply	other threads:[~2022-11-24 19:14 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-11-24 19:13 [PATCH v2 0/3] Nuvoton WPCM450 FIU SPI flash controller Jonathan Neuschäfer
2022-11-24 19:13 ` [PATCH v2 1/3] dt-bindings: spi: Add Nuvoton WPCM450 Flash Interface Unit (FIU) Jonathan Neuschäfer
2022-11-25  8:33   ` Krzysztof Kozlowski
2022-11-25 13:13   ` Mark Brown
2022-11-25 14:33     ` Jonathan Neuschäfer
2022-11-26 22:25   ` Rob Herring
2022-11-28 11:05     ` Conor Dooley
2022-11-28 13:58       ` Jonathan Neuschäfer
2022-11-28 14:09         ` Conor Dooley
2022-11-28 17:57           ` Mark Brown
2022-11-24 19:13 ` [PATCH v2 2/3] spi: wpcm-fiu: Add driver for " Jonathan Neuschäfer
2022-11-25 13:04   ` Mark Brown
2022-11-25 16:33     ` Jonathan Neuschäfer
2022-11-25 16:48       ` Mark Brown
2022-11-24 19:14 ` Jonathan Neuschäfer [this message]

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=20221124191400.287918-4-j.neuschaefer@gmx.net \
    --to=j.neuschaefer@gmx.net \
    --cc=broonie@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=krzysztof.kozlowski+dt@linaro.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-spi@vger.kernel.org \
    --cc=openbmc@lists.ozlabs.org \
    --cc=robh+dt@kernel.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 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).