On 09/11/17 12:14, Borislav Petkov wrote:
On Tue, Nov 07, 2017 at 10:38:58PM +0200, Tero Kristo wrote:
TI Keystone and DRA7xx SoCs have support for EDAC on DDR3 memory that can
correct one bit errors and detect two bit errors. Add EDAC driver for this
feature which plugs into the generic kernel EDAC framework.
Signed-off-by: Tero Kristo <t-kristo@xxxxxx>
Cc: Santosh Shilimkar <ssantosh@xxxxxxxxxx>
Cc: Tony Lindgren <tony@xxxxxxxxxxx>
Please add yourself and whoever else wants to get CCed on bug reports
for this driver, to MAINTAINERS.
Ok, will do that.
Also,
please integrate scripts/checkpatch.pl into your patch creation
workflow. Some of the warnings/errors *actually* make sense.
Oh wow, I actually do that but forgot to do that for the very latest
version of this patch for some reason. >.<
Sorry about that. Will fix all.
---
drivers/edac/Kconfig | 7 ++
drivers/edac/Makefile | 1 +
drivers/edac/ti_edac.c | 306 +++++++++++++++++++++++++++++++++++++++++++++++++
3 files changed, 314 insertions(+)
create mode 100644 drivers/edac/ti_edac.c
diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig
index 96afb2a..54f0184 100644
--- a/drivers/edac/Kconfig
+++ b/drivers/edac/Kconfig
@@ -457,4 +457,11 @@ config EDAC_XGENE
Support for error detection and correction on the
APM X-Gene family of SOCs.
+config EDAC_TI
+ tristate "Texas Instruments DDR3 ECC handler"
s/handler/Controller/
Ok.
+ depends on ARCH_KEYSTONE || SOC_DRA7XX
+ help
+ Support for error detection and correction on the
+ TI SoCs.
+
endif # EDAC
diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile
index 0fd9ffa..b54912e 100644
--- a/drivers/edac/Makefile
+++ b/drivers/edac/Makefile
@@ -78,3 +78,4 @@ obj-$(CONFIG_EDAC_THUNDERX) += thunderx_edac.o
obj-$(CONFIG_EDAC_ALTERA) += altera_edac.o
obj-$(CONFIG_EDAC_SYNOPSYS) += synopsys_edac.o
obj-$(CONFIG_EDAC_XGENE) += xgene_edac.o
+obj-$(CONFIG_EDAC_TI) += ti_edac.o
diff --git a/drivers/edac/ti_edac.c b/drivers/edac/ti_edac.c
new file mode 100644
index 0000000..54bacf3
--- /dev/null
+++ b/drivers/edac/ti_edac.c
@@ -0,0 +1,306 @@
+/*
+ * Copyright (C) 2017 Texas Instruments Incorporated - http://www.ti.com/
+ *
+ * Texas Instruments DDR3 ECC error correction and detection driver
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms and conditions of the GNU General Public License,
+ * version 2, as published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program. If not, see <http://www.gnu.org/licenses/>.
+ */
+
+#include <linux/init.h>
+#include <linux/edac.h>
+#include <linux/io.h>
+#include <linux/interrupt.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/module.h>
+
+#include "edac_module.h"
+
+/* EMIF controller registers */
+#define EMIF_SDRAM_CONFIG 0x008
+#define EMIF_IRQ_STATUS 0x0ac
+#define EMIF_IRQ_ENABLE_SET 0x0b4
+#define EMIF_ECC_CTRL 0x110
+#define EMIF_1B_ECC_ERR_CNT 0x130
+#define EMIF_1B_ECC_ERR_THRSH 0x134
+#define EMIF_1B_ECC_ERR_ADDR_LOG 0x13c
+#define EMIF_2B_ECC_ERR_ADDR_LOG 0x140
+
+/* Bit definitions for EMIF_SDRAM_CONFIG */
+#define SDRAM_TYPE_SHIFT 29
+#define SDRAM_TYPE_MASK GENMASK(31,29)
ERROR: space required after that ',' (ctx:VxV)
#100: FILE: drivers/edac/ti_edac.c:41:
+#define SDRAM_TYPE_MASK GENMASK(31,29)
^
ditto for the rest
Yea sorry about these, they slipped through.
+#define SDRAM_TYPE_DDR3 (3 << SDRAM_TYPE_SHIFT)
+#define SDRAM_TYPE_DDR2 (2 << SDRAM_TYPE_SHIFT)
+#define SDRAM_NARROW_MODE_MASK GENMASK(15,14)
+#define SDRAM_K2_NARROW_MODE_SHIFT 12
+#define SDRAM_K2_NARROW_MODE_MASK GENMASK(13,12)
+#define SDRAM_ROWSIZE_SHIFT 7
+#define SDRAM_ROWSIZE_MASK GENMASK(9,7)
+#define SDRAM_IBANK_SHIFT 4
+#define SDRAM_IBANK_MASK GENMASK(6,4)
+#define SDRAM_K2_IBANK_SHIFT 5
+#define SDRAM_K2_IBANK_MASK GENMASK(6,5)
+#define SDRAM_K2_EBANK_SHIFT 3
+#define SDRAM_K2_EBANK_MASK BIT(SDRAM_K2_EBANK_SHIFT)
+#define SDRAM_PAGESIZE_SHIFT 0
+#define SDRAM_PAGESIZE_MASK GENMASK(2,0)
+#define SDRAM_K2_PAGESIZE_SHIFT 0
+#define SDRAM_K2_PAGESIZE_MASK GENMASK(1,0)
+
+#define EMIF_1B_ECC_ERR_THRSH_SHIFT 24
+
+/* IRQ bit definitions */
+#define EMIF_1B_ECC_ERR BIT(5)
+#define EMIF_2B_ECC_ERR BIT(4)
+#define EMIF_WR_ECC_ERR BIT(3)
+#define EMIF_SYS_ERR BIT(0)
+/* Bit 31 enables ECC and 28 enables RMW */
+#define ECC_ENABLED (BIT(31) | BIT(28))
...
+static void ti_edac_setup_dimm(struct mem_ctl_info *mci, u32 type)
+{
+ struct dimm_info *dimm;
+ struct ti_edac *edac = mci->pvt_info;
+ int bits;
+ u32 val;
+ u32 memsize;
+
+ dimm = EDAC_DIMM_PTR(mci->layers, mci->dimms, mci->n_layers, 0, 0, 0);
+
+ val = ti_edac_readl(edac, EMIF_SDRAM_CONFIG);
+
+ if (type == EMIF_TYPE_DRA7) {
+ bits = ((val & SDRAM_PAGESIZE_MASK) >>
+ SDRAM_PAGESIZE_SHIFT) + 8;
+ bits += ((val & SDRAM_ROWSIZE_MASK) >>
+ SDRAM_ROWSIZE_SHIFT) + 9;
Bah, let those stick out - it is more readable this way:
bits = ((val & SDRAM_PAGESIZE_MASK) >> SDRAM_PAGESIZE_SHIFT) + 8;
bits += ((val & SDRAM_ROWSIZE_MASK) >> SDRAM_ROWSIZE_SHIFT) + 9;
bits += (val & SDRAM_IBANK_MASK) >> SDRAM_IBANK_SHIFT;
Ok.
+ bits += (val & SDRAM_IBANK_MASK) >> SDRAM_IBANK_SHIFT;
+
+ if (val & SDRAM_NARROW_MODE_MASK) {
+ bits++;
+ dimm->dtype = DEV_X16;
+ } else {
+ bits += 2;
+ dimm->dtype = DEV_X32;
+ }
+ } else {
+ bits = 16;
+ bits += ((val & SDRAM_K2_PAGESIZE_MASK) >>
+ SDRAM_K2_PAGESIZE_SHIFT) + 8;
+ bits += (val & SDRAM_K2_IBANK_MASK) >> SDRAM_K2_IBANK_SHIFT;
+ bits += (val & SDRAM_K2_EBANK_MASK) >> SDRAM_K2_EBANK_SHIFT;
+
+ val = (val & SDRAM_K2_NARROW_MODE_MASK) >>
+ SDRAM_K2_NARROW_MODE_SHIFT;
+ switch (val) {
+ case 0:
+ bits += 3;
+ dimm->dtype = DEV_X64;
+ break;
+ case 1:
+ bits += 2;
+ dimm->dtype = DEV_X32;
+ break;
+ case 2:
+ bits ++;
ERROR: space prohibited before that '++' (ctx:WxO)
#233: FILE: drivers/edac/ti_edac.c:174:
+ bits ++;
^
+ dimm->dtype = DEV_X16;
+ break;
+ }
+ }
+
+ memsize = 1 << bits;
+
+ dimm->nr_pages = memsize >> PAGE_SHIFT;
+ dimm->grain = 4;
+ if ((val & SDRAM_TYPE_MASK) == SDRAM_TYPE_DDR2)
+ dimm->mtype = MEM_DDR2;
+ else
+ dimm->mtype = MEM_DDR3;
+
+ val = ti_edac_readl(edac, EMIF_ECC_CTRL);
+ if (val & ECC_ENABLED)
+ dimm->edac_mode = EDAC_SECDED;
+ else
+ dimm->edac_mode = EDAC_NONE;
So if ECC is not enabled, why do you even need to continue loading the
driver here and not return a retval which ti_edac_probe() propagates?
Also, you want to call that function much earlier in ti_edac_probe() so
that you can save yourself all the setup.
Or can you have instances which have ECC disabled and others which have
ECC enabled, on the same platform?
Yeah, the DRA7 SoCs have 2x emif instances, and ECC is only available on
one (and it can be either enabled or disabled.) I thought it might be
useful to have the dimm info populated for both in this case to avoid
confusion and to be able to see the status from userspace.
It is also possible to ditch that and only have dimm info for instances
which have ECC enabled. Which way would you prefer?
+}
+
+static const struct of_device_id ti_edac_of_match[] = {
+ { .compatible = "ti,emif-keystone", .data = (void *)EMIF_TYPE_K2 },
+ { .compatible = "ti,emif-dra7xx", .data = (void *)EMIF_TYPE_DRA7 },
+ {},
+};
+
+static int ti_edac_probe(struct platform_device *pdev)
+{
+ int error_irq = 0, ret = -ENODEV;
+ struct device *dev = &pdev->dev;
+ struct resource *res;
+ void __iomem *reg;
+ struct mem_ctl_info *mci;
+ struct edac_mc_layer layers[1];
+ const struct of_device_id *id;
+ struct ti_edac *edac;
+ static int edac_id;
+ int my_id;
+
+ id = of_match_device(ti_edac_of_match, &pdev->dev);
+ if (!id)
+ return -ENODEV;
+
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ reg = devm_ioremap_resource(dev, res);
+ if (IS_ERR(reg)) {
+ dev_err(dev, "DDR3 controller regs not defined\n");
We have edac_printk() et al macros for printing. Ditto for the remaining
calls.
Ok, I can switch to use those.
+ return PTR_ERR(reg);
+ }
+
+ layers[0].type = EDAC_MC_LAYER_ALL_MEM;
+ layers[0].size = 1;
+
+ /* Allocate ID number for our EMIF controller */
+ mutex_lock(&ti_edac_lock);
+ my_id = edac_id++;
+ mutex_unlock(&ti_edac_lock);
That looks silly. Why not atomic_inc_return()?
Oh didn't quite realize we had such an API also available. Will switch
to that.
+ mci = edac_mc_alloc(my_id, 1, layers, sizeof(*edac));
+ if (!mci)
+ return -ENOMEM;
+
+ mci->pdev = &pdev->dev;
+ edac = mci->pvt_info;
+ edac->reg = reg;
+ platform_set_drvdata(pdev, mci);
...
Thanks for review, and sorry again for missing the checkpatch.
-Tero
--
Texas Instruments Finland Oy, Porkkalankatu 22, 00180 Helsinki. Y-tunnus/Business ID: 0615521-4. Kotipaikka/Domicile: Helsinki
--
To unsubscribe from this list: send the line "unsubscribe linux-omap" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
