On Mon, Mar 17, 2014 at 10:53:44AM +0530, Punnaiah Choudary Kalluri wrote:
> Added EDAC support for reporting the ecc errors of synopsys ddr controller.
> The ddr ecc controller corrects single bit errors and detects double bit
> errors
>
> Signed-off-by: Punnaiah Choudary Kalluri <[email protected]>
> ---
> Changes for v2:
> - Updated the commit header and message
> - Renamed the filenames to synopsys_edac
> - Corrected the compatilble string, commnets
> - Renamed the macros,fucntions and data structures
> ---
> .../devicetree/bindings/edac/synopsys_edac.txt | 18 +
> drivers/edac/Kconfig | 7 +
> drivers/edac/Makefile | 1 +
> drivers/edac/synopsys_edac.c | 614 ++++++++++++++++++++
> 4 files changed, 640 insertions(+), 0 deletions(-)
> create mode 100644 Documentation/devicetree/bindings/edac/synopsys_edac.txt
> create mode 100644 drivers/edac/synopsys_edac.c
>
> diff --git a/Documentation/devicetree/bindings/edac/synopsys_edac.txt b/Documentation/devicetree/bindings/edac/synopsys_edac.txt
> new file mode 100644
> index 0000000..c4a559b
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/edac/synopsys_edac.txt
> @@ -0,0 +1,18 @@
> +Synopsys EDAC driver, it does reports the DDR ECC single bit errors that are
> +corrected and double bit ecc errors that are detected by the DDR ECC controller.
> +ECC support for DDR is available in half-bus width(16 bit) configuration only.
> +
> +Required properties:
> +- compatible: Should be "xlnx,zynq-ddrc-1.04"
> +- reg: Should contain DDR controller registers location and length.
> +
> +Example:
> +++++++++
> +
> +ddrc0: ddrc@f8006000 {
> + compatible = "xlnx,zynq-ddrc-1.04";
> + reg = <0xf8006000 0x1000>;
> +};
> +
> +Synopsys EDAC driver detects the DDR ECC enable state by reading the appropriate
> +control register.
> diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig
> index 878f090..58b69b1 100644
> --- a/drivers/edac/Kconfig
> +++ b/drivers/edac/Kconfig
> @@ -368,4 +368,11 @@ config EDAC_OCTEON_PCI
> Support for error detection and correction on the
> Cavium Octeon family of SOCs.
>
> +config EDAC_SYNOPSYS
> + tristate "Synopsys DDR Memory Controller"
> + depends on EDAC_MM_EDAC && ARCH_ZYNQ
> + help
> + This enables support for EDAC on the ECC memory used
> + with the Synopsys DDR memory controller.
s/This enables s/S/
> +
> endif # EDAC
> diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile
> index 4154ed6..5628a6f 100644
> --- a/drivers/edac/Makefile
> +++ b/drivers/edac/Makefile
> @@ -64,3 +64,4 @@ obj-$(CONFIG_EDAC_OCTEON_PC) += octeon_edac-pc.o
> obj-$(CONFIG_EDAC_OCTEON_L2C) += octeon_edac-l2c.o
> obj-$(CONFIG_EDAC_OCTEON_LMC) += octeon_edac-lmc.o
> obj-$(CONFIG_EDAC_OCTEON_PCI) += octeon_edac-pci.o
> +obj-$(CONFIG_EDAC_SYNOPSYS) += synopsys_edac.o
> diff --git a/drivers/edac/synopsys_edac.c b/drivers/edac/synopsys_edac.c
> new file mode 100644
> index 0000000..7cec331
> --- /dev/null
> +++ b/drivers/edac/synopsys_edac.c
> @@ -0,0 +1,614 @@
> +/*
> + * Synopsys DDR ECC Driver
> + * This driver is based on ppc4xx_edac.c drivers
> + *
> + * Copyright (C) 2012 - 2014 Xilinx, Inc.
The same question to you: are you going to maintain this driver? If so,
please consider adding yourself to the MAINTAINERS file so that people
reporting issues with it can send you a note.
> + *
> + * This program is free software: you can redistribute it and/or modify
> + * it under the terms of the GNU General Public License as published by
> + * the Free Software Foundation, either version 2 of the License, or
> + * (at your option) any later version.
> + *
> + * This program is distributed in the hope that 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/>.
> + */
Please drop the GPL boilerplate - simply refer to the COPYING file in
the kernel source repository instead.
> +
> +#include <linux/edac.h>
> +#include <linux/module.h>
> +#include <linux/platform_device.h>
> +
> +#include "edac_core.h"
> +
> +/* Number of cs_rows needed per memory controller */
> +#define SYNOPSYS_EDAC_NR_CSROWS 1
> +
> +/* Number of channels per memory controller */
> +#define SYNOPSYS_EDAC_NR_CHANS 1
> +
> +/* Granularity of reported error in bytes */
> +#define SYNOPSYS_EDAC_ERROR_GRAIN 1
> +
> +#define SYNOPSYS_EDAC_MESSAGE_SIZE 256
> +
> +/* Synopsys DDR memory controller registers that are relevant to ECC */
> +#define SYNOPSYS_DDRC_CONTROL_REG_OFFSET 0x0 /* Control regsieter */
> +#define SYNOPSYS_DDRC_T_ZQ_REG_OFFSET 0xA4 /* ZQ register */
> +
> +/* ECC control register */
> +#define SYNOPSYS_DDRC_ECC_CONTROL_REG_OFFSET 0xC4
> +/* ECC log register */
> +#define SYNOPSYS_DDRC_ECC_CE_LOG_REG_OFFSET 0xC8
> +/* ECC address register */
> +#define SYNOPSYS_DDRC_ECC_CE_ADDR_REG_OFFSET 0xCC
> +/* ECC data[31:0] register */
> +#define SYNOPSYS_DDRC_ECC_CE_DATA_31_0_REG_OFFSET 0xD0
> +
> +/* Uncorrectable error info regsisters */
> +#define SYNOPSYS_DDRC_ECC_UE_LOG_REG_OFFSET 0xDC /* ECC log register */
> +#define SYNOPSYS_DDRC_ECC_UE_ADDR_REG_OFFSET 0xE0 /* ECC address register */
> +#define SYNOPSYS_DDRC_ECC_UE_DATA_31_0_REG_OFFSET 0xE4 /* ECC data reg */
> +
> +#define SYNOPSYS_DDRC_ECC_STAT_REG_OFFSET 0xF0 /* ECC stats register */
> +#define SYNOPSYS_DDRC_ECC_SCRUB_REG_OFFSET 0xF4 /* ECC scrub register */
> +
> +/* Control regsiter bitfield definitions */
> +#define SYNOPSYS_DDRC_CTRLREG_BUSWIDTH_MASK 0xC
> +#define SYNOPSYS_DDRC_CTRLREG_BUSWIDTH_SHIFT 2
> +
> +#define SYNOPSYS_DDRCTL_WDTH_16 1
> +#define SYNOPSYS_DDRCTL_WDTH_32 0
> +
> +/* ZQ register bitfield definitions */
> +#define SYNOPSYS_DDRC_T_ZQ_REG_DDRMODE_MASK 0x2
> +
> +/* ECC control register bitfield definitions */
> +#define SYNOPSYS_DDRC_ECCCTRL_CLR_CE_ERR 0x2
> +#define SYNOPSYS_DDRC_ECCCTRL_CLR_UE_ERR 0x1
> +
> +/* ECC correctable/uncorrectable error log register definitions */
> +#define SYNOPSYS_DDRC_ECC_CE_LOGREG_VALID 0x1
> +#define SYNOPSYS_DDRC_ECC_CE_LOGREG_BITPOS_MASK 0xFE
> +#define SYNOPSYS_DDRC_ECC_CE_LOGREG_BITPOS_SHIFT 1
> +
> +/* ECC correctable/uncorrectable error address register definitions */
> +#define SYNOPSYS_DDRC_ECC_ADDRREG_COL_MASK 0xFFF
> +#define SYNOPSYS_DDRC_ECC_ADDRREG_ROW_MASK 0xFFFF000
> +#define SYNOPSYS_DDRC_ECC_ADDRREG_ROW_SHIFT 12
> +#define SYNOPSYS_DDRC_ECC_ADDRREG_BANK_MASK 0x70000000
> +#define SYNOPSYS_DDRC_ECC_ADDRREG_BANK_SHIFT 28
> +
> +/* ECC statistic regsiter definitions */
> +#define SYNOPSYS_DDRC_ECC_STATREG_UECOUNT_MASK 0xFF
> +#define SYNOPSYS_DDRC_ECC_STATREG_CECOUNT_MASK 0xFF00
> +#define SYNOPSYS_DDRC_ECC_STATREG_CECOUNT_SHIFT 8
> +
> +/* ECC scrub regsiter definitions */
> +#define SYNOPSYS_DDRC_ECC_SCRUBREG_ECC_MODE_MASK 0x7
> +#define SYNOPSYS_DDRC_ECC_SCRUBREG_ECCMODE_SECDED 0x4
Those defines and their values could use better alignment. Also see
below the note about slimming down those names in general.
> +/**
> + * struct ecc_error_info - ECC error log information
> + * @row: Row number
> + * @col: Column number
> + * @bank: Bank number
> + * @bitpos: Bit position
> + * @data: Data causing the error
> + */
> +struct ecc_error_info {
> + u32 row;
> + u32 col;
> + u32 bank;
> + u32 bitpos;
> + u32 data;
> +};
> +
> +/**
> + * struct synopsys_ecc_status - ECC status information to report
> + * @ce_count: Correctable error count
> + * @ue_count: Uncorrectable error count
> + * @ceinfo: Correctable error log information
> + * @ueinfo: Uncorrectable error log information
> + */
> +struct synopsys_ecc_status {
> + u32 ce_count;
> + u32 ue_count;
> + struct ecc_error_info ceinfo;
> + struct ecc_error_info ueinfo;
> +};
> +
> +/**
> + * struct synopsys_edac_priv - DDR memory controller private instance data
> + * @baseaddr: Base address of the DDR controller
> + * @ce_count: Correctable Error count
> + * @ue_count: Uncorrectable Error count
> + */
> +struct synopsys_edac_priv {
> + void __iomem *baseaddr;
> + u32 ce_count;
> + u32 ue_count;
> +};
> +
> +/**
Why do we need the kernel-doc annotation for all those static functions?
Also, drop the "synopsys_edac_" prefix of all static functions - that'll
slim up the code even further.
> + * synopsys_edac_geterror_info - Get the current ecc error info
> + * @base: Pointer to the base address of the ddr memory controller
> + * @perrstatus: Pointer to the synopsys ecc status structure
> + *
> + * This routine determines there is any ecc error or not
> + *
> + * Return: zero if there is no error otherwise returns 1
> + */
> +static int synopsys_edac_geterror_info(void __iomem *base,
That base is an arg to readl, so it should be "const volatile void
__iomem *base", no?
> + struct synopsys_ecc_status *perrstatus)
Please align function args at the opening brace "("
> +{
> + u32 regval;
> + u32 clearval = 0;
> +
> + regval = readl(base + SYNOPSYS_DDRC_ECC_STAT_REG_OFFSET) &
> + (SYNOPSYS_DDRC_ECC_STATREG_UECOUNT_MASK |
> + SYNOPSYS_DDRC_ECC_STATREG_CECOUNT_MASK);
All your macro definitions start with "SYNOPSYS_DDRC" even though
they're only local to this file. Which makes the code very unreadable
due to their size.
You could slim them up by dropping the prefix and doing:
regval = readl(base + STAT_REG_OFFSET) &
(STATREG_UECOUNT_MASK | STATREG_CECOUNT_MASK);
which is much better IMO.
The same is true for struct names like synopsys_ecc_status and the like.
Shortening them too would slim down your code considerably.
> +
> + if (regval == 0)
> + return 0;
if (!regval)
> +
> + memset(perrstatus, 0, sizeof(struct synopsys_ecc_status));
That perrstatus could also be shortened to errstat and make the whole
code shorter:
memset(errstat, 0, sizeof(*errstat));
and so on...
> +
> + perrstatus->ce_count =
> + (regval & SYNOPSYS_DDRC_ECC_STATREG_CECOUNT_MASK) >>
> + SYNOPSYS_DDRC_ECC_STATREG_CECOUNT_SHIFT;
> + perrstatus->ue_count =
> + (regval & SYNOPSYS_DDRC_ECC_STATREG_UECOUNT_MASK);
> +
> + if (perrstatus->ce_count) {
> + regval = readl(base + SYNOPSYS_DDRC_ECC_CE_LOG_REG_OFFSET);
> + if (regval & SYNOPSYS_DDRC_ECC_CE_LOGREG_VALID) {
> + perrstatus->ceinfo.bitpos = (regval &
> + SYNOPSYS_DDRC_ECC_CE_LOGREG_BITPOS_MASK) >>
> + SYNOPSYS_DDRC_ECC_CE_LOGREG_BITPOS_SHIFT;
> + regval = readl(base +
> + SYNOPSYS_DDRC_ECC_CE_ADDR_REG_OFFSET);
> + perrstatus->ceinfo.row = (regval &
> + SYNOPSYS_DDRC_ECC_ADDRREG_ROW_MASK) >>
> + SYNOPSYS_DDRC_ECC_ADDRREG_ROW_SHIFT;
> + perrstatus->ceinfo.col = (regval &
> + SYNOPSYS_DDRC_ECC_ADDRREG_COL_MASK);
> + perrstatus->ceinfo.bank = (regval &
> + SYNOPSYS_DDRC_ECC_ADDRREG_BANK_MASK) >>
> + SYNOPSYS_DDRC_ECC_ADDRREG_BANK_SHIFT;
> + perrstatus->ceinfo.data = readl(base +
> + SYNOPSYS_DDRC_ECC_CE_DATA_31_0_REG_OFFSET);
> + edac_dbg(3, "ce bitposition: %d data: %d\n",
> + perrstatus->ceinfo.bitpos,
> + perrstatus->ceinfo.data);
> + }
> + clearval = SYNOPSYS_DDRC_ECCCTRL_CLR_CE_ERR;
> + }
> +
> + if (perrstatus->ue_count) {
> + regval = readl(base + SYNOPSYS_DDRC_ECC_UE_LOG_REG_OFFSET);
> + if (regval & SYNOPSYS_DDRC_ECC_CE_LOGREG_VALID) {
> + regval = readl(base +
> + SYNOPSYS_DDRC_ECC_UE_ADDR_REG_OFFSET);
> + perrstatus->ueinfo.row = (regval &
> + SYNOPSYS_DDRC_ECC_ADDRREG_ROW_MASK) >>
> + SYNOPSYS_DDRC_ECC_ADDRREG_ROW_SHIFT;
> + perrstatus->ueinfo.col = (regval &
> + SYNOPSYS_DDRC_ECC_ADDRREG_COL_MASK);
> + perrstatus->ueinfo.bank = (regval &
> + SYNOPSYS_DDRC_ECC_ADDRREG_BANK_MASK) >>
> + SYNOPSYS_DDRC_ECC_ADDRREG_BANK_SHIFT;
> + perrstatus->ueinfo.data = readl(base +
> + SYNOPSYS_DDRC_ECC_UE_DATA_31_0_REG_OFFSET);
All those assignments could be much shorter:
p->ueinfo.row = (regval & ADDRREG_ROW_MASK) >> ADDRREG_ROW_SHIFT;
Also, you can save yourself two indentation levels by flipping the
logic:
if (!(p->ce_count && (regval & CE_LOGREG_VALID)))
goto ue_error;
/* CE assignments */
...
ue_error:
if (!(p->ue_count && (regval & UE_LOGREG_VALID)))
goto out;
/* UE assignments */
...
out:
writel(...);
writel(...);
> + }
> + clearval |= SYNOPSYS_DDRC_ECCCTRL_CLR_UE_ERR;
> + }
> +
> + writel(clearval, base + SYNOPSYS_DDRC_ECC_CONTROL_REG_OFFSET);
> + writel(0x0, base + SYNOPSYS_DDRC_ECC_CONTROL_REG_OFFSET);
> +
> + return 1;
> +}
> +
> +/**
> + * synopsys_edac_generate_message - Generate interpreted ECC status message
> + * @mci: Pointer to the edac memory controller instance
> + * @perrstatus: Pointer to the synopsys ecc status structure
> + * @buffer: Pointer to the buffer in which to generate the
> + * message
> + * @size: The size, in bytes, of space available in buffer
> + *
> + * This routine generates to the provided buffer the portion of the
> + * driver-unique report message associated with the ECC register of
> + * the specified ECC status.
> + */
> +static void synopsys_edac_generate_message(const struct mem_ctl_info *mci,
> + struct synopsys_ecc_status *perrstatus, char *buffer,
> + size_t size)
> +{
> + struct ecc_error_info *pinfo = NULL;
> +
> + if (perrstatus->ce_count > 0)
> + pinfo = &perrstatus->ceinfo;
> + else
> + pinfo = &perrstatus->ueinfo;
> +
> + snprintf(buffer, SYNOPSYS_EDAC_MESSAGE_SIZE,
> + "DDR ECC error type :%s Row %d Bank %d Col %d ",
> + (perrstatus->ce_count > 0) ? "CE" : "UE", pinfo->row,
> + pinfo->bank, pinfo->col);
> +}
This function can be part of synopsys_edac_handle_error() below instead
as it is simple enough and called only once there - no need for a
separate function.
> +
> +/**
> + * synopsys_edac_handle_error - Handle controller error types CE and UE
> + * @mci: Pointer to the edac memory controller instance
> + * @perrstatus: Pointer to the synopsys ecc status structure
> + *
> + * This routine handles the controller ECC correctable error.
It handles also UC errors.
> + */
> +static void synopsys_edac_handle_error(struct mem_ctl_info *mci,
> + struct synopsys_ecc_status *perrstatus)
> +{
> + char message[SYNOPSYS_EDAC_MESSAGE_SIZE];
> +
> + synopsys_edac_generate_message(mci, perrstatus, &message[0],
> + SYNOPSYS_EDAC_MESSAGE_SIZE);
> +
> + if (perrstatus->ce_count)
> + edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci,
> + perrstatus->ce_count, 0, 0, 0, 0, 0, -1,
> + &message[0], "");
string_array == &string_array[0]
> + else
> + edac_mc_handle_error(HW_EVENT_ERR_UNCORRECTED, mci,
> + perrstatus->ue_count, 0, 0, 0, 0, 0, -1,
> + &message[0], "");
Ditto.
> +}
> +
> +/**
> + * synopsys_edac_check - Check controller for ECC errors
> + * @mci: Pointer to the edac memory controller instance
> + *
> + * This routine is used to check and post ECC errors and is called by
> + * the EDAC polling thread
> + */
> +static void synopsys_edac_check(struct mem_ctl_info *mci)
> +{
> + struct synopsys_edac_priv *priv = mci->pvt_info;
> + struct synopsys_ecc_status errstatus;
> + int status;
> +
> + status = synopsys_edac_geterror_info(priv->baseaddr, &errstatus);
> + if (status) {
Save an indentation level:
if (!status)
return;
priv->ce_count += ...
...
> + priv->ce_count += errstatus.ce_count;
> + priv->ue_count += errstatus.ue_count;
> +
> + if (errstatus.ce_count) {
> + synopsys_edac_handle_error(mci, &errstatus);
> + errstatus.ce_count = 0;
> + }
> + if (errstatus.ue_count) {
> + synopsys_edac_handle_error(mci, &errstatus);
> + errstatus.ue_count = 0;
> + }
> + edac_dbg(3, "total error count ce %d ue %d\n",
> + priv->ce_count, priv->ue_count);
> + }
> +}
> +
> +/**
> + * synopsys_edac_get_dtype - Return the controller memory width
> + * @base: Pointer to the ddr memory contoller base address
> + *
> + * This routine returns the EDAC device type width appropriate for the
> + * current controller configuration.
> + *
> + * Return: a device type width enumeration.
> + */
> +static enum dev_type synopsys_edac_get_dtype(void __iomem *base)
> +{
> + enum dev_type dt;
> + u32 width;
> +
> + width = readl(base + SYNOPSYS_DDRC_CONTROL_REG_OFFSET);
> + width = (width & SYNOPSYS_DDRC_CTRLREG_BUSWIDTH_MASK) >>
> + SYNOPSYS_DDRC_CTRLREG_BUSWIDTH_SHIFT;
> +
> + switch (width) {
> + case SYNOPSYS_DDRCTL_WDTH_16:
> + dt = DEV_X2;
> + break;
> + case SYNOPSYS_DDRCTL_WDTH_32:
> + dt = DEV_X4;
> + break;
> + default:
> + dt = DEV_UNKNOWN;
> + }
> +
> + return dt;
> +}
> +
> +/**
> + * synopsys_edac_get_eccstate - Return the controller ecc enable/disable status
> + * @base: Pointer to the ddr memory contoller base address
> + *
> + * This routine returns the ECC enable/diable status for the controller
> + *
> + * Return: a ecc status boolean i.e true/false - enabled/disabled.
> + */
> +static bool synopsys_edac_get_eccstate(void __iomem *base)
> +{
> + enum dev_type dt;
> + u32 ecctype;
> + bool state = false;
> +
> + dt = synopsys_edac_get_dtype(base);
Shouldn't you handle the error case where it returns DEV_UNKNOWN type?
> +
> + ecctype = (readl(base + SYNOPSYS_DDRC_ECC_SCRUB_REG_OFFSET) &
> + SYNOPSYS_DDRC_ECC_SCRUBREG_ECC_MODE_MASK);
> +
> + if ((ecctype == SYNOPSYS_DDRC_ECC_SCRUBREG_ECCMODE_SECDED)
> + && (dt == DEV_X2)) {
> + state = true;
> + writel(0x0, base + SYNOPSYS_DDRC_ECC_CONTROL_REG_OFFSET);
> + } else {
> + state = false;
Why that else branch? state is already initialized to false at function
entry time.
> + }
> +
> + return state;
> +}
> +
> +/**
> + * synopsys_edac_get_memsize - reads the size of the attached memory device
> + *
> + * Return: the memory size in bytes
> + *
> + * This routine returns the size of the system memory by reading the sysinfo
> + * information
> + */
No need for that comment.
> +static u32 synopsys_edac_get_memsize(void)
> +{
> + struct sysinfo inf;
> +
> + /* Reading the system memory size from the global meminfo structure */
Same here.
For simple functions where it is paramount what the function does, you
don't need to comment every line.
> + si_meminfo(&inf);
> +
> + return inf.totalram * inf.mem_unit;
> +}
> +
> +/**
> + * synopsys_edac_get_mtype - Returns controller memory type
> + * @base: pointer to the synopsys ecc status structure
> + *
> + * This routine returns the EDAC memory type appropriate for the
> + * current controller configuration.
> + *
> + * Return: a memory type enumeration.
> + */
> +static enum mem_type synopsys_edac_get_mtype(void __iomem *base)
> +{
> + enum mem_type mt;
> + u32 memtype;
> +
> + memtype = readl(base + SYNOPSYS_DDRC_T_ZQ_REG_OFFSET);
> +
> + if (memtype & SYNOPSYS_DDRC_T_ZQ_REG_DDRMODE_MASK)
> + mt = MEM_DDR3;
> + else
> + mt = MEM_DDR2;
> +
> + return mt;
> +}
> +
> +/**
> + * synopsys_edac_init_csrows - Initialize the cs row data
> + * @mci: Pointer to the edac memory controller instance
> + *
> + * This routine initializes the chip select rows associated
> + * with the EDAC memory controller instance
> + *
> + * Return: 0 if OK; otherwise, -EINVAL if the memory bank size
Where is that -EINVAL being returned?
> + * configuration cannot be determined.
> + */
> +static int synopsys_edac_init_csrows(struct mem_ctl_info *mci)
> +{
> + struct csrow_info *csi;
> + struct dimm_info *dimm;
> + struct synopsys_edac_priv *priv = mci->pvt_info;
> + u32 size;
> + int row, j;
> +
> + for (row = 0; row < mci->nr_csrows; row++) {
> + csi = mci->csrows[row];
> + size = synopsys_edac_get_memsize();
> +
> + for (j = 0; j < csi->nr_channels; j++) {
> + dimm = csi->channels[j]->dimm;
> + dimm->edac_mode = EDAC_FLAG_SECDED;
> + dimm->mtype = synopsys_edac_get_mtype(priv->baseaddr);
> + dimm->nr_pages =
> + (size >> PAGE_SHIFT) / csi->nr_channels;
> + dimm->grain = SYNOPSYS_EDAC_ERROR_GRAIN;
> + dimm->dtype = synopsys_edac_get_dtype(priv->baseaddr);
> + }
> +
> + }
> +
> + return 0;
> +}
> +
> +/**
> + * synopsys_edac_mc_init - Initialize driver instance
> + * @mci: Pointer to the edac memory controller instance
> + * @pdev: Pointer to the platform_device struct
> + *
> + * This routine performs initialization of the EDAC memory controller
> + * instance and related driver-private data associated with the
> + * memory controller the instance is bound to.
> + *
> + * Return: 0 if OK; otherwise, < 0 on error.
> + */
> +static int synopsys_edac_mc_init(struct mem_ctl_info *mci,
> + struct platform_device *pdev)
> +{
> + int status;
> + struct synopsys_edac_priv *priv;
> +
> + /* Initial driver pointers and private data */
Useless comment.
> + mci->pdev = &pdev->dev;
> + priv = mci->pvt_info;
> + platform_set_drvdata(pdev, mci);
> +
> + /* Initialize controller capabilities and configuration */
> + mci->mtype_cap = MEM_FLAG_DDR3 | MEM_FLAG_DDR2;
> + mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_SECDED;
> + mci->scrub_cap = SCRUB_HW_SRC;
> + /* Check the scrub setting from the controller */
Useless comment.
> + mci->scrub_mode = SCRUB_NONE;
> +
> + mci->edac_cap = EDAC_FLAG_SECDED;
> + /* Initialize strings */
Useless comment.
> + mci->ctl_name = "synopsys_ddr_controller";
> + mci->dev_name = dev_name(&pdev->dev);
> + mci->mod_name = "synopsys_edac";
Do a
#define EDAC_MOD_STR "synopsys_edac"
like the other drivers do and use that here and below.
> + mci->mod_ver = "1";
> +
> + /* Initialize callbacks */
Useless comment.
> + edac_op_state = EDAC_OPSTATE_POLL;
> + mci->edac_check = synopsys_edac_check;
> + mci->ctl_page_to_phys = NULL;
> +
> + /*
> + * Initialize the MC control structure 'csrows' table
> + * with the mapping and control information.
> + */
> + status = synopsys_edac_init_csrows(mci);
> + if (status)
> + pr_err("Failed to initialize rows!\n");
edac_printk
> +
> + return status;
> +}
> +
> +/**
> + * synopsys_edac_mc_probe - Check controller and bind driver
> + * @pdev: Pointer to the platform_device struct
> + *
> + * This routine probes a specific controller
> + * instance for binding with the driver.
> + *
> + * Return: 0 if the controller instance was successfully bound to the
> + * driver; otherwise, < 0 on error.
> + */
> +static int synopsys_edac_mc_probe(struct platform_device *pdev)
> +{
> + struct mem_ctl_info *mci;
> + struct edac_mc_layer layers[2];
> + struct synopsys_edac_priv *priv;
> + int rc;
> + struct resource *res;
> + void __iomem *baseaddr;
> +
> + /* Get the data from the platform device */
> + res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
You're not handling the case where this function returns NULL.
> + baseaddr = devm_ioremap_resource(&pdev->dev, res);
> + if (IS_ERR(baseaddr))
> + return PTR_ERR(baseaddr);
> +
> + /* Check for the ecc enable status */
> + if (synopsys_edac_get_eccstate(baseaddr) == false) {
if (!get_eccstate(...))
> + dev_err(&pdev->dev, "ecc not enabled\n");
edac_printk pls.
> + return -ENXIO;
> + }
> +
> + /*
> + * At this point, we know ECC is enabled, allocate an EDAC
> + * controller instance and perform the appropriate
> + * initialization.
> + */
> + layers[0].type = EDAC_MC_LAYER_CHIP_SELECT;
> + layers[0].size = SYNOPSYS_EDAC_NR_CSROWS;
> + layers[0].is_virt_csrow = true;
> + layers[1].type = EDAC_MC_LAYER_CHANNEL;
> + layers[1].size = SYNOPSYS_EDAC_NR_CHANS;
> + layers[1].is_virt_csrow = false;
> +
> + mci = edac_mc_alloc(0, ARRAY_SIZE(layers), layers,
> + sizeof(struct synopsys_edac_priv));
> + if (mci == NULL) {
if (!mci)
> + pr_err("Failed memory allocation for mci instance!\n");
edac_printk
Before you return, devm_iounmap(... baseaddr). Just do it with nice goto
labels.
> + return -ENOMEM;
> + }
> +
> + priv = mci->pvt_info;
> + priv->baseaddr = baseaddr;
> + rc = synopsys_edac_mc_init(mci, pdev);
> + if (rc) {
> + pr_err("Failed to initialize instance!\n");
edac_printk
> + goto free_edac_mc;
> + }
> +
> + /*
> + * We have a valid, initialized EDAC instance bound to the
> + * controller. Attempt to register it with the EDAC subsystem
> + */
> + rc = edac_mc_add_mc(mci);
> + if (rc) {
> + dev_err(&pdev->dev, "failed to register with EDAC core\n");
Ditto.
> + goto del_edac_mc;
> + }
> +
> + return rc;
> +
> +del_edac_mc:
> + edac_mc_del_mc(&pdev->dev);
> +free_edac_mc:
> + edac_mc_free(mci);
> +
> + return rc;
> +}
> +
> +/**
> + * synopsys_edac_mc_remove - Unbind driver from controller
> + * @pdev: Pointer to the platform_device struct
> + *
> + * This routine unbinds the EDAC memory controller instance associated
> + * with the specified controller described by the
> + * OpenFirmware device tree node passed as a parameter.
> + *
> + * Return: Unconditionally 0
> + */
No need for that whole comment at all.
> +static int synopsys_edac_mc_remove(struct platform_device *pdev)
> +{
> + struct mem_ctl_info *mci = platform_get_drvdata(pdev);
> +
> + edac_mc_del_mc(&pdev->dev);
> + edac_mc_free(mci);
> +
> + return 0;
> +}
> +
> +/* Device tree node type and compatible tuples this driver can match on */
Ditto.
> +static struct of_device_id synopsys_edac_match[] = {
> + { .compatible = "xlnx,zynq-ddrc-1.04", },
> + { /* end of table */ }
> +};
> +
> +MODULE_DEVICE_TABLE(of, synopsys_edac_match);
> +
> +static struct platform_driver synopsys_edac_mc_driver = {
> + .driver = {
> + .name = "synopsys-edac",
> + .owner = THIS_MODULE,
> + .of_match_table = synopsys_edac_match,
> + },
> + .probe = synopsys_edac_mc_probe,
> + .remove = synopsys_edac_mc_remove,
> +};
> +
> +module_platform_driver(synopsys_edac_mc_driver);
> +
> +MODULE_AUTHOR("Xilinx, Inc.");
> +MODULE_DESCRIPTION("Synopsys DDR ECC driver");
> +MODULE_LICENSE("GPL v2");
Ok, that should be it so far, more review with the next submission.
--
Regards/Gruss,
Boris.
Sent from a fat crate under my desk. Formatting is fine.
--
On Wed, Apr 9, 2014 at 2:10 AM, Borislav Petkov <[email protected]> wrote:
> On Mon, Mar 17, 2014 at 10:53:44AM +0530, Punnaiah Choudary Kalluri wrote:
>> Added EDAC support for reporting the ecc errors of synopsys ddr controller.
>> The ddr ecc controller corrects single bit errors and detects double bit
>> errors
>>
>> Signed-off-by: Punnaiah Choudary Kalluri <[email protected]>
>> ---
>> Changes for v2:
>> - Updated the commit header and message
>> - Renamed the filenames to synopsys_edac
>> - Corrected the compatilble string, commnets
>> - Renamed the macros,fucntions and data structures
>> ---
>> .../devicetree/bindings/edac/synopsys_edac.txt | 18 +
>> drivers/edac/Kconfig | 7 +
>> drivers/edac/Makefile | 1 +
>> drivers/edac/synopsys_edac.c | 614 ++++++++++++++++++++
>> 4 files changed, 640 insertions(+), 0 deletions(-)
>> create mode 100644 Documentation/devicetree/bindings/edac/synopsys_edac.txt
>> create mode 100644 drivers/edac/synopsys_edac.c
>>
>> diff --git a/Documentation/devicetree/bindings/edac/synopsys_edac.txt b/Documentation/devicetree/bindings/edac/synopsys_edac.txt
>> new file mode 100644
>> index 0000000..c4a559b
>> --- /dev/null
>> +++ b/Documentation/devicetree/bindings/edac/synopsys_edac.txt
>> @@ -0,0 +1,18 @@
>> +Synopsys EDAC driver, it does reports the DDR ECC single bit errors that are
>> +corrected and double bit ecc errors that are detected by the DDR ECC controller.
>> +ECC support for DDR is available in half-bus width(16 bit) configuration only.
>> +
>> +Required properties:
>> +- compatible: Should be "xlnx,zynq-ddrc-1.04"
>> +- reg: Should contain DDR controller registers location and length.
>> +
>> +Example:
>> +++++++++
>> +
>> +ddrc0: ddrc@f8006000 {
>> + compatible = "xlnx,zynq-ddrc-1.04";
>> + reg = <0xf8006000 0x1000>;
>> +};
>> +
>> +Synopsys EDAC driver detects the DDR ECC enable state by reading the appropriate
>> +control register.
>> diff --git a/drivers/edac/Kconfig b/drivers/edac/Kconfig
>> index 878f090..58b69b1 100644
>> --- a/drivers/edac/Kconfig
>> +++ b/drivers/edac/Kconfig
>> @@ -368,4 +368,11 @@ config EDAC_OCTEON_PCI
>> Support for error detection and correction on the
>> Cavium Octeon family of SOCs.
>>
>> +config EDAC_SYNOPSYS
>> + tristate "Synopsys DDR Memory Controller"
>> + depends on EDAC_MM_EDAC && ARCH_ZYNQ
>> + help
>> + This enables support for EDAC on the ECC memory used
>> + with the Synopsys DDR memory controller.
>
> s/This enables s/S/
>
>> +
>> endif # EDAC
>> diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile
>> index 4154ed6..5628a6f 100644
>> --- a/drivers/edac/Makefile
>> +++ b/drivers/edac/Makefile
>> @@ -64,3 +64,4 @@ obj-$(CONFIG_EDAC_OCTEON_PC) += octeon_edac-pc.o
>> obj-$(CONFIG_EDAC_OCTEON_L2C) += octeon_edac-l2c.o
>> obj-$(CONFIG_EDAC_OCTEON_LMC) += octeon_edac-lmc.o
>> obj-$(CONFIG_EDAC_OCTEON_PCI) += octeon_edac-pci.o
>> +obj-$(CONFIG_EDAC_SYNOPSYS) += synopsys_edac.o
>> diff --git a/drivers/edac/synopsys_edac.c b/drivers/edac/synopsys_edac.c
>> new file mode 100644
>> index 0000000..7cec331
>> --- /dev/null
>> +++ b/drivers/edac/synopsys_edac.c
>> @@ -0,0 +1,614 @@
>> +/*
>> + * Synopsys DDR ECC Driver
>> + * This driver is based on ppc4xx_edac.c drivers
>> + *
>> + * Copyright (C) 2012 - 2014 Xilinx, Inc.
>
> The same question to you: are you going to maintain this driver? If so,
> please consider adding yourself to the MAINTAINERS file so that people
> reporting issues with it can send you a note.
Let me check internally with team and get back to you on this.
>
>> + *
>> + * This program is free software: you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License as published by
>> + * the Free Software Foundation, either version 2 of the License, or
>> + * (at your option) any later version.
>> + *
>> + * This program is distributed in the hope that 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/>.
>> + */
>
> Please drop the GPL boilerplate - simply refer to the COPYING file in
> the kernel source repository instead.
>
>> +
>> +#include <linux/edac.h>
>> +#include <linux/module.h>
>> +#include <linux/platform_device.h>
>> +
>> +#include "edac_core.h"
>> +
>> +/* Number of cs_rows needed per memory controller */
>> +#define SYNOPSYS_EDAC_NR_CSROWS 1
>> +
>> +/* Number of channels per memory controller */
>> +#define SYNOPSYS_EDAC_NR_CHANS 1
>> +
>> +/* Granularity of reported error in bytes */
>> +#define SYNOPSYS_EDAC_ERROR_GRAIN 1
>> +
>> +#define SYNOPSYS_EDAC_MESSAGE_SIZE 256
>> +
>> +/* Synopsys DDR memory controller registers that are relevant to ECC */
>> +#define SYNOPSYS_DDRC_CONTROL_REG_OFFSET 0x0 /* Control regsieter */
>> +#define SYNOPSYS_DDRC_T_ZQ_REG_OFFSET 0xA4 /* ZQ register */
>> +
>> +/* ECC control register */
>> +#define SYNOPSYS_DDRC_ECC_CONTROL_REG_OFFSET 0xC4
>> +/* ECC log register */
>> +#define SYNOPSYS_DDRC_ECC_CE_LOG_REG_OFFSET 0xC8
>> +/* ECC address register */
>> +#define SYNOPSYS_DDRC_ECC_CE_ADDR_REG_OFFSET 0xCC
>> +/* ECC data[31:0] register */
>> +#define SYNOPSYS_DDRC_ECC_CE_DATA_31_0_REG_OFFSET 0xD0
>> +
>> +/* Uncorrectable error info regsisters */
>> +#define SYNOPSYS_DDRC_ECC_UE_LOG_REG_OFFSET 0xDC /* ECC log register */
>> +#define SYNOPSYS_DDRC_ECC_UE_ADDR_REG_OFFSET 0xE0 /* ECC address register */
>> +#define SYNOPSYS_DDRC_ECC_UE_DATA_31_0_REG_OFFSET 0xE4 /* ECC data reg */
>> +
>> +#define SYNOPSYS_DDRC_ECC_STAT_REG_OFFSET 0xF0 /* ECC stats register */
>> +#define SYNOPSYS_DDRC_ECC_SCRUB_REG_OFFSET 0xF4 /* ECC scrub register */
>> +
>> +/* Control regsiter bitfield definitions */
>> +#define SYNOPSYS_DDRC_CTRLREG_BUSWIDTH_MASK 0xC
>> +#define SYNOPSYS_DDRC_CTRLREG_BUSWIDTH_SHIFT 2
>> +
>> +#define SYNOPSYS_DDRCTL_WDTH_16 1
>> +#define SYNOPSYS_DDRCTL_WDTH_32 0
>> +
>> +/* ZQ register bitfield definitions */
>> +#define SYNOPSYS_DDRC_T_ZQ_REG_DDRMODE_MASK 0x2
>> +
>> +/* ECC control register bitfield definitions */
>> +#define SYNOPSYS_DDRC_ECCCTRL_CLR_CE_ERR 0x2
>> +#define SYNOPSYS_DDRC_ECCCTRL_CLR_UE_ERR 0x1
>> +
>> +/* ECC correctable/uncorrectable error log register definitions */
>> +#define SYNOPSYS_DDRC_ECC_CE_LOGREG_VALID 0x1
>> +#define SYNOPSYS_DDRC_ECC_CE_LOGREG_BITPOS_MASK 0xFE
>> +#define SYNOPSYS_DDRC_ECC_CE_LOGREG_BITPOS_SHIFT 1
>> +
>> +/* ECC correctable/uncorrectable error address register definitions */
>> +#define SYNOPSYS_DDRC_ECC_ADDRREG_COL_MASK 0xFFF
>> +#define SYNOPSYS_DDRC_ECC_ADDRREG_ROW_MASK 0xFFFF000
>> +#define SYNOPSYS_DDRC_ECC_ADDRREG_ROW_SHIFT 12
>> +#define SYNOPSYS_DDRC_ECC_ADDRREG_BANK_MASK 0x70000000
>> +#define SYNOPSYS_DDRC_ECC_ADDRREG_BANK_SHIFT 28
>> +
>> +/* ECC statistic regsiter definitions */
>> +#define SYNOPSYS_DDRC_ECC_STATREG_UECOUNT_MASK 0xFF
>> +#define SYNOPSYS_DDRC_ECC_STATREG_CECOUNT_MASK 0xFF00
>> +#define SYNOPSYS_DDRC_ECC_STATREG_CECOUNT_SHIFT 8
>> +
>> +/* ECC scrub regsiter definitions */
>> +#define SYNOPSYS_DDRC_ECC_SCRUBREG_ECC_MODE_MASK 0x7
>> +#define SYNOPSYS_DDRC_ECC_SCRUBREG_ECCMODE_SECDED 0x4
>
> Those defines and their values could use better alignment. Also see
> below the note about slimming down those names in general.
>
>> +/**
>> + * struct ecc_error_info - ECC error log information
>> + * @row: Row number
>> + * @col: Column number
>> + * @bank: Bank number
>> + * @bitpos: Bit position
>> + * @data: Data causing the error
>> + */
>> +struct ecc_error_info {
>> + u32 row;
>> + u32 col;
>> + u32 bank;
>> + u32 bitpos;
>> + u32 data;
>> +};
>> +
>> +/**
>> + * struct synopsys_ecc_status - ECC status information to report
>> + * @ce_count: Correctable error count
>> + * @ue_count: Uncorrectable error count
>> + * @ceinfo: Correctable error log information
>> + * @ueinfo: Uncorrectable error log information
>> + */
>> +struct synopsys_ecc_status {
>> + u32 ce_count;
>> + u32 ue_count;
>> + struct ecc_error_info ceinfo;
>> + struct ecc_error_info ueinfo;
>> +};
>> +
>> +/**
>> + * struct synopsys_edac_priv - DDR memory controller private instance data
>> + * @baseaddr: Base address of the DDR controller
>> + * @ce_count: Correctable Error count
>> + * @ue_count: Uncorrectable Error count
>> + */
>> +struct synopsys_edac_priv {
>> + void __iomem *baseaddr;
>> + u32 ce_count;
>> + u32 ue_count;
>> +};
>> +
>> +/**
>
> Why do we need the kernel-doc annotation for all those static functions?
Since it is recommended in Documentation/kernel-doc-nano-HOWTO.txt
but also said it is low priority and at the discretion of the MAINTAINER of
that kernel source file
So, if you recommend not use kernel-doc annotation then i will take care
in next version.
>
> Also, drop the "synopsys_edac_" prefix of all static functions - that'll
> slim up the code even further.
Ok. it means using geterr_info is sufficient than synopsys_edac_geterror_info
and applicable to other static functions
>
>> + * synopsys_edac_geterror_info - Get the current ecc error info
>> + * @base: Pointer to the base address of the ddr memory controller
>> + * @perrstatus: Pointer to the synopsys ecc status structure
>> + *
>> + * This routine determines there is any ecc error or not
>> + *
>> + * Return: zero if there is no error otherwise returns 1
>> + */
>> +static int synopsys_edac_geterror_info(void __iomem *base,
>
> That base is an arg to readl, so it should be "const volatile void
> __iomem *base", no?
Yes. I will fix
>
>> + struct synopsys_ecc_status *perrstatus)
>
> Please align function args at the opening brace "("
>
>> +{
>> + u32 regval;
>> + u32 clearval = 0;
>> +
>> + regval = readl(base + SYNOPSYS_DDRC_ECC_STAT_REG_OFFSET) &
>> + (SYNOPSYS_DDRC_ECC_STATREG_UECOUNT_MASK |
>> + SYNOPSYS_DDRC_ECC_STATREG_CECOUNT_MASK);
>
> All your macro definitions start with "SYNOPSYS_DDRC" even though
> they're only local to this file. Which makes the code very unreadable
> due to their size.
>
> You could slim them up by dropping the prefix and doing:
>
> regval = readl(base + STAT_REG_OFFSET) &
> (STATREG_UECOUNT_MASK | STATREG_CECOUNT_MASK);
>
> which is much better IMO.
>
> The same is true for struct names like synopsys_ecc_status and the like.
> Shortening them too would slim down your code considerably.
>
>> +
>> + if (regval == 0)
>> + return 0;
>
> if (!regval)
>
>> +
>> + memset(perrstatus, 0, sizeof(struct synopsys_ecc_status));
>
> That perrstatus could also be shortened to errstat and make the whole
> code shorter:
>
> memset(errstat, 0, sizeof(*errstat));
>
> and so on...
>
>> +
>> + perrstatus->ce_count =
>> + (regval & SYNOPSYS_DDRC_ECC_STATREG_CECOUNT_MASK) >>
>> + SYNOPSYS_DDRC_ECC_STATREG_CECOUNT_SHIFT;
>> + perrstatus->ue_count =
>> + (regval & SYNOPSYS_DDRC_ECC_STATREG_UECOUNT_MASK);
>> +
>> + if (perrstatus->ce_count) {
>> + regval = readl(base + SYNOPSYS_DDRC_ECC_CE_LOG_REG_OFFSET);
>> + if (regval & SYNOPSYS_DDRC_ECC_CE_LOGREG_VALID) {
>> + perrstatus->ceinfo.bitpos = (regval &
>> + SYNOPSYS_DDRC_ECC_CE_LOGREG_BITPOS_MASK) >>
>> + SYNOPSYS_DDRC_ECC_CE_LOGREG_BITPOS_SHIFT;
>> + regval = readl(base +
>> + SYNOPSYS_DDRC_ECC_CE_ADDR_REG_OFFSET);
>> + perrstatus->ceinfo.row = (regval &
>> + SYNOPSYS_DDRC_ECC_ADDRREG_ROW_MASK) >>
>> + SYNOPSYS_DDRC_ECC_ADDRREG_ROW_SHIFT;
>> + perrstatus->ceinfo.col = (regval &
>> + SYNOPSYS_DDRC_ECC_ADDRREG_COL_MASK);
>> + perrstatus->ceinfo.bank = (regval &
>> + SYNOPSYS_DDRC_ECC_ADDRREG_BANK_MASK) >>
>> + SYNOPSYS_DDRC_ECC_ADDRREG_BANK_SHIFT;
>> + perrstatus->ceinfo.data = readl(base +
>> + SYNOPSYS_DDRC_ECC_CE_DATA_31_0_REG_OFFSET);
>> + edac_dbg(3, "ce bitposition: %d data: %d\n",
>> + perrstatus->ceinfo.bitpos,
>> + perrstatus->ceinfo.data);
>> + }
>> + clearval = SYNOPSYS_DDRC_ECCCTRL_CLR_CE_ERR;
>> + }
>> +
>> + if (perrstatus->ue_count) {
>> + regval = readl(base + SYNOPSYS_DDRC_ECC_UE_LOG_REG_OFFSET);
>> + if (regval & SYNOPSYS_DDRC_ECC_CE_LOGREG_VALID) {
>> + regval = readl(base +
>> + SYNOPSYS_DDRC_ECC_UE_ADDR_REG_OFFSET);
>> + perrstatus->ueinfo.row = (regval &
>> + SYNOPSYS_DDRC_ECC_ADDRREG_ROW_MASK) >>
>> + SYNOPSYS_DDRC_ECC_ADDRREG_ROW_SHIFT;
>> + perrstatus->ueinfo.col = (regval &
>> + SYNOPSYS_DDRC_ECC_ADDRREG_COL_MASK);
>> + perrstatus->ueinfo.bank = (regval &
>> + SYNOPSYS_DDRC_ECC_ADDRREG_BANK_MASK) >>
>> + SYNOPSYS_DDRC_ECC_ADDRREG_BANK_SHIFT;
>> + perrstatus->ueinfo.data = readl(base +
>> + SYNOPSYS_DDRC_ECC_UE_DATA_31_0_REG_OFFSET);
>
> All those assignments could be much shorter:
>
> p->ueinfo.row = (regval & ADDRREG_ROW_MASK) >> ADDRREG_ROW_SHIFT;
>
> Also, you can save yourself two indentation levels by flipping the
> logic:
>
> if (!(p->ce_count && (regval & CE_LOGREG_VALID)))
> goto ue_error;
>
> /* CE assignments */
>
> ...
>
> ue_error:
>
> if (!(p->ue_count && (regval & UE_LOGREG_VALID)))
> goto out;
>
> /* UE assignments */
>
> ...
>
> out:
>
> writel(...);
> writel(...);
>
>
>> + }
>> + clearval |= SYNOPSYS_DDRC_ECCCTRL_CLR_UE_ERR;
>> + }
>> +
>> + writel(clearval, base + SYNOPSYS_DDRC_ECC_CONTROL_REG_OFFSET);
>> + writel(0x0, base + SYNOPSYS_DDRC_ECC_CONTROL_REG_OFFSET);
>> +
>> + return 1;
>> +}
>> +
>> +/**
>> + * synopsys_edac_generate_message - Generate interpreted ECC status message
>> + * @mci: Pointer to the edac memory controller instance
>> + * @perrstatus: Pointer to the synopsys ecc status structure
>> + * @buffer: Pointer to the buffer in which to generate the
>> + * message
>> + * @size: The size, in bytes, of space available in buffer
>> + *
>> + * This routine generates to the provided buffer the portion of the
>> + * driver-unique report message associated with the ECC register of
>> + * the specified ECC status.
>> + */
>> +static void synopsys_edac_generate_message(const struct mem_ctl_info *mci,
>> + struct synopsys_ecc_status *perrstatus, char *buffer,
>> + size_t size)
>> +{
>> + struct ecc_error_info *pinfo = NULL;
>> +
>> + if (perrstatus->ce_count > 0)
>> + pinfo = &perrstatus->ceinfo;
>> + else
>> + pinfo = &perrstatus->ueinfo;
>> +
>> + snprintf(buffer, SYNOPSYS_EDAC_MESSAGE_SIZE,
>> + "DDR ECC error type :%s Row %d Bank %d Col %d ",
>> + (perrstatus->ce_count > 0) ? "CE" : "UE", pinfo->row,
>> + pinfo->bank, pinfo->col);
>> +}
>
> This function can be part of synopsys_edac_handle_error() below instead
> as it is simple enough and called only once there - no need for a
> separate function.
>
>> +
>> +/**
>> + * synopsys_edac_handle_error - Handle controller error types CE and UE
>> + * @mci: Pointer to the edac memory controller instance
>> + * @perrstatus: Pointer to the synopsys ecc status structure
>> + *
>> + * This routine handles the controller ECC correctable error.
>
> It handles also UC errors.
>
>> + */
>> +static void synopsys_edac_handle_error(struct mem_ctl_info *mci,
>> + struct synopsys_ecc_status *perrstatus)
>> +{
>> + char message[SYNOPSYS_EDAC_MESSAGE_SIZE];
>> +
>> + synopsys_edac_generate_message(mci, perrstatus, &message[0],
>> + SYNOPSYS_EDAC_MESSAGE_SIZE);
>> +
>> + if (perrstatus->ce_count)
>> + edac_mc_handle_error(HW_EVENT_ERR_CORRECTED, mci,
>> + perrstatus->ce_count, 0, 0, 0, 0, 0, -1,
>> + &message[0], "");
>
> string_array == &string_array[0]
>
>> + else
>> + edac_mc_handle_error(HW_EVENT_ERR_UNCORRECTED, mci,
>> + perrstatus->ue_count, 0, 0, 0, 0, 0, -1,
>> + &message[0], "");
>
> Ditto.
>
>> +}
>> +
>> +/**
>> + * synopsys_edac_check - Check controller for ECC errors
>> + * @mci: Pointer to the edac memory controller instance
>> + *
>> + * This routine is used to check and post ECC errors and is called by
>> + * the EDAC polling thread
>> + */
>> +static void synopsys_edac_check(struct mem_ctl_info *mci)
>> +{
>> + struct synopsys_edac_priv *priv = mci->pvt_info;
>> + struct synopsys_ecc_status errstatus;
>> + int status;
>> +
>> + status = synopsys_edac_geterror_info(priv->baseaddr, &errstatus);
>> + if (status) {
>
> Save an indentation level:
>
> if (!status)
> return;
>
> priv->ce_count += ...
> ...
>
>> + priv->ce_count += errstatus.ce_count;
>> + priv->ue_count += errstatus.ue_count;
>> +
>> + if (errstatus.ce_count) {
>> + synopsys_edac_handle_error(mci, &errstatus);
>> + errstatus.ce_count = 0;
>> + }
>> + if (errstatus.ue_count) {
>> + synopsys_edac_handle_error(mci, &errstatus);
>> + errstatus.ue_count = 0;
>> + }
>> + edac_dbg(3, "total error count ce %d ue %d\n",
>> + priv->ce_count, priv->ue_count);
>> + }
>> +}
>> +
>> +/**
>> + * synopsys_edac_get_dtype - Return the controller memory width
>> + * @base: Pointer to the ddr memory contoller base address
>> + *
>> + * This routine returns the EDAC device type width appropriate for the
>> + * current controller configuration.
>> + *
>> + * Return: a device type width enumeration.
>> + */
>> +static enum dev_type synopsys_edac_get_dtype(void __iomem *base)
>> +{
>> + enum dev_type dt;
>> + u32 width;
>> +
>> + width = readl(base + SYNOPSYS_DDRC_CONTROL_REG_OFFSET);
>> + width = (width & SYNOPSYS_DDRC_CTRLREG_BUSWIDTH_MASK) >>
>> + SYNOPSYS_DDRC_CTRLREG_BUSWIDTH_SHIFT;
>> +
>> + switch (width) {
>> + case SYNOPSYS_DDRCTL_WDTH_16:
>> + dt = DEV_X2;
>> + break;
>> + case SYNOPSYS_DDRCTL_WDTH_32:
>> + dt = DEV_X4;
>> + break;
>> + default:
>> + dt = DEV_UNKNOWN;
>> + }
>> +
>> + return dt;
>> +}
>> +
>> +/**
>> + * synopsys_edac_get_eccstate - Return the controller ecc enable/disable status
>> + * @base: Pointer to the ddr memory contoller base address
>> + *
>> + * This routine returns the ECC enable/diable status for the controller
>> + *
>> + * Return: a ecc status boolean i.e true/false - enabled/disabled.
>> + */
>> +static bool synopsys_edac_get_eccstate(void __iomem *base)
>> +{
>> + enum dev_type dt;
>> + u32 ecctype;
>> + bool state = false;
>> +
>> + dt = synopsys_edac_get_dtype(base);
>
> Shouldn't you handle the error case where it returns DEV_UNKNOWN type?
>
>> +
>> + ecctype = (readl(base + SYNOPSYS_DDRC_ECC_SCRUB_REG_OFFSET) &
>> + SYNOPSYS_DDRC_ECC_SCRUBREG_ECC_MODE_MASK);
>> +
>> + if ((ecctype == SYNOPSYS_DDRC_ECC_SCRUBREG_ECCMODE_SECDED)
>> + && (dt == DEV_X2)) {
>> + state = true;
>> + writel(0x0, base + SYNOPSYS_DDRC_ECC_CONTROL_REG_OFFSET);
>> + } else {
>> + state = false;
>
> Why that else branch? state is already initialized to false at function
> entry time.
>
>> + }
>> +
>> + return state;
>> +}
>> +
>> +/**
>> + * synopsys_edac_get_memsize - reads the size of the attached memory device
>> + *
>> + * Return: the memory size in bytes
>> + *
>> + * This routine returns the size of the system memory by reading the sysinfo
>> + * information
>> + */
>
> No need for that comment.
>
>> +static u32 synopsys_edac_get_memsize(void)
>> +{
>> + struct sysinfo inf;
>> +
>> + /* Reading the system memory size from the global meminfo structure */
>
> Same here.
>
> For simple functions where it is paramount what the function does, you
> don't need to comment every line.
>
>> + si_meminfo(&inf);
>> +
>> + return inf.totalram * inf.mem_unit;
>> +}
>> +
>> +/**
>> + * synopsys_edac_get_mtype - Returns controller memory type
>> + * @base: pointer to the synopsys ecc status structure
>> + *
>> + * This routine returns the EDAC memory type appropriate for the
>> + * current controller configuration.
>> + *
>> + * Return: a memory type enumeration.
>> + */
>> +static enum mem_type synopsys_edac_get_mtype(void __iomem *base)
>> +{
>> + enum mem_type mt;
>> + u32 memtype;
>> +
>> + memtype = readl(base + SYNOPSYS_DDRC_T_ZQ_REG_OFFSET);
>> +
>> + if (memtype & SYNOPSYS_DDRC_T_ZQ_REG_DDRMODE_MASK)
>> + mt = MEM_DDR3;
>> + else
>> + mt = MEM_DDR2;
>> +
>> + return mt;
>> +}
>> +
>> +/**
>> + * synopsys_edac_init_csrows - Initialize the cs row data
>> + * @mci: Pointer to the edac memory controller instance
>> + *
>> + * This routine initializes the chip select rows associated
>> + * with the EDAC memory controller instance
>> + *
>> + * Return: 0 if OK; otherwise, -EINVAL if the memory bank size
>
> Where is that -EINVAL being returned?
>
>> + * configuration cannot be determined.
>> + */
>> +static int synopsys_edac_init_csrows(struct mem_ctl_info *mci)
>> +{
>> + struct csrow_info *csi;
>> + struct dimm_info *dimm;
>> + struct synopsys_edac_priv *priv = mci->pvt_info;
>> + u32 size;
>> + int row, j;
>> +
>> + for (row = 0; row < mci->nr_csrows; row++) {
>> + csi = mci->csrows[row];
>> + size = synopsys_edac_get_memsize();
>> +
>> + for (j = 0; j < csi->nr_channels; j++) {
>> + dimm = csi->channels[j]->dimm;
>> + dimm->edac_mode = EDAC_FLAG_SECDED;
>> + dimm->mtype = synopsys_edac_get_mtype(priv->baseaddr);
>> + dimm->nr_pages =
>> + (size >> PAGE_SHIFT) / csi->nr_channels;
>> + dimm->grain = SYNOPSYS_EDAC_ERROR_GRAIN;
>> + dimm->dtype = synopsys_edac_get_dtype(priv->baseaddr);
>> + }
>> +
>> + }
>> +
>> + return 0;
>> +}
>> +
>> +/**
>> + * synopsys_edac_mc_init - Initialize driver instance
>> + * @mci: Pointer to the edac memory controller instance
>> + * @pdev: Pointer to the platform_device struct
>> + *
>> + * This routine performs initialization of the EDAC memory controller
>> + * instance and related driver-private data associated with the
>> + * memory controller the instance is bound to.
>> + *
>> + * Return: 0 if OK; otherwise, < 0 on error.
>> + */
>> +static int synopsys_edac_mc_init(struct mem_ctl_info *mci,
>> + struct platform_device *pdev)
>> +{
>> + int status;
>> + struct synopsys_edac_priv *priv;
>> +
>> + /* Initial driver pointers and private data */
>
> Useless comment.
>
>> + mci->pdev = &pdev->dev;
>> + priv = mci->pvt_info;
>> + platform_set_drvdata(pdev, mci);
>> +
>> + /* Initialize controller capabilities and configuration */
>> + mci->mtype_cap = MEM_FLAG_DDR3 | MEM_FLAG_DDR2;
>> + mci->edac_ctl_cap = EDAC_FLAG_NONE | EDAC_FLAG_SECDED;
>> + mci->scrub_cap = SCRUB_HW_SRC;
>> + /* Check the scrub setting from the controller */
>
> Useless comment.
>
>> + mci->scrub_mode = SCRUB_NONE;
>> +
>> + mci->edac_cap = EDAC_FLAG_SECDED;
>> + /* Initialize strings */
>
> Useless comment.
>
>> + mci->ctl_name = "synopsys_ddr_controller";
>> + mci->dev_name = dev_name(&pdev->dev);
>> + mci->mod_name = "synopsys_edac";
>
> Do a
>
> #define EDAC_MOD_STR "synopsys_edac"
>
> like the other drivers do and use that here and below.
>
>> + mci->mod_ver = "1";
>> +
>> + /* Initialize callbacks */
>
> Useless comment.
>
>> + edac_op_state = EDAC_OPSTATE_POLL;
>> + mci->edac_check = synopsys_edac_check;
>> + mci->ctl_page_to_phys = NULL;
>> +
>> + /*
>> + * Initialize the MC control structure 'csrows' table
>> + * with the mapping and control information.
>> + */
>> + status = synopsys_edac_init_csrows(mci);
>> + if (status)
>> + pr_err("Failed to initialize rows!\n");
>
> edac_printk
>
>> +
>> + return status;
>> +}
>> +
>> +/**
>> + * synopsys_edac_mc_probe - Check controller and bind driver
>> + * @pdev: Pointer to the platform_device struct
>> + *
>> + * This routine probes a specific controller
>> + * instance for binding with the driver.
>> + *
>> + * Return: 0 if the controller instance was successfully bound to the
>> + * driver; otherwise, < 0 on error.
>> + */
>> +static int synopsys_edac_mc_probe(struct platform_device *pdev)
>> +{
>> + struct mem_ctl_info *mci;
>> + struct edac_mc_layer layers[2];
>> + struct synopsys_edac_priv *priv;
>> + int rc;
>> + struct resource *res;
>> + void __iomem *baseaddr;
>> +
>> + /* Get the data from the platform device */
>> + res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
>
> You're not handling the case where this function returns NULL.
devm_ioremap_resource function will check this condition, so not
checking for the NULL explicitly
>
>> + baseaddr = devm_ioremap_resource(&pdev->dev, res);
>> + if (IS_ERR(baseaddr))
>> + return PTR_ERR(baseaddr);
>> +
>> + /* Check for the ecc enable status */
>> + if (synopsys_edac_get_eccstate(baseaddr) == false) {
>
> if (!get_eccstate(...))
>
>> + dev_err(&pdev->dev, "ecc not enabled\n");
>
> edac_printk pls.
>
>> + return -ENXIO;
>> + }
>> +
>> + /*
>> + * At this point, we know ECC is enabled, allocate an EDAC
>> + * controller instance and perform the appropriate
>> + * initialization.
>> + */
>> + layers[0].type = EDAC_MC_LAYER_CHIP_SELECT;
>> + layers[0].size = SYNOPSYS_EDAC_NR_CSROWS;
>> + layers[0].is_virt_csrow = true;
>> + layers[1].type = EDAC_MC_LAYER_CHANNEL;
>> + layers[1].size = SYNOPSYS_EDAC_NR_CHANS;
>> + layers[1].is_virt_csrow = false;
>> +
>> + mci = edac_mc_alloc(0, ARRAY_SIZE(layers), layers,
>> + sizeof(struct synopsys_edac_priv));
>> + if (mci == NULL) {
>
> if (!mci)
>
>> + pr_err("Failed memory allocation for mci instance!\n");
>
> edac_printk
>
> Before you return, devm_iounmap(... baseaddr). Just do it with nice goto
> labels.
Since the probe returns error, it it that the devm_ framework will clean these
resources automatically?
>
>> + return -ENOMEM;
>> + }
>> +
>> + priv = mci->pvt_info;
>> + priv->baseaddr = baseaddr;
>> + rc = synopsys_edac_mc_init(mci, pdev);
>> + if (rc) {
>> + pr_err("Failed to initialize instance!\n");
>
> edac_printk
>
>> + goto free_edac_mc;
>> + }
>> +
>> + /*
>> + * We have a valid, initialized EDAC instance bound to the
>> + * controller. Attempt to register it with the EDAC subsystem
>> + */
>> + rc = edac_mc_add_mc(mci);
>> + if (rc) {
>> + dev_err(&pdev->dev, "failed to register with EDAC core\n");
>
> Ditto.
>
>> + goto del_edac_mc;
>> + }
>> +
>> + return rc;
>> +
>> +del_edac_mc:
>> + edac_mc_del_mc(&pdev->dev);
>> +free_edac_mc:
>> + edac_mc_free(mci);
>> +
>> + return rc;
>> +}
>> +
>> +/**
>> + * synopsys_edac_mc_remove - Unbind driver from controller
>> + * @pdev: Pointer to the platform_device struct
>> + *
>> + * This routine unbinds the EDAC memory controller instance associated
>> + * with the specified controller described by the
>> + * OpenFirmware device tree node passed as a parameter.
>> + *
>> + * Return: Unconditionally 0
>> + */
>
> No need for that whole comment at all.
>
>> +static int synopsys_edac_mc_remove(struct platform_device *pdev)
>> +{
>> + struct mem_ctl_info *mci = platform_get_drvdata(pdev);
>> +
>> + edac_mc_del_mc(&pdev->dev);
>> + edac_mc_free(mci);
>> +
>> + return 0;
>> +}
>> +
>> +/* Device tree node type and compatible tuples this driver can match on */
>
> Ditto.
>
>> +static struct of_device_id synopsys_edac_match[] = {
>> + { .compatible = "xlnx,zynq-ddrc-1.04", },
>> + { /* end of table */ }
>> +};
>> +
>> +MODULE_DEVICE_TABLE(of, synopsys_edac_match);
>> +
>> +static struct platform_driver synopsys_edac_mc_driver = {
>> + .driver = {
>> + .name = "synopsys-edac",
>> + .owner = THIS_MODULE,
>> + .of_match_table = synopsys_edac_match,
>> + },
>> + .probe = synopsys_edac_mc_probe,
>> + .remove = synopsys_edac_mc_remove,
>> +};
>> +
>> +module_platform_driver(synopsys_edac_mc_driver);
>> +
>> +MODULE_AUTHOR("Xilinx, Inc.");
>> +MODULE_DESCRIPTION("Synopsys DDR ECC driver");
>> +MODULE_LICENSE("GPL v2");
>
> Ok, that should be it so far, more review with the next submission.
Thanks for the detailed review and i will implement all other comments and will
send the next version once we conclude the resolutions for my above concerns
and comments.
Regrads,
Punnaiah
>
> --
> Regards/Gruss,
> Boris.
>
> Sent from a fat crate under my desk. Formatting is fine.
> --
On Wed, Apr 09, 2014 at 11:34:31AM +0530, punnaiah choudary kalluri wrote:
> Since it is recommended in Documentation/kernel-doc-nano-HOWTO.txt
> but also said it is low priority and at the discretion of the MAINTAINER of
> that kernel source file
>
> So, if you recommend not use kernel-doc annotation then i will take care
> in next version.
I don't think there's need for documenting those functions for
kernel-doc - it is an EDAC driver and not some subsystem function with a
lot of users.
> devm_ioremap_resource function will check this condition, so not
> checking for the NULL explicitly
And it would say
dev_err(dev, "invalid resource\n");
so we'll know what fails, ok.
> Since the probe returns error, it it that the devm_ framework will
> clean these resources automatically?
Ah, that's the managed thing, ok.
Thanks.
--
Regards/Gruss,
Boris.
Sent from a fat crate under my desk. Formatting is fine.
--
On Wed, Apr 9, 2014 at 4:33 PM, Borislav Petkov <[email protected]> wrote:
> On Wed, Apr 09, 2014 at 11:34:31AM +0530, punnaiah choudary kalluri wrote:
>> Since it is recommended in Documentation/kernel-doc-nano-HOWTO.txt
>> but also said it is low priority and at the discretion of the MAINTAINER of
>> that kernel source file
>>
>> So, if you recommend not use kernel-doc annotation then i will take care
>> in next version.
>
> I don't think there's need for documenting those functions for
> kernel-doc - it is an EDAC driver and not some subsystem function with a
> lot of users.
OK. I just checked at highbank_mc_edac and other drivers found no documentation
for static functions. So, is it ok if i remove documentation
completely for these static
functions ?
Thanks,
Punnaiah
>
>> devm_ioremap_resource function will check this condition, so not
>> checking for the NULL explicitly
>
> And it would say
>
> dev_err(dev, "invalid resource\n");
>
> so we'll know what fails, ok.
>
>> Since the probe returns error, it it that the devm_ framework will
>> clean these resources automatically?
>
> Ah, that's the managed thing, ok.
>
> Thanks.
>
> --
> Regards/Gruss,
> Boris.
>
> Sent from a fat crate under my desk. Formatting is fine.
> --
> --
> To unsubscribe from this list: send the line "unsubscribe linux-edac" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
On Wed, Apr 09, 2014 at 11:15:00PM +0530, punnaiah choudary kalluri wrote:
> OK. I just checked at highbank_mc_edac and other drivers found
> no documentation for static functions. So, is it ok if i remove
> documentation completely for these static functions?
Yes. Didn't I already suggested that, like, twice already?
Just drop them.
--
Regards/Gruss,
Boris.
Sent from a fat crate under my desk. Formatting is fine.
--
Hi Borislav and Punnaiah,
some comments below.
>> diff --git a/drivers/edac/synopsys_edac.c b/drivers/edac/synopsys_edac.c
>> new file mode 100644
>> index 0000000..7cec331
>> --- /dev/null
>> +++ b/drivers/edac/synopsys_edac.c
>> @@ -0,0 +1,614 @@
>> +/*
>> + * Synopsys DDR ECC Driver
>> + * This driver is based on ppc4xx_edac.c drivers
>> + *
>> + * Copyright (C) 2012 - 2014 Xilinx, Inc.
>
> The same question to you: are you going to maintain this driver? If so,
> please consider adding yourself to the MAINTAINERS file so that people
> reporting issues with it can send you a note.
Just add it to Zynq maintainer fragment as we are doing for non zynq/xilinx
drivers.
>> +/**
>> + * struct synopsys_edac_priv - DDR memory controller private instance data
>> + * @baseaddr: Base address of the DDR controller
>> + * @ce_count: Correctable Error count
>> + * @ue_count: Uncorrectable Error count
>> + */
>> +struct synopsys_edac_priv {
>> + void __iomem *baseaddr;
>> + u32 ce_count;
>> + u32 ue_count;
>> +};
>> +
>> +/**
>
> Why do we need the kernel-doc annotation for all those static functions?
I don't want to fight on this but having documentation help with reading
the code. I also don't think that only non static functions are documented.
I can't see any problem to have kernel-doc for static functions.
At least the is the first time when someone saying that only some functions
should be documented.
> Also, drop the "synopsys_edac_" prefix of all static functions - that'll
> slim up the code even further.
I don't think this is good to do. When we remove this prefix entirely
it is bigger chance that the same function name will be used by another driver.
It is not a problem with linking but the same functions names will be listed
in System.map which will complicate debugging.
For ARM multiplatform kernel I think that it is almost must to use
vendor prefixes for all platform/driver code.
All drivers can use instead of name_probe just probe but I don't think
this is the right way to go.
The rest was commented by Punnaiah.
Thanks,
Michal
--
Michal Simek, Ing. (M.Eng), OpenPGP -> KeyID: FE3D1F91
w: http://www.monstr.eu p: +42-0-721842854
Maintainer of Linux kernel - Microblaze cpu - http://www.monstr.eu/fdt/
Maintainer of Linux kernel - Xilinx Zynq ARM architecture
Microblaze U-BOOT custodian and responsible for u-boot arm zynq platform
On Thu, Apr 10, 2014 at 08:01:36AM +0200, Michal Simek wrote:
> Just add it to Zynq maintainer fragment as we are doing for non
> zynq/xilinx drivers.
Yep, as a file pattern F:
> At least the is the first time when someone saying that only some
> functions should be documented.
I don't see the need to document simple functions which are *obvious*.
But if you want to do that, I'm not going to stop you so be my guest.
> I don't think this is good to do. When we remove this prefix entirely
> it is bigger chance that the same function name will be used by
> another driver. It is not a problem with linking but the same
> functions names will be listed in System.map which will complicate
> debugging.
Ok, fair enough. You probably still want to make them shorter, though,
so that your code doesn't have to be line-broken too much. For example,
since we only load a single edac driver per system, you can shorten the
prefix to "edac_" only.
Thanks.
--
Regards/Gruss,
Boris.
Sent from a fat crate under my desk. Formatting is fine.
--
On 04/10/2014 11:21 AM, Borislav Petkov wrote:
> On Thu, Apr 10, 2014 at 08:01:36AM +0200, Michal Simek wrote:
>> Just add it to Zynq maintainer fragment as we are doing for non
>> zynq/xilinx drivers.
>
> Yep, as a file pattern F:
yep.
>> At least the is the first time when someone saying that only some
>> functions should be documented.
>
> I don't see the need to document simple functions which are *obvious*.
> But if you want to do that, I'm not going to stop you so be my guest.
ok. thanks.
>> I don't think this is good to do. When we remove this prefix entirely
>> it is bigger chance that the same function name will be used by
>> another driver. It is not a problem with linking but the same
>> functions names will be listed in System.map which will complicate
>> debugging.
>
> Ok, fair enough. You probably still want to make them shorter, though,
> so that your code doesn't have to be line-broken too much. For example,
> since we only load a single edac driver per system, you can shorten the
> prefix to "edac_" only.
I agree with you that we can use shorter name.
I have checked this with Punnaiah and I can't see any problem to have
more edac drivers in the system.
For zynq there is primary ddr controller which is target by this driver.
Then we are talking about L2 in another thread.
Because zynq has programmable logic next to chip where you can add
soft memory controller (we have these hw designs) on the same bus
then there could be others edac drivers.
I would say how many you like till pins on the package.
Thanks,
Michal
On Thu, Apr 10, 2014 at 11:49:57AM +0200, Michal Simek wrote:
> I agree with you that we can use shorter name. I have checked this
> with Punnaiah and I can't see any problem to have more edac drivers
> in the system. For zynq there is primary ddr controller which is
> target by this driver. Then we are talking about L2 in another thread.
> Because zynq has programmable logic next to chip where you can add
> soft memory controller (we have these hw designs) on the same bus then
> there could be others edac drivers. I would say how many you like till
> pins on the package.
I knew you were gonna say something like that! :-)
Then you'll have to check whether edac can even stomach more than one
drivers in parallel. TTBOMK, no one has ever done this so far... Yadda
yadda.
But before you do all that, you need to check whether edac is really
what you're looking for, as I put it in my other mail.
--
Regards/Gruss,
Boris.
Sent from a fat crate under my desk. Formatting is fine.
--
On Thu, Apr 10, 2014 at 12:00:47PM +0200, Borislav Petkov wrote:
> Then you'll have to check whether edac can even stomach more than one
> drivers in parallel. TTBOMK, no one has ever done this so far... Yadda
> yadda.
Alternatively, you could put together a single edac driver which
contains both zynq and L2 functionality.
--
Regards/Gruss,
Boris.
Sent from a fat crate under my desk. Formatting is fine.
--
On 04/10/2014 12:00 PM, Borislav Petkov wrote:
> On Thu, Apr 10, 2014 at 11:49:57AM +0200, Michal Simek wrote:
>> I agree with you that we can use shorter name. I have checked this
>> with Punnaiah and I can't see any problem to have more edac drivers
>> in the system. For zynq there is primary ddr controller which is
>> target by this driver. Then we are talking about L2 in another thread.
>> Because zynq has programmable logic next to chip where you can add
>> soft memory controller (we have these hw designs) on the same bus then
>> there could be others edac drivers. I would say how many you like till
>> pins on the package.
>
> I knew you were gonna say something like that! :-)
:-) good. We are getting famous by non standard cases.
>
> Then you'll have to check whether edac can even stomach more than one
> drivers in parallel. TTBOMK, no one has ever done this so far... Yadda
> yadda.
It can there is just a problem with some IDs. :-)
> But before you do all that, you need to check whether edac is really
> what you're looking for, as I put it in my other mail.
If edac is the interface for reporting problems like this. It is probably
just a limitation of current edac core and should be fixed.
Thanks,
Michal
On 04/10/2014 12:05 PM, Borislav Petkov wrote:
> On Thu, Apr 10, 2014 at 12:00:47PM +0200, Borislav Petkov wrote:
>> Then you'll have to check whether edac can even stomach more than one
>> drivers in parallel. TTBOMK, no one has ever done this so far... Yadda
>> yadda.
>
> Alternatively, you could put together a single edac driver which
> contains both zynq and L2 functionality.
That's seems to more problematic because we just don't know what our
customers will do and what combinations they will use.
Thanks,
Michal