Em Tue, 16 Jul 2024 16:03:27 +0100
<shiju.jose@xxxxxxxxxx> escreveu:
> From: Shiju Jose <shiju.jose@xxxxxxxxxx>
>
> Add EDAC ECS (Error Check Scrub) control driver supports configuring
> the memory device's ECS feature.
>
> The Error Check Scrub (ECS) is a feature defined in JEDEC DDR5 SDRAM
> Specification (JESD79-5) and allows the DRAM to internally read, correct
> single-bit errors, and write back corrected data bits to the DRAM array
> while providing transparency to error counts.
>
> The DDR5 device contains number of memory media FRUs per device. The
> DDR5 ECS feature and thus the ECS control driver supports configuring
> the ECS parameters per FRU.
>
> The memory devices supports ECS feature register with the EDAC ECS driver
typo:
supports -> support
> and thus with the generic EDAC RAS feature driver, which adds the sysfs
> ECS control interface. The ECS control attributes are exposed to the
> userspace in /sys/bus/edac/devices/<dev-name>/ecs_fruX/.
>
> Generic EDAC ECS driver and the common sysfs ECS interface promotes
> unambiguous control from the userspace irrespective of the underlying
> devices, support ECS feature.
>
> The support for ECS feature is added separately because the DDR5 ECS
> feature's control attributes are dissimilar from those of the scrub
> feature.
>
> Note: Documentation can be added if necessary.
Please document.
>
> Co-developed-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@xxxxxxxxxx>
> Signed-off-by: Shiju Jose <shiju.jose@xxxxxxxxxx>
> ---
> drivers/edac/Makefile | 2 +-
> drivers/edac/edac_ecs.c | 396 +++++++++++++++++++++++++++++++
> drivers/edac/edac_ras_feature.c | 5 +
> include/linux/edac_ras_feature.h | 36 +++
> 4 files changed, 438 insertions(+), 1 deletion(-)
> create mode 100755 drivers/edac/edac_ecs.c
>
> diff --git a/drivers/edac/Makefile b/drivers/edac/Makefile
> index de56cbd039eb..c1412c7d3efb 100644
> --- a/drivers/edac/Makefile
> +++ b/drivers/edac/Makefile
> @@ -10,7 +10,7 @@ obj-$(CONFIG_EDAC) := edac_core.o
>
> edac_core-y := edac_mc.o edac_device.o edac_mc_sysfs.o
> edac_core-y += edac_module.o edac_device_sysfs.o wq.o
> -edac_core-y += edac_ras_feature.o edac_scrub.o
> +edac_core-y += edac_ras_feature.o edac_scrub.o edac_ecs.o
>
> edac_core-$(CONFIG_EDAC_DEBUG) += debugfs.o
>
> diff --git a/drivers/edac/edac_ecs.c b/drivers/edac/edac_ecs.c
> new file mode 100755
> index 000000000000..37dabd053c36
> --- /dev/null
> +++ b/drivers/edac/edac_ecs.c
> @@ -0,0 +1,396 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * ECS driver supporting controlling on die error check scrub
> + * (e.g. DDR5 ECS). The common sysfs ECS interface promotes
> + * unambiguous access from the userspace.
> + *
> + * Copyright (c) 2024 HiSilicon Limited.
> + */
> +
> +#define pr_fmt(fmt) "EDAC ECS: " fmt
> +
> +#include <linux/edac_ras_feature.h>
> +
> +#define EDAC_ECS_FRU_NAME "ecs_fru"
> +
> +enum edac_ecs_attributes {
> + ecs_log_entry_type,
> + ecs_log_entry_type_per_dram,
> + ecs_log_entry_type_per_memory_media,
> + ecs_mode,
> + ecs_mode_counts_rows,
> + ecs_mode_counts_codewords,
> + ecs_reset,
> + ecs_name,
> + ecs_threshold,
> + ecs_max_attrs
> +};
Please use uppercase for enums.
> +
> +struct edac_ecs_dev_attr {
> + struct device_attribute dev_attr;
> + int fru_id;
> +};
> +
> +struct edac_ecs_fru_context {
> + char name[EDAC_RAS_NAME_LEN];
> + struct edac_ecs_dev_attr ecs_dev_attr[ecs_max_attrs];
> + struct attribute *ecs_attrs[ecs_max_attrs + 1];
> + struct attribute_group group;
> +};
> +
> +struct edac_ecs_context {
> + u16 num_media_frus;
> + struct edac_ecs_fru_context *fru_ctxs;
> +};
> +
> +#define to_ecs_dev_attr(_dev_attr) \
> + container_of(_dev_attr, struct edac_ecs_dev_attr, dev_attr)
> +
> +static ssize_t log_entry_type_show(struct device *ras_feat_dev,
> + struct device_attribute *attr,
> + char *buf)
> +{
> + struct edac_ecs_dev_attr *ecs_dev_attr = to_ecs_dev_attr(attr);
> + struct edac_ras_feat_ctx *ctx = dev_get_drvdata(ras_feat_dev);
> + const struct edac_ecs_ops *ops = ctx->ecs.ops;
> + u64 val;
> + int ret;
> +
> + ret = ops->get_log_entry_type(ras_feat_dev->parent, ctx->ecs.private,
> + ecs_dev_attr->fru_id, &val);
> + if (ret)
> + return ret;
Same notes as patch 2/11 with regards to sysfs documentation/store/show.
Also, it is hard to review this patch without the ABI documentation.
Regards,
Mauro
Thanks,
Mauro
