Re: [PATCH net-next 4/7] net: hibmcge: Add register dump supported in this module

[Andrew Lunn]

On Wed, Oct 23, 2024 at 09:42:10PM +0800, Jijie Shao wrote:
> With the ethtool of a specific version,
> the following effects are achieved:
> 
> [root@localhost sjj]# ./ethtool -d enp131s0f1
> [SPEC] VALID                    [0x0000]: 0x00000001
> [SPEC] EVENT_REQ                [0x0004]: 0x00000000
> [SPEC] MAC_ID                   [0x0008]: 0x00000002
> [SPEC] PHY_ADDR                 [0x000c]: 0x00000002
> [SPEC] MAC_ADDR_L               [0x0010]: 0x00000808
> [SPEC] MAC_ADDR_H               [0x0014]: 0x08080802
> [SPEC] UC_MAX_NUM               [0x0018]: 0x00000004
> [SPEC] MAX_MTU                  [0x0028]: 0x00000fc2
> [SPEC] MIN_MTU                  [0x002c]: 0x00000100

Seems like this makes your debugfs patches redundant?

> +static u32 hbg_get_reg_info(struct hbg_priv *priv,
> +			    const struct hbg_reg_type_info *type_info,
> +			    const struct hbg_reg_offset_name_map *reg_map,
> +			    struct hbg_reg_info *info)
> +{
> +	info->val = hbg_reg_read(priv, reg_map->reg_offset);
> +	info->offset = reg_map->reg_offset - type_info->offset_base;
> +	snprintf(info->name, sizeof(info->name),
> +		 "[%s] %s", type_info->name, reg_map->name);
> +
> +	return sizeof(*info);
> +}
> +
> +static void hbg_ethtool_get_regs(struct net_device *netdev,
> +				 struct ethtool_regs *regs, void *data)
> +{
> +	struct hbg_priv *priv = netdev_priv(netdev);
> +	const struct hbg_reg_type_info *info;
> +	u32 i, j, offset = 0;
> +
> +	regs->version = 0;
> +	for (i = 0; i < ARRAY_SIZE(hbg_type_infos); i++) {
> +		info = &hbg_type_infos[i];
> +		for (j = 0; j < info->reg_num; j++)
> +			offset += hbg_get_reg_info(priv, info,
> +						   &info->reg_maps[j],
> +						   data + offset);
> +	}
> +}

data is supposed to be just raw values, dumped from registers in the
device. You appear to be passing back ASCII text. It is supposed to be
ethtool which does the pretty print, not the kernel driver.

    Andrew

---
pw-bot: cr