netdev - Re: [PATCH] ethtool: igb: dump RR2DCDELAY register

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 15 Jul 2019 11:13:34 +0200
From:   Michal Kubecek <mkubecek@...e.cz>
To:     Artem Bityutskiy <artem.bityutskiy@...ux.intel.com>
Cc:     "John W . Linville" <linville@...driver.com>,
        netdev@...r.kernel.org
Subject: Re: [PATCH] ethtool: igb: dump RR2DCDELAY register

On Mon, Jul 15, 2019 at 09:52:28AM +0300, Artem Bityutskiy wrote:
> Decode 'RR2DCDELAY' register which Linux kernel provides starting from version
> 5.3. The corresponding commit in the Linux kernel is:
>     cd502a7f7c9c igb: add RR2DCDELAY to ethtool registers dump
> 
> The RR2DCDELAY register is present in I210 and I211 Intel Gigabit Ethernet
> chips and it stands for "Read Request To Data Completion Delay". Here is how
> this register is described in the I210 datasheet:
> 
> "This field captures the maximum PCIe split time in 16 ns units, which is the
> maximum delay between the read request to the first data completion. This is
> giving an estimation of the PCIe round trip time."
> 
> In practice, this register can be used to measure the time it takes the NIC to
> read data from the host memory.
> 
> Signed-off-by: Artem Bityutskiy <artem.bityutskiy@...ux.intel.com>
> ---
>  igb.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
> 
> diff --git a/igb.c b/igb.c
> index e0ccef9..ab0896f 100644
> --- a/igb.c
> +++ b/igb.c
> @@ -859,6 +859,18 @@ igb_dump_regs(struct ethtool_drvinfo *info, struct ethtool_regs *regs)
>  		"0x03430: TDFPC       (Tx data FIFO packet count)      0x%08X\n",
>  		regs_buff[550]);
>  
> +	/*
> +	 * Starting from kernel version 5.3 the registers dump buffer grew from
> +	 * 739 4-byte words to 740 words, and word 740 contains the RR2DCDLAY

nit: "E" missing here:                                                    ^

> +	 * register.
> +	 */
> +	if (regs->len < 740)
> +		return 0;
> +
> +	fprintf(stdout,
> +		"0x05BF4: RR2DCDELAY  (Max. DMA read delay)            0x%08X\n",
> +		regs_buff[739]);

Showing a delay as hex number doesn't seem very user friendly but that
also applies to many existing registers so it's probably better to be
consistent and perhaps do an overall cleanup later.

Michal