| 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
| ||
|
Message-ID: <5f03bbd7-9c21-6d62-20b2-56a4ba1819e0@kernel.org>
Date: Wed, 8 Feb 2023 11:17:21 +0200
From: Roger Quadros <rogerq@...nel.org>
To: Md Danish Anwar <a0501179@...com>, Andrew Lunn <andrew@...n.ch>,
MD Danish Anwar <danishanwar@...com>
Cc: "Andrew F. Davis" <afd@...com>, Suman Anna <s-anna@...com>,
YueHaibing <yuehaibing@...wei.com>,
Vignesh Raghavendra <vigneshr@...com>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Rob Herring <robh+dt@...nel.org>,
Paolo Abeni <pabeni@...hat.com>,
Jakub Kicinski <kuba@...nel.org>,
Eric Dumazet <edumazet@...gle.com>,
"David S. Miller" <davem@...emloft.net>, nm@...com,
ssantosh@...nel.org, srk@...com, linux-kernel@...r.kernel.org,
devicetree@...r.kernel.org, netdev@...r.kernel.org,
linux-omap@...r.kernel.org, linux-arm-kernel@...ts.infradead.org
Subject: Re: [EXTERNAL] Re: [EXTERNAL] Re: [PATCH v4 2/2] net: ti:
icssg-prueth: Add ICSSG ethernet driver
Danish,
On 08/02/2023 09:46, Md Danish Anwar wrote:
> Hi Roger and Andrew,
>
> On 08/02/23 01:26, Roger Quadros wrote:
>> Danish,
>>
>> On 07/02/2023 17:29, Md Danish Anwar wrote:
>>> Hi Andrew,
>>>
>>> On 06/02/23 19:45, Andrew Lunn wrote:
>>>>> +enum mii_mode {
>>>>> + MII_MODE_MII = 0,
>>>>> + MII_MODE_RGMII,
>>>>> + MII_MODE_SGMII
>>>>
>>>> There is no mention of SGMII anywhere else. And in a couple of places,
>>>> the code makes the assumption that if it is not RGMII it is MII.
>>>>
>>>> Does the hardware really support SGMII?
>>>>
>>>
>>> As far as I know, the hardware does support SGMII but it's not yet supported by
>>> the driver. I will drop the SGMII because it's not needed as of now. If in
>>> future support for SGMII is there, I'll add it.
>>>
>>>>> +static int prueth_config_rgmiidelay(struct prueth *prueth,
>>>>> + struct device_node *eth_np,
>>>>> + phy_interface_t phy_if)
>>>>> +{
>>>>
>>>> ...
>>>>
>>>>> + if (phy_if == PHY_INTERFACE_MODE_RGMII_ID ||
>>>>> + phy_if == PHY_INTERFACE_MODE_RGMII_TXID)
>>>>> + rgmii_tx_id |= ICSSG_CTRL_RGMII_ID_MODE;
>>>>> +
>>>>> + regmap_update_bits(ctrl_mmr, icssgctrl_reg, ICSSG_CTRL_RGMII_ID_MODE, rgmii_tx_id);
>>
>> This is only applicable to some devices so you need to restrict this only
>> to those devices.
>>
>
> Currently ICSSG driver is getting upstreamed for AM65 SR2.0 device, so I don't
> think there is any need for any device related restriction. Once support for
> other devices are enabled for upstream, we can modify this accordingly.
OK then please get rid of prueth_config_rgmiidelay() entirely.
>
>> And only when you enable MAC TX delay you need to change emac->phy_if to PHY_INTERFACE_MODE_RGMII_RXID
>> if it was PHY_INTERFACE_MODE_RGMII_ID.
>>
>>>>
>>>> Here you are adding the TX delay if the phy-mode indicates it should
>>>> be added.
>>>>
>>>>> +static int prueth_netdev_init(struct prueth *prueth,
>>>>> + struct device_node *eth_node)
>>>>> +{
>>>>
>>>>> + ret = of_get_phy_mode(eth_node, &emac->phy_if);
>>>>> + if (ret) {
>>>>> + dev_err(prueth->dev, "could not get phy-mode property\n");
>>>>> + goto free;
>>>>> + }
>>>>
>>>>> + ret = prueth_config_rgmiidelay(prueth, eth_node, emac->phy_if);
>>>>> + if (ret)
>>>>> + goto free;
>>>>> +
>>>>
>>>> Reading it from DT and calling the delay function.
>>>>
>>>>> +static int prueth_probe(struct platform_device *pdev)
>>>>> +{
>>>>
>>>>
>>>>> + /* register the network devices */
>>>>> + if (eth0_node) {
>>>>> + ret = register_netdev(prueth->emac[PRUETH_MAC0]->ndev);
>>>>> + if (ret) {
>>>>> + dev_err(dev, "can't register netdev for port MII0");
>>>>> + goto netdev_exit;
>>>>> + }
>>>>> +
>>>>> + prueth->registered_netdevs[PRUETH_MAC0] = prueth->emac[PRUETH_MAC0]->ndev;
>>>>> +
>>>>> + emac_phy_connect(prueth->emac[PRUETH_MAC0]);
>>>>
>>>> And this is connecting the MAC and the PHY, where emac_phy_connect()
>>>> passes emac->phy_if to phylib.
>>>>
>>>> What i don't see anywhere is you changing emac->phy_if to indicate the
>>>> MAC has inserted the TX delay, and so the PHY should not.
>>>>
>>>
>>> Yes, there is no indication whether MAC has enabled TX delay or not. I have
>>> changed the phy-mode in DT from "rgmii-rxid" to "rgmii-id" as per your
>>> suggestion in previous revision. I will keep Tx Internal delay as it is(getting
>>> configured in MAC) and inside emac_phy_connect() API, while calling
>>> of_phy_connect() instead of passing emac->phy_if (which is rgmii-id as per DT),
>>> I will pass PHY_INTERFACE_MODE_RGMII_RXID. This will make sure that phy only
>>> enables Rx delay and keep the existing approach of keepping Tx delay in MAC.
>>>
>>> Currently, in emac_phy_connect() API,
>>>
>>> /* connect PHY */
>>> ndev->phydev = of_phy_connect(emac->ndev, emac->phy_node,
>>> &emac_adjust_link, 0,
>>> emac->phy_if);
>>> I will change it to,
>>>
>>> /* connect PHY */
>>> ndev->phydev = of_phy_connect(emac->ndev, emac->phy_node,
>>> &emac_adjust_link, 0,
>>> PHY_INTERFACE_MODE_RGMII_RXID);
>>>
>>> Let me know if this looks OK.
>>
>> No, this is not OK.
>>
>> Please keep this as emac->phy_if.
>>
>> In prueth_config_rgmiidelay(), you can change emac->phy_if to
>> PHY_INTERFACE_MODE_RGMII_RXID only if it was RGMII mode
>> *and* MAC TX delay was enabled.
>>
>
> I checked the latest Technical Reference Manual [1] (Section 5.1.3.4.49, Table
> 5-624) for AM65 Silicon Revision 2.0.
>
> Below is the description in Table 5-624
>
> BIT : 24
> Field : RGMII0_ID_MODE
> Type : R/W
> Reset : 0h
> Description : Controls the PRU_ICSSG0 RGMII0 port internal transmit delay
> 0h - Internal transmit delay is enabled
> 1h - Reserved
>
> The TX internal delay is always enabled and couldn't be disabled as 1h is
> reserved. So hardware support for disabling TX internal delay is not there.
>
> As, TX internal delay is always there, there is no need to enable it in MAC or
> PHY. So no need of API prueth_config_rgmiidelay().
>
> My approach to handle delay would be as below.
>
> *) Keep phy-mode = "rgmii-id" in DT as asked by Andrew.
> *) Let TX internal delay enabled in Hardware.
You mean in TX delay in MAC? Why? If Silicon does not have an issue
then it should be enabled in PHY not in MAC.
> *) Let PHY configure RX internal delay.
No. PHY should do whatever is asked in the DT. In this case both TX and RX delay.
> *) Remove prueth_config_rgmiidelay() API is there is no use of this. TX
Agreed.
> Internal delay is always enabled.
no.
> *) Instead of calling prueth_config_rgmiidelay() API in prueth_netdev_init()
> API, add below if condition.
>
> if(emac->phy_if == PHY_INTERFACE_MODE_RGMII_ID)
> emac->phy_if == PHY_INTERFACE_MODE_RGMII_RXID
No. This will remain emac->phy_if.
>
> This will check if phy-mode is "rgmii-id", then change the phy-mode to
> "rgmii-rxid". The same emac->phy_if is passed to emac_phy_connect which passes
> emac->phy_if to phylib. So that PHY will only enable RX internal delay and TX
> internal delay is always enabled by default(in Hardware)
>
No need of all this complexity.
> As of now ICSSG driver is getting upstreamed with support for only AM65 SR2.0,
OK. Then let's just deal with SR2.0 for now.
Pre-production devices are never up-streamed.
> In future when ICSSG driver starts supporting other devices, we can modify this
> condition to something like below
>
> if(device_is_AM65 && emac->phy_if == PHY_INTERFACE_MODE_RGMII_ID)
> emac->phy_if == PHY_INTERFACE_MODE_RGMII_RXID
>
> So that, for other devices, phy-mode should remain as it is, as other devices
> don't have hardware restriction.
>
> Please let me know if this is the right approach.
>
> [1]
> https://www.ti.com/lit/ug/spruid7e/spruid7e.pdf?ts=1675841023830&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FAM6548#%5B%7B%22num%22%3A1706%2C%22gen%22%3A0%7D%2C%7B%22name%22%3A%22XYZ%22%7D%2C0%2C717.9%2C0%5D
>
>> cheers,
>> -roger
cheers,
-roger.
Powered by blists - more mailing lists