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:   Wed, 13 Dec 2023 17:28:12 +0530
From:   Sanath S <sanaths2@....com>
To:     Mika Westerberg <mika.westerberg@...ux.intel.com>,
        Sanath S <Sanath.S@....com>
Cc:     mario.limonciello@....com, andreas.noever@...il.com,
        michael.jamet@...el.com, YehezkelShB@...il.com,
        linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [Patch v2 1/2] thunderbolt: Introduce tb_switch_reset_ports(),
 tb_port_reset() and usb4_port_reset()


On 12/13/2023 11:29 AM, Mika Westerberg wrote:
> On Wed, Dec 13, 2023 at 12:46:34AM +0530, Sanath S wrote:
>> Introduce the tb_switch_reset_ports() function that resets the
>> downstream ports of a given switch. This helps us reset the USB4
>> links created by boot firmware during the init sequence.
>>
>> Introduce the tb_port_reset() helper function that resets the
>> given port.
>>
>> Introduce the usb4_port_reset() function that performs the DPR
>> of a given port. This function follows the CM guide specification 7.3
>>
>> Suggested-by: Mario Limonciello <mario.limonciello@....com>
>> Signed-off-by: Sanath S <Sanath.S@....com>
>> ---
>>   drivers/thunderbolt/switch.c  | 35 +++++++++++++++++++++++++++++++
>>   drivers/thunderbolt/tb.h      |  2 ++
>>   drivers/thunderbolt/tb_regs.h |  1 +
>>   drivers/thunderbolt/usb4.c    | 39 +++++++++++++++++++++++++++++++++++
>>   4 files changed, 77 insertions(+)
>>
>> diff --git a/drivers/thunderbolt/switch.c b/drivers/thunderbolt/switch.c
>> index 44e9b09de47a..ef7ed92fd48e 100644
>> --- a/drivers/thunderbolt/switch.c
>> +++ b/drivers/thunderbolt/switch.c
>> @@ -626,6 +626,19 @@ int tb_port_unlock(struct tb_port *port)
>>   	return 0;
>>   }
>>   
>> +/**
>> + * tb_port_reset() - Reset downstream port
>> + * @port: Port to reset
>> + *
>> + * Helps to reconfigure the USB4 link by resetting the downstream port.
>> + *
>> + * Return: Returns 0 on success or an error code on failure.
>> + */
>> +static int tb_port_reset(struct tb_port *port)
>> +{
>> +	return usb4_port_reset(port);
>> +}
>> +
>>   static int __tb_port_enable(struct tb_port *port, bool enable)
>>   {
>>   	int ret;
>> @@ -1547,6 +1560,28 @@ static void tb_dump_switch(const struct tb *tb, const struct tb_switch *sw)
>>   	       regs->__unknown1, regs->__unknown4);
>>   }
>>   
>> +/**
>> + * tb_switch_reset_ports() - Reset downstream ports of switch.
> Drop the '.'
Ack.
>> + * @sw: Switch whose ports need to be reset.
>> + *
>> + * This is applicable only for USB4 routers.
>> + * tb_switch_is_usb4() needs to be called before calling this
>> + * function.
> This should fit into 2 lines:
>
> 	* This is applicable only for USB4 routers. tb_switch_is_usb4()
> 	* needs to be called before calling this function.
Ack.
>> + *
>> + * Return: Returns 0 on success or an error code on failure.
> Specifically returns %-EOPNOTSUPP if the router does not support this
> (e.g is not USB4 router).
>
>> + */
>> +int tb_switch_reset_ports(struct tb_switch *sw)
>> +{
>> +	struct tb_port *port;
>> +	int ret = -EOPNOTSUPP;
> Reverse christmas tree:
>
> 	int ret = -EOPNOTSUPP;
> 	struct tb_port *port;
Ack.
>> +
>> +	tb_switch_for_each_port(sw, port) {
>> +		if (tb_port_is_null(port) && port->cap_usb4)
>> +			return tb_port_reset(port);
> Now you run this only once for the first lane adapter it finds.
>
> How much you actually tested this patch series? :(

I've tested this patch on multiple TBT4 docks like OWC, Kenningston and 
Lenovo on
reboot/suspend-resume/hotplug scenarios.

> Since we are already in -rc5 it is unlikely that behavioral changes like
> this will go to the next release (v6.8-rc1), so you have all that time
> to make sure your patches work as expected.
Agreed. Since this is a behavioral change it needs more testing and 
refinement on how
we can reset the switch and its ports.
>> +	}
>> +	return ret;
>> +}
>> +
>>   /**
>>    * tb_switch_reset() - reconfigure route, enable and send TB_CFG_PKG_RESET
>>    * @sw: Switch to reset
>> diff --git a/drivers/thunderbolt/tb.h b/drivers/thunderbolt/tb.h
>> index e299e53473ae..f2687ec4ac53 100644
>> --- a/drivers/thunderbolt/tb.h
>> +++ b/drivers/thunderbolt/tb.h
>> @@ -797,6 +797,7 @@ void tb_switch_remove(struct tb_switch *sw);
>>   void tb_switch_suspend(struct tb_switch *sw, bool runtime);
>>   int tb_switch_resume(struct tb_switch *sw);
>>   int tb_switch_reset(struct tb_switch *sw);
>> +int tb_switch_reset_ports(struct tb_switch *sw);
>>   int tb_switch_wait_for_bit(struct tb_switch *sw, u32 offset, u32 bit,
>>   			   u32 value, int timeout_msec);
>>   void tb_sw_set_unplugged(struct tb_switch *sw);
>> @@ -1281,6 +1282,7 @@ struct tb_port *usb4_switch_map_usb3_down(struct tb_switch *sw,
>>   int usb4_switch_add_ports(struct tb_switch *sw);
>>   void usb4_switch_remove_ports(struct tb_switch *sw);
>>   
>> +int usb4_port_reset(struct tb_port *port);
>>   int usb4_port_unlock(struct tb_port *port);
>>   int usb4_port_hotplug_enable(struct tb_port *port);
>>   int usb4_port_configure(struct tb_port *port);
>> diff --git a/drivers/thunderbolt/tb_regs.h b/drivers/thunderbolt/tb_regs.h
>> index 87e4795275fe..d49530bc0d53 100644
>> --- a/drivers/thunderbolt/tb_regs.h
>> +++ b/drivers/thunderbolt/tb_regs.h
>> @@ -389,6 +389,7 @@ struct tb_regs_port_header {
>>   #define PORT_CS_18_CSA				BIT(22)
>>   #define PORT_CS_18_TIP				BIT(24)
>>   #define PORT_CS_19				0x13
>> +#define PORT_CS_19_DPR				BIT(0)
>>   #define PORT_CS_19_PC				BIT(3)
>>   #define PORT_CS_19_PID				BIT(4)
>>   #define PORT_CS_19_WOC				BIT(16)
>> diff --git a/drivers/thunderbolt/usb4.c b/drivers/thunderbolt/usb4.c
>> index 4277733d0021..c8a4bf33ed1c 100644
>> --- a/drivers/thunderbolt/usb4.c
>> +++ b/drivers/thunderbolt/usb4.c
>> @@ -1073,6 +1073,45 @@ void usb4_switch_remove_ports(struct tb_switch *sw)
>>   	}
>>   }
>>   
>> +/**
>> + * usb4_port_reset() - Reset USB4 downsteam port
>> + * @port: USB4 port to reset.
>> + *
>> + * Helps to reconfigure USB4 link by resetting downstream port.
>> + *
>> + * Return: Returns 0 on success or an error code on failure.
>> + */
>> +int usb4_port_reset(struct tb_port *port)
>> +{
>> +	u32 val = 0;
> This initialization is actually not needed.
Ack.
>
>> +	int ret;
>> +
>> +	ret = tb_port_read(port, &val, TB_CFG_PORT,
>> +			port->cap_usb4 + PORT_CS_19, 1);
>> +	if (ret)
>> +		return ret;
>> +
>> +	val = val | PORT_CS_19_DPR;
> val |= PORT_CS_19_DPR;
>
> Same as you do below with &= ~..
Ack.
>
>> +	ret = tb_port_write(port, &val, TB_CFG_PORT,
>> +			port->cap_usb4 + PORT_CS_19, 1);
>> +	if (ret)
>> +		return ret;
>> +
>> +	/* Wait for 10ms after requesting downstream port reset */
>> +	usleep_range(10000, 15000);
>> +
>> +	ret = tb_port_read(port, &val, TB_CFG_PORT,
>> +			port->cap_usb4 + PORT_CS_19, 1);
>> +	if (ret)
>> +		return ret;
> Do you really need to read it back from the hardware here?

I don't think it's necessary. But I added this just to be safe if 
performing DPR has
made changes to other CS_19 bits.

>> +
>> +	val &= ~PORT_CS_19_DPR;
>> +	ret = tb_port_write(port, &val, TB_CFG_PORT,
>> +			port->cap_usb4 + PORT_CS_19, 1);
>> +
>> +	return ret;
> This can be simply
>
> 	return tb_port_write(port, &val, TB_CFG_PORT,
> 			     port->cap_usb4 + PORT_CS_19, 1);
Ack.
>> +}
>> +
>>   /**
>>    * usb4_port_unlock() - Unlock USB4 downstream port
>>    * @port: USB4 port to unlock
>> -- 
>> 2.34.1

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ