netdev - Re: [PATCH net-next v4 13/12] net: lan865x: optional hardware reset

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: Tue, 30 Apr 2024 10:04:50 +0200
From: Ramón Nordin Rodriguez <ramon.nordin.rodriguez@...roamp.se>
To: Andrew Lunn <andrew@...n.ch>
Cc: Parthiban.Veerasooran@...rochip.com, davem@...emloft.net,
	edumazet@...gle.com, kuba@...nel.org, pabeni@...hat.com,
	horms@...nel.org, saeedm@...dia.com, anthony.l.nguyen@...el.com,
	netdev@...r.kernel.org, linux-kernel@...r.kernel.org,
	corbet@....net, linux-doc@...r.kernel.org, robh+dt@...nel.org,
	krzysztof.kozlowski+dt@...aro.org, conor+dt@...nel.org,
	devicetree@...r.kernel.org, Horatiu.Vultur@...rochip.com,
	ruanjinjie@...wei.com, Steen.Hegelund@...rochip.com,
	vladimir.oltean@....com, UNGLinuxDriver@...rochip.com,
	Thorsten.Kummermehr@...rochip.com, Pier.Beruto@...emi.com,
	Selvamani.Rajagopal@...emi.com, Nicolas.Ferre@...rochip.com,
	benjamin.bigler@...nformulastudent.ch
Subject: Re: [PATCH net-next v4 13/12] net: lan865x: optional hardware reset

> > Additionally I figured out why my setup did not work without the HW
> > reset, I had missed a pull resistor in the schematic that held the IC in
> > reset.
> 
> Having a reset controlled by software is a pretty common
> design. Something needs to ensure the device is out of reset. It could
> be the bootloader, but i don't particularly like that, hiding away
> critical things where they are hard to see. So i think having it in
> the Linux driver is better.

Wholeheartedly agree, beyond the basics of bringing up ram, cores etc.
it becomes really weird when/if the bootloaders behaviour defines kernel
functionality.

In this case the oa_tc6 module does a soft reset and waits for a status
reg to signal ready.

What seems to be missing here is that the chip signals ready/out of
reset by asserting the irq pin and setting the RESETC bit of ther OA_STATUS0
reg (which is defined behaviour in the OA spec as I understand it).

Neither the lan865x_probe or oa_tc6_init checks for the initial
condition, but I'm guessing it's a fair assumption that the chip is out
of reset by the point when the oa_tc6_sw_reset_machphy function is
invoked.

Far as I can tell no timing information is given in the datasheet.

Might be unecessary but the setup could be made more explicit/clear
with a func such as:

int oa_tc6_out_of_reset(struct oa_tc*);

Which should be invoked before any reg access/modification code in
either oa_tc6 or the mac driver.
If this fails my (opinionated) preferred style would be to do one hw
reset, recheck, and on subsequent failure bail.

Such a change would probably lead to the HW reset being invoked on
reboots (if there is enough capacitors to keep the IC powered) and
definetly as a result of kexec calls.

> 
> There is an open question of does the driver need to actually reset
> the device, or is it sufficient to ensure it is out of reset? The
> wording of the standard suggests a hardware reset cycle is probably
> not required, but why did Microchip provide a reset pin?
> 

This has me tripped up. The lan8650/1 has a configuration write
protection mechanism with an unlock sequence, described in "4.6.3
Configuration Protection" of the datasheet.
The unlocking can be bypassed/simplified with a HW reset, still does
not seem like an explanation for the functionality.

I can't define one scenario where the HW reset is definetly necessary,
but will probably do it anyways on the systems I work on.

Ramón