netdev - Re: [PATCH net-next 14/15] net: sparx5: add compatible strings for lan969x and verify the target

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]
Message-ID: <24b18f9e-6fbd-48cc-96bd-e634d0af9824@kernel.org>
Date: Wed, 23 Oct 2024 14:06:33 +0200
From: Krzysztof Kozlowski <krzk@...nel.org>
To: Daniel Machon <daniel.machon@...rochip.com>
Cc: "David S. Miller" <davem@...emloft.net>,
 Eric Dumazet <edumazet@...gle.com>, Jakub Kicinski <kuba@...nel.org>,
 Paolo Abeni <pabeni@...hat.com>, andrew@...n.ch,
 Lars Povlsen <lars.povlsen@...rochip.com>,
 Steen Hegelund <Steen.Hegelund@...rochip.com>, horatiu.vultur@...rochip.com,
 jensemil.schulzostergaard@...rochip.com,
 Parthiban.Veerasooran@...rochip.com, Raju.Lakkaraju@...rochip.com,
 UNGLinuxDriver@...rochip.com, Richard Cochran <richardcochran@...il.com>,
 Rob Herring <robh@...nel.org>, Krzysztof Kozlowski <krzk+dt@...nel.org>,
 Conor Dooley <conor+dt@...nel.org>, jacob.e.keller@...el.com,
 ast@...erby.net, maxime.chevallier@...tlin.com, netdev@...r.kernel.org,
 linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
 devicetree@...r.kernel.org
Subject: Re: [PATCH net-next 14/15] net: sparx5: add compatible strings for
 lan969x and verify the target

On 23/10/2024 13:00, Daniel Machon wrote:
> Hi Krzysztof,
> 
>>> Add compatible strings for the twelve lan969x SKU's (Stock Keeping Unit)
>>> that we support, and verify that the devicetree target is supported by
>>> the chip target.
>>>
>>> Each SKU supports different bandwidths and features (see [1] for
>>> details). We want to be able to run a SKU with a lower bandwidth and/or
>>> feature set, than what is supported by the actual chip. In order to
>>> accomplish this we:
>>>
>>>     - add new field sparx5->target_dt that reflects the target from the
>>>       devicetree (compatible string).
>>>
>>>     - compare the devicetree target with the actual chip target. If the
>>>       bandwidth and features provided by the devicetree target is
>>>       supported by the chip, we approve - otherwise reject.
>>>
>>>     - set the core clock and features based on the devicetree target
>>>
>>> [1] https://www.microchip.com/en-us/product/lan9698
>>>
>>> Reviewed-by: Steen Hegelund <Steen.Hegelund@...rochip.com>
>>> Signed-off-by: Daniel Machon <daniel.machon@...rochip.com>
>>> ---
>>>  drivers/net/ethernet/microchip/sparx5/Makefile     |   1 +
>>>  .../net/ethernet/microchip/sparx5/sparx5_main.c    | 194 ++++++++++++++++++++-
>>>  .../net/ethernet/microchip/sparx5/sparx5_main.h    |   1 +
>>>  3 files changed, 193 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/drivers/net/ethernet/microchip/sparx5/Makefile b/drivers/net/ethernet/microchip/sparx5/Makefile
>>> index 3435ca86dd70..8fe302415563 100644
>>> --- a/drivers/net/ethernet/microchip/sparx5/Makefile
>>> +++ b/drivers/net/ethernet/microchip/sparx5/Makefile
>>> @@ -19,3 +19,4 @@ sparx5-switch-$(CONFIG_DEBUG_FS) += sparx5_vcap_debugfs.o
>>>  # Provide include files
>>>  ccflags-y += -I$(srctree)/drivers/net/ethernet/microchip/vcap
>>>  ccflags-y += -I$(srctree)/drivers/net/ethernet/microchip/fdma
>>> +ccflags-y += -I$(srctree)/drivers/net/ethernet/microchip/lan969x
>>> diff --git a/drivers/net/ethernet/microchip/sparx5/sparx5_main.c b/drivers/net/ethernet/microchip/sparx5/sparx5_main.c
>>> index 5c986c373b3e..edbe639d98c5 100644
>>> --- a/drivers/net/ethernet/microchip/sparx5/sparx5_main.c
>>> +++ b/drivers/net/ethernet/microchip/sparx5/sparx5_main.c
>>> @@ -24,6 +24,8 @@
>>>  #include <linux/types.h>
>>>  #include <linux/reset.h>
>>>
>>> +#include "lan969x.h" /* lan969x_desc */
>>> +
>>>  #include "sparx5_main_regs.h"
>>>  #include "sparx5_main.h"
>>>  #include "sparx5_port.h"
>>> @@ -227,6 +229,168 @@ bool is_sparx5(struct sparx5 *sparx5)
>>>       }
>>>  }
>>>
>>> +/* Set the devicetree target based on the compatible string */
>>> +static int sparx5_set_target_dt(struct sparx5 *sparx5)
>>> +{
>>> +     struct device_node *node = sparx5->pdev->dev.of_node;
>>> +
>>> +     if (is_sparx5(sparx5))
>>> +             /* For Sparx5 the devicetree target is always the chip target */
>>> +             sparx5->target_dt = sparx5->target_ct;
>>> +     else if (of_device_is_compatible(node, "microchip,lan9691-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9691VAO;
>>> +     else if (of_device_is_compatible(node, "microchip,lan9692-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9692VAO;
>>> +     else if (of_device_is_compatible(node, "microchip,lan9693-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9693VAO;
>>> +     else if (of_device_is_compatible(node, "microchip,lan9694-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9694;
>>> +     else if (of_device_is_compatible(node, "microchip,lan9695-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9694TSN;
>>> +     else if (of_device_is_compatible(node, "microchip,lan9696-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9696;
>>> +     else if (of_device_is_compatible(node, "microchip,lan9697-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9696TSN;
>>> +     else if (of_device_is_compatible(node, "microchip,lan9698-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9698;
>>> +     else if (of_device_is_compatible(node, "microchip,lan9699-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9698TSN;
>>> +     else if (of_device_is_compatible(node, "microchip,lan969a-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9694RED;
>>> +     else if (of_device_is_compatible(node, "microchip,lan969b-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9696RED;
>>> +     else if (of_device_is_compatible(node, "microchip,lan969c-switch"))
>>> +             sparx5->target_dt = SPX5_TARGET_CT_LAN9698RED;
>>> +     else
>>> +             return -EINVAL;
>>> +
>>> +     return 0;
>>> +}
>>> +
>>> +/* Compare the devicetree target with the chip target.
>>> + * Make sure the chip target supports the features and bandwidth requested
>>> + * from the devicetree target.
>>> + */
>>> +static int sparx5_verify_target(struct sparx5 *sparx5)
>>> +{
>>> +     switch (sparx5->target_dt) {
>>> +     case SPX5_TARGET_CT_7546:
>>> +     case SPX5_TARGET_CT_7549:
>>> +     case SPX5_TARGET_CT_7552:
>>> +     case SPX5_TARGET_CT_7556:
>>> +     case SPX5_TARGET_CT_7558:
>>> +     case SPX5_TARGET_CT_7546TSN:
>>> +     case SPX5_TARGET_CT_7549TSN:
>>> +     case SPX5_TARGET_CT_7552TSN:
>>> +     case SPX5_TARGET_CT_7556TSN:
>>> +     case SPX5_TARGET_CT_7558TSN:
>>> +             return 0;
>>
>> All this is weird. Why would you verify? You were matched, it cannot be
>> mis-matching.
> 
> We are verifying that the match (target/compatible string) from the
> device tree is supported by the chip. Maybe I wasn't too clear about the
> intend in v1.
> 
> Each target supports different bandwidths and features. If you have a
> lan9698 chip, it must, obviously, be possible to run it as a lan9698
> target. However, some targets can be run on chip targets other than
> themselves, given that the chip supports the bandwidth and features of
> the provided target. In contrary, trying to run as a target with a
> feature not supported by the chip, or a bandwidth higher than what the
> chip supports, should be rejected.

But you are not supposed to compare DT with what you auto-detected.
Detect your hardware, test if it is supported and then bail out.

None of above explains the code.

> 
> Without this logic, the chip id is read and a target is determined. That
> means on a lan9698 chip you will always match the lan9698 target.

That's not the job of kernel.

> 
> With the new logic, it is possible to run as a different target than
> what is read from the chip id, given that the target you are trying to
> run as, is supported by the chip.

So just run on different target.

> 
>>
>>> +     case SPX5_TARGET_CT_LAN9698RED:
>>> +             if (sparx5->target_ct == SPX5_TARGET_CT_LAN9698RED)
>>
>> What is "ct"? sorry, all this code is a big no.
> 
> In this case we were matched as a SPX5_TARGET_CT_LAN9698RED target. We
> are verifying that the chip target (target_ct, which is read from the
> chip) supports the target we were matched as.
> 
>> Krzysztof
>>
> 
> This is a feature that we would like, as it gives the flexibility of
> running different targets on the same chip. Now if this is something
> that cannot be accepted, I will have to ditch this part.

I have no clue what the "target" is but so far it looks like you try to
validate DT against detected device. That's not how it should work.

Best regards,
Krzysztof