Devicetree
 help / color / mirror / Atom feed
From: sashiko-bot@kernel.org
To: "Christian Marangi" <ansuelsmth@gmail.com>, devicetree@vger.kernel.org
Cc: robh@kernel.org, conor+dt@kernel.org
Subject: Re: [PATCH net-next v20 07/15] net: phylink: add capabilities helper and track modified advertising
Date: Sun, 15 Nov 2026 10:45:16 +0000	[thread overview]
Message-ID: <sashiko-outbox-215398@kernel.org> (raw)
In-Reply-To: <20261114104426.235620-8-ansuelsmth@gmail.com>

Thank you for your contribution! Sashiko AI review found 2 potential
issues to consider:
- [High] net: phylink: unchecked phylink_validate() in
  phylink_calculate_capabilities()
- [High] net: phylink: broken advertising tracking in
  phylink_calculate_capabilities()
--

commit fb3c4a8238a09c214077ba885092205d6fc987a8
Author: Christian Marangi <ansuelsmth@gmail.com>

net: phylink: add capabilities helper and track modified advertising

This commit introduces phylink_calculate_capabilities() and a new
advertising mask in struct phylink to recalculate supported and advertised
link modes while preserving user-disabled advertisements.

> diff --git a/drivers/net/phy/phylink.c b/drivers/net/phy/phylink.c
> index d1bd7109c5668..46b2aba336bd1 100644
> --- a/drivers/net/phy/phylink.c
> +++ b/drivers/net/phy/phylink.c
> @@ -61,6 +61,7 @@ struct phylink {
>  	u8 link_port;			/* The current non-phy ethtool port */
>  	__ETHTOOL_DECLARE_LINK_MODE_MASK(supported);
>  	__ETHTOOL_DECLARE_LINK_MODE_MASK(supported_lpi);
> +	__ETHTOOL_DECLARE_LINK_MODE_MASK(advertising);
>  
>  	/* The link configuration settings */
>  	struct phylink_link_state link_config;
> @@ -717,6 +718,27 @@ static int phylink_validate(struct phylink *pl, unsigned long *supported,
>  	return ret;
>  }
>  
> +static void phylink_calculate_capabilities(struct phylink *pl)
> +{
> +	__ETHTOOL_DECLARE_LINK_MODE_MASK(disabled_advertising);
> +
> +	/* Get user-disabled advertising modes */
> +	linkmode_xor(disabled_advertising, pl->advertising,
> +		     pl->link_config.advertising);
> +
> +	/* Calculate the full supported advertisement */
> +	linkmode_fill(pl->supported);
> +	linkmode_copy(pl->link_config.advertising, pl->supported);
> +	phylink_validate(pl, pl->supported, &pl->link_config);
> +
> +	/* Save the newly calculated full advertisement */
> +	linkmode_copy(pl->advertising, pl->link_config.advertising);

[Severity: High]
What happens to pl->supported, pl->advertising, and
pl->link_config.advertising when phylink_validate() fails?

When the PCS supporting pl->link_config.interface has not attached yet or
has just been released, __phylink_validate() fails the
test_bit(state->interface, interfaces) check and returns -EINVAL before
masking supported or state->advertising.  Because the return value of
phylink_validate() is ignored, pl->supported, pl->advertising, and
pl->link_config.advertising remain filled with all ones, corrupting both
the active capabilities and the pl->advertising baseline used by later
phylink_calculate_capabilities() calls.

> +
> +	/* Keep user-disabled modes out of the active advertisement. */
> +	linkmode_andnot(pl->link_config.advertising,
> +			pl->link_config.advertising, disabled_advertising);
> +}

[Severity: High]
Does phylink_calculate_capabilities() preserve user-disabled modes and
mode-specific capabilities across PCS release and re-attach?

First, disabled_advertising is computed from pl->advertising and
pl->link_config.advertising on each call, and then pl->advertising is
overwritten with the new capabilities.  When a PCS is removed, its modes
are cleared from both masks; when it is later re-attached, linkmode_xor()
sees 0 ^ 0 = 0 for those modes and re-enables any modes the user had
disabled before removal.

Second, because linkmode_xor() is symmetric rather than
linkmode_andnot(disabled_advertising, pl->advertising,
pl->link_config.advertising), any bit present in
pl->link_config.advertising but absent from pl->advertising (such as pause
bits updated by phylink_update_pause_state()) is placed into
disabled_advertising and stripped.

Third, phylink_create() calls phylink_calculate_capabilities() before
phylink_parse_mode() and phylink_parse_fixedlink(), which (along with
phylink_set_fixed_link(), phylink_bringup_phy(), and
phylink_sfp_set_config()) overwrite pl->supported and
pl->link_config.advertising without updating pl->advertising.  A later
phylink_calculate_capabilities() call then treats the difference between
the stale initial pl->advertising and the mode-specific
pl->link_config.advertising as user-disabled modes, while
linkmode_fill(pl->supported) clobbers the mode-specific pl->supported
mask.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261114104426.235620-1-ansuelsmth@gmail.com?part=7

  reply	other threads:[~2026-11-15 10:45 UTC|newest]

Thread overview: 28+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-11-14 10:43 [PATCH net-next v20 00/15] net: pcs: Introduce support for fwnode PCS Christian Marangi
2026-11-14 10:43 ` [PATCH net-next v20 01/15] net: phylink: keep and use MAC supported_interfaces in phylink struct Christian Marangi
2026-11-15 10:45   ` sashiko-bot
2026-11-14 10:43 ` [PATCH net-next v20 02/15] net: phylink: make PCS phylink instance access RCU protected Christian Marangi
2026-11-15 10:45   ` sashiko-bot
2026-11-14 10:43 ` [PATCH net-next v20 03/15] net: phylink: introduce internal phylink PCS handling Christian Marangi
2026-11-15 10:45   ` sashiko-bot
2026-11-14 10:43 ` [PATCH net-next v20 04/15] net: pcs: implement Firmware node support for PCS driver Christian Marangi
2026-11-15 10:45   ` sashiko-bot
2026-11-14 10:43 ` [PATCH net-next v20 05/15] net: phylink: save phylink instance fwnode on phylink_create Christian Marangi
2026-11-14 10:43 ` [PATCH net-next v20 06/15] net: phylink: make phylink_validate state locked Christian Marangi
2026-11-15 10:45   ` sashiko-bot
2026-11-14 10:43 ` [PATCH net-next v20 07/15] net: phylink: add capabilities helper and track modified advertising Christian Marangi
2026-11-15 10:45   ` sashiko-bot [this message]
2026-11-14 10:43 ` [PATCH net-next v20 08/15] net: phylink: support PCS provider release Christian Marangi
2026-11-15 10:45   ` sashiko-bot
2026-11-14 10:43 ` [PATCH net-next v20 09/15] net: phylink: support late PCS provider attach Christian Marangi
2026-11-15 10:45   ` sashiko-bot
2026-11-14 10:43 ` [PATCH net-next v20 10/15] net: Document PCS subsystem Christian Marangi
2026-11-14 10:43 ` [PATCH net-next v20 11/15] MAINTAINERS: add myself as PCS subsystem maintainer Christian Marangi
2026-11-14 10:43 ` [PATCH net-next v20 12/15] net: phylink: add .pcs_link_down PCS OP Christian Marangi
2026-11-14 10:43 ` [PATCH net-next v20 13/15] dt-bindings: net: pcs: Document support for Airoha Ethernet PCS Christian Marangi
2026-11-15 10:45   ` sashiko-bot
2026-11-14 10:43 ` [PATCH net-next v20 14/15] net: pcs: airoha: add PCS driver for Airoha AN7581 SoC Christian Marangi
2026-11-15 10:45   ` sashiko-bot
2026-11-14 10:43 ` [PATCH net-next v20 15/15] net: airoha: add phylink support Christian Marangi
2026-11-15 10:45   ` sashiko-bot
2026-11-14 13:27 ` [PATCH net-next v20 00/15] net: pcs: Introduce support for fwnode PCS Frank Wunderlich

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=sashiko-outbox-215398@kernel.org \
    --to=sashiko-bot@kernel.org \
    --cc=ansuelsmth@gmail.com \
    --cc=conor+dt@kernel.org \
    --cc=devicetree@vger.kernel.org \
    --cc=robh@kernel.org \
    --cc=sashiko-reviews@lists.linux.dev \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox