On 11/17/26 20:58, [email protected] wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential
> issue to consider:
> - [Medium] phy: amlogic: broken rate and submode check in
>   meson_hdmi_phy_validate()
> --
> 
> commit 0bf49ea77c864cdcf7af96e59e1ffba5194d9636
> Author: Ronald Claveau <[email protected]>
> 
> phy: amlogic: Add Meson HDMI TX PHY driver
> 
> This commit adds a generic PHY driver for the HDMI TX PHY on Amlogic GXBB,
> GXL, GXM, and G12A SoCs, along with a legacy helper for device trees
> lacking a "phys" property.
> 
>> diff --git a/drivers/phy/amlogic/phy-meson-hdmi-tx.c 
>> b/drivers/phy/amlogic/phy-meson-hdmi-tx.c
>> new file mode 100644
>> index 0000000000000..69833a080dc35
>> --- /dev/null
>> +++ b/drivers/phy/amlogic/phy-meson-hdmi-tx.c
> [ ... ]
>> +/* GXBB */
>> +static const struct meson_hdmi_phy_rate gxbb_rates[] = {
>> +    /* 5.94Gbps, 3.7125Gbps */
>> +    { 371250000, 0x33353245, 0x2100115b },
>> +    /* 2.97Gbps */
>> +    { 297000000, 0x33634283, 0xb000115b },
>> +    /* 1.485Gbps, and below */
>> +    { 0,         0x33632122, 0x2000115b },
>> +};
> [ ... ]
>> +static const struct meson_hdmi_phy_rate *
>> +meson_hdmi_phy_find_rate(const struct meson_hdmi_phy_data *data, u64 rate)
>> +{
>> +    unsigned int i;
>> +
>> +    for (i = 0; i < data->num_rates; i++)
>> +            if (rate >= data->rates[i].min_rate)
>> +                    return &data->rates[i];
>> +
>> +    return NULL;
>> +}
> [ ... ]
>> +static int meson_hdmi_phy_validate(struct phy *phy, enum phy_mode mode,
>> +                               int submode, union phy_configure_opts *opts)
>> +{
>> +    struct meson_hdmi_phy *priv = phy_get_drvdata(phy);
>> +
>> +    if (mode != PHY_MODE_HDMI)
>> +            return -EINVAL;
>> +
>> +    if (opts->hdmi.bpc != 8)
>> +            return -EINVAL;
>> +
>> +    if (!meson_hdmi_phy_find_rate(priv->data, opts->hdmi.tmds_char_rate))
>> +            return -EINVAL;
>> +
>> +    return 0;
>> +}
> 
> [Severity: Medium]
> Can meson_hdmi_phy_find_rate() ever return NULL here?
> 
> Every per-SoC rate table ends with an entry where min_rate is 0, and rate
> is an unsigned u64, so rate >= data->rates[i].min_rate is always true on
> the final entry.  That makes the NULL return and the callers' NULL checks
> in meson_hdmi_phy_validate() and meson_hdmi_phy_configure() dead code, and
> meson_hdmi_phy_validate() accepts any tmds_char_rate value (including 0 Hz
> or rates above the 594 MHz hardware limit).
> 
> In addition, meson_hdmi_phy_validate() does not check that submode is
> PHY_HDMI_MODE_TMDS before reading opts->hdmi.tmds_char_rate, which shares
> an anonymous union with opts->hdmi.frl when submode is PHY_HDMI_MODE_FRL.
> 

I will add min_rate 25175000 in table end and a check on max_rate 600MHz
in v2.

-- 
Best regards,
Ronald

Reply via email to