Re: [PATCH 1/1] [brcmfmac] Fix regulatory domain handling to reset bands properly

Re: [PATCH 1/1] [brcmfmac] Fix regulatory domain handling to reset bands properly

[Daniel Berlin]

As an aside, the alternative would be to simply not ignore any
attempts to set the regulatory domain, regardless of whether it's 00
or the chip is already set to that country.
It doesn't happen that often, so it's not clear it's worth avoiding it at all.
There are some things i'd have to fix to make it work resiliently well
(for example, i know we set the 2g bw cap where we do because it has
to be done with the interface down), but i can fix those if needed.

On Thu, Oct 19, 2023 at 10:18 AM Daniel Berlin <dberlin@dberlin.org> wrote:
>
> Currently, we ignore the default country in the reg notifier.
> We also register a custom regulatory domain, which is set
> as the default.
> As a result, the chip is likely to be set to the correct country,
> but the regulatory domain will not match it.
>
> When the regulatory notifier is then called, we see the countries
> are the same and do not change anything, even though the domain
> is wrong.
>
> This patch forces us to reset the bands on the first country change
> even if the chip is already set to that country.
>
> We also restore the original band info before reconstructing channel
> info, as the new regdom power limits may be higher than what is
> currently set.
>
> Signed-off-by: Daniel Berlin <dberlin@dberlin.org>
> ---
>  .../broadcom/brcm80211/brcmfmac/cfg80211.c    | 38 +++++++++++++++----
>  .../broadcom/brcm80211/brcmfmac/cfg80211.h    |  2 +
>  2 files changed, 33 insertions(+), 7 deletions(-)
>
> diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
> index 3656790ec4c9..7bc479ccc24b 100644
> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.c
> @@ -7199,11 +7199,23 @@ static int brcmf_construct_chaninfo(struct brcmf_cfg80211_info *cfg,
>                 goto fail_pbuf;
>         }
>
> +       /* Changing regulatory domain may change channels and limits
> +        * To ensure that we correctly set the new band info, copy the original
> +        * info first.
> +        */
>         band = wiphy->bands[NL80211_BAND_2GHZ];
> -       if (band)
> +       if (band) {
> +               memcpy(band->channels, &__wl_2ghz_channels,
> +                      sizeof(__wl_2ghz_channels));
> +               band->n_channels = ARRAY_SIZE(__wl_2ghz_channels);
>                 for (i = 0; i < band->n_channels; i++)
>                         band->channels[i].flags = IEEE80211_CHAN_DISABLED;
> +       }
>         band = wiphy->bands[NL80211_BAND_5GHZ];
> -       if (band)
> +       if (band) {
> +               memcpy(band->channels, &__wl_5ghz_channels,
> +                      sizeof(__wl_5ghz_channels));
> +               band->n_channels = ARRAY_SIZE(__wl_5ghz_channels);
>                 for (i = 0; i < band->n_channels; i++)
>                         band->channels[i].flags = IEEE80211_CHAN_DISABLED;
> +       }
> @@ -7210,8 +7222,11 @@
>         band = wiphy->bands[NL80211_BAND_6GHZ];
> -       if (band)
> +       if (band) {
> +               memcpy(band->channels, &__wl_6ghz_channels,
> +                      sizeof(__wl_6ghz_channels));
> +               band->n_channels = ARRAY_SIZE(__wl_6ghz_channels);
>                 for (i = 0; i < band->n_channels; i++)
>                         band->channels[i].flags = IEEE80211_CHAN_DISABLED;
> -
> +       }
>         total = le32_to_cpu(list->count);
>         if (total > BRCMF_MAX_CHANSPEC_LIST) {
>                 bphy_err(drvr, "Invalid count of channel Spec. (%u)\n",
> @@ -8601,9 +8616,17 @@ static void brcmf_cfg80211_reg_notifier(struct wiphy *wiphy,
>         }
>
>         err = brcmf_translate_country_code(ifp->drvr, req->alpha2, &ccreq);
> -       if (err)
> -               return;
> -
> +       if (err) {
> +               /* Because we ignore the default country code above,
> +                * we will start out in our custom reg domain, but the chip
> +                * may already be set to the right country.
> +                * As such, we force the bands to be re-set the first
> +                * time we try to set a country for real.
> +                */
> +               if (err != -EAGAIN || !cfg->force_band_setup)
> +                       return;
> +       }
> +       cfg->force_band_setup = false;
>         err = brcmf_fil_iovar_data_set(ifp, "country", &ccreq, sizeof(ccreq));
>         if (err) {
>                 bphy_err(drvr, "Firmware rejected country setting\n");
> @@ -8670,6 +8693,7 @@ struct brcmf_cfg80211_info *brcmf_cfg80211_attach(struct brcmf_pub *drvr,
>         cfg->pub = drvr;
>         init_vif_event(&cfg->vif_event);
>         INIT_LIST_HEAD(&cfg->vif_list);
> +       cfg->force_band_setup = true;
>
>         vif = brcmf_alloc_vif(cfg, NL80211_IFTYPE_STATION);
>         if (IS_ERR(vif))
> diff --git a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.h b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.h
> index 0e1fa3f0dea2..7e60ceeeeb3a 100644
> --- a/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.h
> +++ b/drivers/net/wireless/broadcom/brcm80211/brcmfmac/cfg80211.h
> @@ -327,6 +327,7 @@ struct brcmf_cfg80211_wowl {
>   * @dongle_up: indicate whether dongle up or not.
>   * @roam_on: on/off switch for dongle self-roaming.
>   * @scan_tried: indicates if first scan attempted.
> + * @force_band_setup: indicates if we should force band setup
>   * @dcmd_buf: dcmd buffer.
>   * @extra_buf: mainly to grab assoc information.
>   * @debugfsdir: debugfs folder for this device.
> @@ -357,6 +358,7 @@ struct brcmf_cfg80211_info {
>         bool pwr_save;
>         bool dongle_up;
>         bool scan_tried;
> +       bool force_band_setup;
>         u8 *dcmd_buf;
>         u8 *extra_buf;
>         struct dentry *debugfsdir;
> --
> 2.41.0
>

Re: [PATCH 1/1] [brcmfmac] Fix regulatory domain handling to reset bands properly

[Arend van Spriel]

[-- Attachment #1: Type: text/plain, Size: 2187 bytes --]

On 10/21/2023 4:40 PM, Daniel Berlin wrote:
> As an aside, the alternative would be to simply not ignore any
> attempts to set the regulatory domain, regardless of whether it's 00
> or the chip is already set to that country.
> It doesn't happen that often, so it's not clear it's worth avoiding it at all.
> There are some things i'd have to fix to make it work resiliently well
> (for example, i know we set the 2g bw cap where we do because it has
> to be done with the interface down), but i can fix those if needed.

Please do not top post.

The firmware has its own regulatory information (actually the CLM blob 
holds that information) which is independent of the wireless regulatory 
database in the kernel. As such trying to configure country '00' will 
simply fail as it is not known inside CLM blob. The firmware needs a 
country abbreviation *and* a revision. Hence we require a mapping of 
country code to <abbrev,rev> tuple. If that mapping is not in place we 
bail out setting the country in the regulatory notifier.

 From the description below it is not clear what problem is fixed here 
in terms of user experience.

Regards,
Arend

> On Thu, Oct 19, 2023 at 10:18 AM Daniel Berlin <dberlin@dberlin.org> wrote:
>>
>> Currently, we ignore the default country in the reg notifier.
>> We also register a custom regulatory domain, which is set
>> as the default.
>> As a result, the chip is likely to be set to the correct country,
>> but the regulatory domain will not match it.
>>
>> When the regulatory notifier is then called, we see the countries
>> are the same and do not change anything, even though the domain
>> is wrong.
>>
>> This patch forces us to reset the bands on the first country change
>> even if the chip is already set to that country.
>>
>> We also restore the original band info before reconstructing channel
>> info, as the new regdom power limits may be higher than what is
>> currently set.
>>
>> Signed-off-by: Daniel Berlin <dberlin@dberlin.org>
>> ---
>>   .../broadcom/brcm80211/brcmfmac/cfg80211.c    | 38 +++++++++++++++----
>>   .../broadcom/brcm80211/brcmfmac/cfg80211.h    |  2 +
>>   2 files changed, 33 insertions(+), 7 deletions(-)

[-- Attachment #2: S/MIME Cryptographic Signature --]
[-- Type: application/pkcs7-signature, Size: 4219 bytes --]

Re: [PATCH 1/1] [brcmfmac] Fix regulatory domain handling to reset bands properly

[Daniel Berlin]

On Mon, Oct 23, 2023 at 4:38 AM Arend van Spriel
<arend.vanspriel@broadcom.com> wrote:
>
> On 10/21/2023 4:40 PM, Daniel Berlin wrote:
> > As an aside, the alternative would be to simply not ignore any
> > attempts to set the regulatory domain, regardless of whether it's 00
> > or the chip is already set to that country.
> > It doesn't happen that often, so it's not clear it's worth avoiding it at all.
> > There are some things i'd have to fix to make it work resiliently well
> > (for example, i know we set the 2g bw cap where we do because it has
> > to be done with the interface down), but i can fix those if needed.
>
> Please do not top post.
>
> The firmware has its own regulatory information (actually the CLM blob
> holds that information) which is independent of the wireless regulatory
> database in the kernel. As such trying to configure country '00' will
> simply fail as it is not known inside CLM blob.
>
> The firmware needs a
> country abbreviation *and* a revision. Hence we require a mapping of
> country code to <abbrev,rev> tuple. If that mapping is not in place we
> bail out setting the country in the regulatory notifier.
>
>  From the description below it is not clear what problem is fixed here
> in terms of user experience.


As far as the stack and kernel are concerned, it gets stuck in the
custom regulatory domain ('99') and sets the max power levels,
bandwidth caps, etc in the capabilities to match.

If you look at the phy info, it will clearly show you this.
You are correct that the chip maintains its own view of power levels
(though most drivers set them, we do not), but the bandwidth caps are
respected, for example.
Note also that '00' is recognized by at least this firmware, though I
agree it depends on whether/what is done in the CLM blob.

All that said, even if it changed nothing in the chip itself, it would
still be displaying completely wrong info to the user.
I'm happy to update the description if that is your main concern.


>
> Regards,
> Arend
>
> > On Thu, Oct 19, 2023 at 10:18 AM Daniel Berlin <dberlin@dberlin.org> wrote:
> >>
> >> Currently, we ignore the default country in the reg notifier.
> >> We also register a custom regulatory domain, which is set
> >> as the default.
> >> As a result, the chip is likely to be set to the correct country,
> >> but the regulatory domain will not match it.
> >>
> >> When the regulatory notifier is then called, we see the countries
> >> are the same and do not change anything, even though the domain
> >> is wrong.
> >>
> >> This patch forces us to reset the bands on the first country change
> >> even if the chip is already set to that country.
> >>
> >> We also restore the original band info before reconstructing channel
> >> info, as the new regdom power limits may be higher than what is
> >> currently set.
> >>
> >> Signed-off-by: Daniel Berlin <dberlin@dberlin.org>
> >> ---
> >>   .../broadcom/brcm80211/brcmfmac/cfg80211.c    | 38 +++++++++++++++----
> >>   .../broadcom/brcm80211/brcmfmac/cfg80211.h    |  2 +
> >>   2 files changed, 33 insertions(+), 7 deletions(-)