public inbox for linux-kernel@vger.kernel.org
 help / color / mirror / Atom feed
From: Ping-Ke Shih <pkshih@realtek.com>
To: "Ondřej Jirman" <megi@xff.cz>,
	"linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>
Cc: "open list:REALTEK WIRELESS DRIVER (rtw89)"
	<linux-wireless@vger.kernel.org>
Subject: RE: [PATCH] wifi: rtw89: Fix inadverent sharing of struct ieee80211_supported_band data
Date: Mon, 28 Apr 2025 01:53:44 +0000	[thread overview]
Message-ID: <d3c6e149a2794551ba4570bdd7f1b7e7@realtek.com> (raw)
In-Reply-To: <20250427002414.410791-1-megi@xff.cz>

Ondřej Jirman <megi@xff.cz> wrote:
> 
> Internally wiphy writes to individual channels in this structure,
> so we must not share one static definition of channel list between
> multiple device instances, because that causes hard to debug
> breakage.
> 
> For example, with two rtw89 driven devices in the system, channel
> information may get incoherent, preventing channel use.
> 
> Signed-off-by: Ondrej Jirman <megi@xff.cz>
> ---
> 
> This patch relates to this report of mine:
> 
> 
> https://lore.kernel.org/linux-wireless/2goskmst4na36v42p2bs47uernp6kh3gzpadhr3u3r2yvyoxlg@bfprgq2qae7p
> /T/#u
> 
>  drivers/net/wireless/realtek/rtw89/core.c | 48 ++++++++++++++++++-----
>  1 file changed, 38 insertions(+), 10 deletions(-)
> 
> diff --git a/drivers/net/wireless/realtek/rtw89/core.c b/drivers/net/wireless/realtek/rtw89/core.c
> index cc9b014457ac..ae22954f5f5c 100644
> --- a/drivers/net/wireless/realtek/rtw89/core.c
> +++ b/drivers/net/wireless/realtek/rtw89/core.c
> @@ -4398,16 +4398,44 @@ static void rtw89_init_he_eht_cap(struct rtw89_dev *rtwdev,
>         _ieee80211_set_sband_iftype_data(sband, iftype_data, idx);
>  }
> 
> +static struct ieee80211_supported_band *rtw89_copy_sband(const struct ieee80211_supported_band *sband)

prefer naming rtw89_core_sband_dup().

> +{
> +       struct ieee80211_supported_band *copy = kmemdup(sband, sizeof(*sband), GFP_KERNEL);

Then, '*dup'. 

> +
> +       copy->channels = kmemdup(sband->channels, sizeof(struct ieee80211_channel) * sband->n_channels,
> GFP_KERNEL);

I'm planning to use devm_ series to manage sband data, so we don't need to
free them one by one. Do you interest to adjust that along with this patchset? 
I mean adding additional patches to adjust the code before this patch, and
make them as a patchset. 

For kmemdup, the corresponding one is devm_kmemdup. 

The line is too long. Less than 80 characters is preferred. 

> +       if (!copy->channels) {
> +               kfree(copy);
> +               return NULL;
> +       }
> +
> +       copy->bitrates = kmemdup(sband->bitrates, sizeof(struct ieee80211_rate) * sband->n_bitrates,
> GFP_KERNEL);

Since you have duplicated arrays of channels and bitrate, we should add const
to them, like:

  static const struct ieee80211_channel rtw89_channels_{2ghz,5ghz,6ghz}[]
  static const struct ieee80211_rate rtw89_bitrates[]


> +       if (!copy->bitrates) {
> +               kfree(copy->channels);
> +               kfree(copy);
> +               return NULL;
> +       }
> +
> +       return copy;
> +}
> +
> +static void rtw89_free_sband(const struct ieee80211_supported_band *sband)
> +{
> +       if (sband) {
> +               kfree(sband->bitrates);
> +               kfree(sband->channels);
> +               kfree(sband);
> +       }
> +}
> +
>  static int rtw89_core_set_supported_band(struct rtw89_dev *rtwdev)
>  {
>         struct ieee80211_hw *hw = rtwdev->hw;
>         struct ieee80211_supported_band *sband_2ghz = NULL, *sband_5ghz = NULL;
>         struct ieee80211_supported_band *sband_6ghz = NULL;
> -       u32 size = sizeof(struct ieee80211_supported_band);
>         u8 support_bands = rtwdev->chip->support_bands;
> 
>         if (support_bands & BIT(NL80211_BAND_2GHZ)) {
> -               sband_2ghz = kmemdup(&rtw89_sband_2ghz, size, GFP_KERNEL);
> +               sband_2ghz = rtw89_copy_sband(&rtw89_sband_2ghz);
>                 if (!sband_2ghz)
>                         goto err;
>                 rtw89_init_ht_cap(rtwdev, &sband_2ghz->ht_cap);
> @@ -4416,7 +4444,7 @@ static int rtw89_core_set_supported_band(struct rtw89_dev *rtwdev)
>         }
> 
>         if (support_bands & BIT(NL80211_BAND_5GHZ)) {
> -               sband_5ghz = kmemdup(&rtw89_sband_5ghz, size, GFP_KERNEL);
> +               sband_5ghz = rtw89_copy_sband(&rtw89_sband_5ghz);
>                 if (!sband_5ghz)
>                         goto err;
>                 rtw89_init_ht_cap(rtwdev, &sband_5ghz->ht_cap);
> @@ -4426,7 +4454,7 @@ static int rtw89_core_set_supported_band(struct rtw89_dev *rtwdev)
>         }
> 
>         if (support_bands & BIT(NL80211_BAND_6GHZ)) {
> -               sband_6ghz = kmemdup(&rtw89_sband_6ghz, size, GFP_KERNEL);
> +               sband_6ghz = rtw89_copy_sband(&rtw89_sband_6ghz);
>                 if (!sband_6ghz)
>                         goto err;
>                 rtw89_init_he_eht_cap(rtwdev, NL80211_BAND_6GHZ, sband_6ghz);
> @@ -4445,9 +4473,9 @@ static int rtw89_core_set_supported_band(struct rtw89_dev *rtwdev)
>                 kfree((__force void *)sband_5ghz->iftype_data);
>         if (sband_6ghz)
>                 kfree((__force void *)sband_6ghz->iftype_data);
> -       kfree(sband_2ghz);
> -       kfree(sband_5ghz);
> -       kfree(sband_6ghz);
> +       rtw89_free_sband(sband_2ghz);
> +       rtw89_free_sband(sband_5ghz);
> +       rtw89_free_sband(sband_6ghz);
>         return -ENOMEM;
>  }
> 
> @@ -4461,9 +4489,9 @@ static void rtw89_core_clr_supported_band(struct rtw89_dev *rtwdev)
>                 kfree((__force void *)hw->wiphy->bands[NL80211_BAND_5GHZ]->iftype_data);
>         if (hw->wiphy->bands[NL80211_BAND_6GHZ])
>                 kfree((__force void *)hw->wiphy->bands[NL80211_BAND_6GHZ]->iftype_data);
> -       kfree(hw->wiphy->bands[NL80211_BAND_2GHZ]);
> -       kfree(hw->wiphy->bands[NL80211_BAND_5GHZ]);
> -       kfree(hw->wiphy->bands[NL80211_BAND_6GHZ]);
> +       rtw89_free_sband(hw->wiphy->bands[NL80211_BAND_2GHZ]);
> +       rtw89_free_sband(hw->wiphy->bands[NL80211_BAND_5GHZ]);
> +       rtw89_free_sband(hw->wiphy->bands[NL80211_BAND_6GHZ]);
>         hw->wiphy->bands[NL80211_BAND_2GHZ] = NULL;
>         hw->wiphy->bands[NL80211_BAND_5GHZ] = NULL;
>         hw->wiphy->bands[NL80211_BAND_6GHZ] = NULL;

Like I mentioned above, with devm_ series, I suppose this function can be
removed entirely. 



  reply	other threads:[~2025-04-28  1:53 UTC|newest]

Thread overview: 4+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-04-27  0:24 [PATCH] wifi: rtw89: Fix inadverent sharing of struct ieee80211_supported_band data Ondřej Jirman
2025-04-28  1:53 ` Ping-Ke Shih [this message]
2025-04-28 21:50   ` Ondřej Jirman
2025-04-29  0:21     ` Ping-Ke Shih

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=d3c6e149a2794551ba4570bdd7f1b7e7@realtek.com \
    --to=pkshih@realtek.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-wireless@vger.kernel.org \
    --cc=megi@xff.cz \
    /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