linux-kernel - Re: [EXT] Re: [PATCH v2] wifi: mwifiex: avoid AP and STA running on different channel

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: <ZuAD5fKyKgHPFN8s@pengutronix.de>
Date: Tue, 10 Sep 2024 10:31:33 +0200
From: Sascha Hauer <s.hauer@...gutronix.de>
To: David Lin <yu-hao.lin@....com>
Cc: "linux-wireless@...r.kernel.org" <linux-wireless@...r.kernel.org>,
	"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
	"briannorris@...omium.org" <briannorris@...omium.org>,
	"kvalo@...nel.org" <kvalo@...nel.org>,
	"francesco@...cini.it" <francesco@...cini.it>,
	Pete Hsieh <tsung-hsien.hsieh@....com>
Subject: Re: [EXT] Re: [PATCH v2] wifi: mwifiex: avoid AP and STA running on
 different channel

On Tue, Sep 10, 2024 at 08:19:59AM +0000, David Lin wrote:
> > From: Sascha Hauer <s.hauer@...gutronix.de>
> > Sent: Tuesday, September 10, 2024 4:00 PM
> > To: David Lin <yu-hao.lin@....com>
> > Cc: linux-wireless@...r.kernel.org; linux-kernel@...r.kernel.org;
> > briannorris@...omium.org; kvalo@...nel.org; francesco@...cini.it; Pete
> > Hsieh <tsung-hsien.hsieh@....com>
> > Subject: Re: [EXT] Re: [PATCH v2] wifi: mwifiex: avoid AP and STA running on
> > different channel
> > 
> > Caution: This is an external email. Please take care when clicking links or
> > opening attachments. When in doubt, report the message using the 'Report
> > this email' button
> > 
> > 
> > On Tue, Sep 10, 2024 at 01:52:02AM +0000, David Lin wrote:
> > > > From: Sascha Hauer <s.hauer@...gutronix.de>
> > > > Sent: Tuesday, September 10, 2024 5:05 AM
> > > > To: David Lin <yu-hao.lin@....com>
> > > > Cc: linux-wireless@...r.kernel.org; linux-kernel@...r.kernel.org;
> > > > briannorris@...omium.org; kvalo@...nel.org; francesco@...cini.it;
> > > > Pete Hsieh <tsung-hsien.hsieh@....com>
> > > > Subject: [EXT] Re: [PATCH v2] wifi: mwifiex: avoid AP and STA
> > > > running on different channel
> > > >
> > > > Caution: This is an external email. Please take care when clicking
> > > > links or opening attachments. When in doubt, report the message
> > > > using the 'Report this email' button
> > > >
> > > >
> > > > On Mon, Sep 02, 2024 at 04:43:11PM +0800, David Lin wrote:
> > > > > Current firmware doesn't support AP and STA running on different
> > > > > channels simultaneously.
> > > > > FW crash would occur in such case.
> > > > > This patch avoids the issue by disabling AP and STA to run on
> > > > > different channels.
> > > > >
> > > > > Signed-off-by: David Lin <yu-hao.lin@....com>
> > > > > ---
> > > > >
> > > > > v2:
> > > > >    - clean up code.
> > > > >
> > > > > ---
> > > > >  .../net/wireless/marvell/mwifiex/cfg80211.c   | 17 ++++---
> > > > >  drivers/net/wireless/marvell/mwifiex/util.c   | 44
> > +++++++++++++++++++
> > > > >  drivers/net/wireless/marvell/mwifiex/util.h   | 13 ++++++
> > > > >  3 files changed, 69 insertions(+), 5 deletions(-)
> > > > >
> > > > > diff --git a/drivers/net/wireless/marvell/mwifiex/cfg80211.c
> > > > > b/drivers/net/wireless/marvell/mwifiex/cfg80211.c
> > > > > index 722ead51e912..3dbcab463445 100644
> > > > > --- a/drivers/net/wireless/marvell/mwifiex/cfg80211.c
> > > > > +++ b/drivers/net/wireless/marvell/mwifiex/cfg80211.c
> > > > > @@ -781,11 +781,9 @@ mwifiex_cfg80211_set_wiphy_params(struct
> > > > > wiphy
> > > > *wiphy, u32 changed)
> > > > >               break;
> > > > >
> > > > >       case MWIFIEX_BSS_ROLE_STA:
> > > > > -             if (priv->media_connected) {
> > > > > -                     mwifiex_dbg(adapter, ERROR,
> > > > > -                                 "cannot change wiphy params
> > > > when connected");
> > > > > -                     return -EINVAL;
> > > > > -             }
> > > > > +             if (priv->media_connected)
> > > > > +                     break;
> > > >
> > > > This hunk seems unrelated to this patch. If this is needed then it
> > > > deserves an extra patch along with an explanation why this is necessary.
> > > >
> > > > Sascha
> > > >
> > >
> > > Without this hunk, AP and STA can't run on the same channel if some
> > > wiphy parameters are setting.
> > 
> > Ok, I now see where you are aiming at. Here's the problematic function:
> > 
> > > static int
> > > mwifiex_cfg80211_set_wiphy_params(struct wiphy *wiphy, u32 changed) {
> > >       ...
> > >
> > >       priv = mwifiex_get_priv(adapter, MWIFIEX_BSS_ROLE_ANY);
> > >
> > >       switch (priv->bss_role) {
> > >       case MWIFIEX_BSS_ROLE_UAP:
> > >               if (priv->bss_started) {
> > >                       mwifiex_dbg(adapter, ERROR,
> > >                                   "cannot change wiphy params
> > when bss started");
> > >                       return -EINVAL;
> > >               }
> > >
> > >               ...
> > >               mwifiex_send_cmd(priv,
> > HostCmd_CMD_UAP_SYS_CONFIG, ...);
> > >
> > >               break;
> > >       case MWIFIEX_BSS_ROLE_STA:
> > >               if (priv->media_connected) {
> > >                       mwifiex_dbg(adapter, ERROR,
> > >                                   "cannot change wiphy params
> > when connected");
> > >                       return -EINVAL;
> > >               }
> > >
> > >               ...
> > >               mwifiex_send_cmd(priv,
> > HostCmd_CMD_802_11_SNMP_MIB,
> > > ...);
> > >
> > >               break;
> > >       }
> > >
> > >       return 0;
> > > }
> > 
> > This function is for setting wiphy params like rts_threshold and others.
> > 
> > mwifiex_get_priv(adapter, MWIFIEX_BSS_ROLE_ANY) returns the first priv
> > which by default is in station mode. Now if you start priv0 in station mode,
> > then afterwards start priv1 in AP mode *and* have rts_threshold = xy in your
> > config, then you run into the "cannot change wiphy params when connected"
> > case.
> > 
> > I really wonder if the settings done in this function are per priv or per adapter.
> > Is there one rts_threshold setting in a mwifiex chip or are there multiple (per
> > vif/priv)?
> > 
> > If it's a global setting, then why are we interested in the media_connected
> > state of one specific priv? Shouldn't we check all privs?
> > 
> > If it's a setting per priv, then why do we choose the same priv everytime in this
> > function?
> > 
> > Either way, this function looks fishy and changing it should be done with an
> > explanation, just dropping the error message and returning success is not
> > enough.
> > 
> > Sascha
> > 
> > --
> 
> O.K. I will add comment in patch v3.

Really I hoped that you could clarify a bit how this function works and
answer some of the questions I raised above. That could help us to get
a better driver over time. The same code is in the nxpwifi driver as
well, so it's not only a matter of which driver we decide to maintain
in the future.

Sascha

-- 
Pengutronix e.K.                           |                             |
Steuerwalder Str. 21                       | http://www.pengutronix.de/  |
31137 Hildesheim, Germany                  | Phone: +49-5121-206917-0    |
Amtsgericht Hildesheim, HRA 2686           | Fax:   +49-5121-206917-5555 |