| Message ID | 20211130153943.3531922-1-andrei.otcheretianski@intel.com |
|---|---|
| State | Changes Requested |
| Headers | show |
| Series | nl80211: set NL80211_SCAN_FLAG_COLOCATED_6GHZ in scan | expand |
On Tue, Nov 30, 2021 at 05:39:43PM +0200, Andrei Otcheretianski wrote: > Set NL80211_SCAN_FLAG_COLOCATED_6GHZ in the scan params to enable > scanning for co-located AP's found in 2.4/5 GHz bands. > diff --git a/src/drivers/driver_nl80211_scan.c b/src/drivers/driver_nl80211_scan.c > @@ -300,6 +300,9 @@ nl80211_scan_common(struct i802_bss *bss, u8 cmd, > + /* scan for 6ghz colocated APs that reported by other APs */ > + scan_flags |= NL80211_SCAN_FLAG_COLOCATED_6GHZ; > + > if (scan_flags && > nla_put_u32(msg, NL80211_ATTR_SCAN_FLAGS, scan_flags)) Is it really correct to do this unconditionally? Wouldn't this result in undesired extra latency in scanning when the 6 GHz band is not used (e.g., for most P2P use cases) or when there is an explicit configuration (e.g., SETBAND) to disable use of the 6 GHz band?
Hi, > -----Original Message----- > From: Hostap <hostap-bounces@lists.infradead.org> On Behalf Of Jouni > Malinen > Sent: Saturday, December 11, 2021 13:21 > To: Otcheretianski, Andrei <andrei.otcheretianski@intel.com> > Cc: hostap@lists.infradead.org; Mussai, Tova <tova.mussai@intel.com> > Subject: Re: [PATCH] nl80211: set NL80211_SCAN_FLAG_COLOCATED_6GHZ > in scan > > On Tue, Nov 30, 2021 at 05:39:43PM +0200, Andrei Otcheretianski wrote: > > Set NL80211_SCAN_FLAG_COLOCATED_6GHZ in the scan params to enable > > scanning for co-located AP's found in 2.4/5 GHz bands. > > > diff --git a/src/drivers/driver_nl80211_scan.c > > b/src/drivers/driver_nl80211_scan.c > > @@ -300,6 +300,9 @@ nl80211_scan_common(struct i802_bss *bss, u8 > cmd, > > + /* scan for 6ghz colocated APs that reported by other APs */ > > + scan_flags |= NL80211_SCAN_FLAG_COLOCATED_6GHZ; > > + > > if (scan_flags && > > nla_put_u32(msg, NL80211_ATTR_SCAN_FLAGS, scan_flags)) > > Is it really correct to do this unconditionally? Wouldn't this result in undesired > extra latency in scanning when the 6 GHz band is not used (e.g., for most P2P > use cases) or when there is an explicit configuration (e.g., SETBAND) to > disable use of the 6 GHz band? > Without this flag, when user space requests the kernel to scan 6GHz channels, the 6GHz non-PSC channels would be scanned passively which is not desired. Setting this flag instructs the kernel to only scan non-PSC channels if relevant information was found in the RNR elements on previously found BSSs (e.g., matching SSID etc.). Thus, I think that this flag should always be set if not scanning passively. I'll prepare a fix that does not set this flag for passive scan. As for use cases that 6GHz are not used, additional logic is indeed needed to avoid scanning these channels. Regards, Ilan.
diff --git a/src/drivers/driver_nl80211_scan.c b/src/drivers/driver_nl80211_scan.c index 1316084805..356fbcebfa 100644 --- a/src/drivers/driver_nl80211_scan.c +++ b/src/drivers/driver_nl80211_scan.c @@ -300,6 +300,9 @@ nl80211_scan_common(struct i802_bss *bss, u8 cmd, NL80211_SCAN_FLAG_OCE_PROBE_REQ_DEFERRAL_SUPPRESSION; } + /* scan for 6ghz colocated APs that reported by other APs */ + scan_flags |= NL80211_SCAN_FLAG_COLOCATED_6GHZ; + if (scan_flags && nla_put_u32(msg, NL80211_ATTR_SCAN_FLAGS, scan_flags)) goto fail;