Hi Brad, [I will now drop the CC to the bug from Bill, as we do not know yet if it is the same issue, Bill if you can test and this fixes the same for you we can merge the two, otherwise bisecting your issue would be helpful as well]
On Mon, May 04, 2026 at 05:11:01AM -0400, Brad Barnett wrote: > > > > # git bisect good > 0c4f1c02d27a880b10b58c63f574f13bed4f711d is the first bad commit > commit 0c4f1c02d27a880b10b58c63f574f13bed4f711d > Author: Miri Korenblit <[email protected]> > Date: Wed Jan 7 14:04:51 2026 +0200 > > wifi: cfg80211: stop NAN and P2P in cfg80211_leave > > [ Upstream commit e1696c8bd0056bc1a5f7766f58ac333adc203e8a ] > > Seems that there is an assumption that this function should be called > only for netdev interfaces, but it can also be called in suspend, or > from nl80211_netlink_notify (indirectly). > Note that the documentation of NL80211_ATTR_SOCKET_OWNER explicitly > says that NAN interfaces would be destroyed as well in the > nl80211_netlink_notify case. > > Fix this by also stopping P2P and NAN. > > Fixes: cb3b7d87652a ("cfg80211: add start / stop NAN commands") > Reviewed-by: Johannes Berg <[email protected]> > Signed-off-by: Miri Korenblit <[email protected]> > Link: > https://patch.msgid.link/20260107140430.dab142cbef0b.I290cc47836d56dd7e35012ce06bec36c6da688cd@changeid > Signed-off-by: Johannes Berg <[email protected]> > Signed-off-by: Sasha Levin <[email protected]> > > net/wireless/core.c | 4 +++- > 1 file changed, 3 insertions(+), 1 deletion(-) > > > > > # git bisect log > git bisect start > # status: waiting for both good and bad commits > # good: [779f9571ac3e3b2c969690d09e5353f56b7ed4ef] Linux 6.1.164 > git bisect good 779f9571ac3e3b2c969690d09e5353f56b7ed4ef > # status: waiting for bad commit, 1 good commit known > # bad: [4931e0e1673d2592ba7ab7365a25d1248b6a41b8] Linux 6.1.170 > git bisect bad 4931e0e1673d2592ba7ab7365a25d1248b6a41b8 > # bad: [eb44a35b6da3f0d393ef87eb27c33626865ec2ca] ASoC: core: Exit all links > before removing their components > git bisect bad eb44a35b6da3f0d393ef87eb27c33626865ec2ca > # bad: [da6b1bf764cc169cf27ff2474f315df71f698ab2] PCI: Mark ASM1164 SATA > controller to avoid bus reset > git bisect bad da6b1bf764cc169cf27ff2474f315df71f698ab2 > # bad: [64303b92d94c0c7845a273acd8d84b796d6f1db7] SUNRPC: auth_gss: fix > memory leaks in XDR decoding error paths > git bisect bad 64303b92d94c0c7845a273acd8d84b796d6f1db7 > # good: [03438554ee7b74891634069989190a4bee9f9325] PCI: mediatek: Fix IRQ > domain leak when MSI allocation fails > git bisect good 03438554ee7b74891634069989190a4bee9f9325 > # bad: [7fe1872ab6c465b2a20d573898e667b5e0b5d6bf] IB/cache: update gid cache > on client reregister event > git bisect bad 7fe1872ab6c465b2a20d573898e667b5e0b5d6bf > # bad: [f3b70cc7994be7378c34dac01a7cfec4b001072d] PCI: Initialize RCB from > pci_configure_device() > git bisect bad f3b70cc7994be7378c34dac01a7cfec4b001072d > # bad: [ad45a1c6d3c50c930ffffd76157e58598c914e62] netfilter: nft_compat: add > more restrictions on netlink attributes > git bisect bad ad45a1c6d3c50c930ffffd76157e58598c914e62 > # good: [40089ce7dfe3c4db9b239e89b172ee302c5f6a4f] net: mctp-i2c: fix > duplicate reception of old data > git bisect good 40089ce7dfe3c4db9b239e89b172ee302c5f6a4f > # bad: [0c4f1c02d27a880b10b58c63f574f13bed4f711d] wifi: cfg80211: stop NAN > and P2P in cfg80211_leave > git bisect bad 0c4f1c02d27a880b10b58c63f574f13bed4f711d > # good: [93e01e837e105299f1c259ef71f6e1ec4fe806e3] mctp i2c: initialise event > handler read bytes > git bisect good 93e01e837e105299f1c259ef71f6e1ec4fe806e3 > # first bad commit: [0c4f1c02d27a880b10b58c63f574f13bed4f711d] wifi: > cfg80211: stop NAN and P2P in cfg80211_leave Thanks, that was helpful. A revert of this commit is queued upstream as per https://lore.kernel.org/regressions/2026042343-unhearing-thee-53d1@gregkh/ So we can expect it to be in the next update for the 6.1.y stable series. Regards, Salvatore
