Return-path: Received: from [217.148.43.144] ([217.148.43.144]:50622 "EHLO mnementh.co.uk" rhost-flags-FAIL-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1751320AbdGQTD2 (ORCPT ); Mon, 17 Jul 2017 15:03:28 -0400 To: "linux-wireless@vger.kernel.org" Cc: Arend Van Spriel , Franky Lin , Hante Meuleman , davem@davemloft.net From: Ian Molton Subject: BUG: brcmfmac: rework headroom check in .start_xmit() causes 43430a1 to fail. Message-ID: <86761c42-c91d-d4a7-3ba4-314612dd7fd1@mnementh.co.uk> (sfid-20170717_210332_180416_8095F9DA) Date: Mon, 17 Jul 2017 20:03:20 +0100 MIME-Version: 1.0 Content-Type: text/plain; charset=windows-1252 Sender: linux-wireless-owner@vger.kernel.org List-ID: Hi folks, I've found that: 270a6c1f65fe brcmfmac: rework headroom check in .start_xmit() Causes the 43430a1 to become non-responsive. The log starts emitting these messages: brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 net_ratelimit: 3 callbacks suppressed brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 brcmfmac: brcmf_proto_bcdc_hdrpull: wlan0: non-BCDC packet received, flags 0x0 Without this commit, the card works fine. Again, Firmware version = wl0: Aug 29 2016 20:48:16 version 7.45.41.26 (r640327) FWID 01-4527cfab Does anyone actually test these commits before applying to mainline? Thats two total breakages I've found that are verifiable at module load time in the last couple of hours. Even if this is a case of me having old firmware - this should not cause this issue- at worst the driver should refuse to load. Better, it should disable or work around known bugs. -Ian