From patchwork Fri Jul 26 18:01:57 2024 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: James Prestwood X-Patchwork-Id: 13743013 Received: from mail-qv1-f45.google.com (mail-qv1-f45.google.com [209.85.219.45]) (using TLSv1.2 with cipher ECDHE-RSA-AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id D6A923BBF1 for ; Fri, 26 Jul 2024 18:02:04 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=209.85.219.45 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1722016926; cv=none; b=kpcvkn4J2nPwwe1dFps97QT2nF5lK6Gxty65FWYKUPib804mQXSnEyvHQrp3oxu4vcTgZd297aSwliByvQmVKdIruJYFMRDTOPnJnXTdidILgZnkqb4F2F18C9tkDii+IIPJPwEvfJduhA8WSZH/+Nj9eIEwAcv9/v6CHA+cdn8= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1722016926; c=relaxed/simple; bh=/QBvcYRv/QuBLo2peiVZ6trFiOhu6YMZJfPICZCtz+M=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=uZPLTe4av6JZ+8ZkWZmUElQjMPNYs5Cpj8rXg2+OoHYIqpeQjZmSOmHoFa3RLDBoj07YjUOtntBGn7C32SIJFpVCOJXvPRrwqiWgkFIz1C/qybLU3ndRTiIfiesvLDr8kdbgQ67qhSvv28aAMJd0IQmKR8nkDG1PpvoMeLdjZoY= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com; spf=pass smtp.mailfrom=gmail.com; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b=GishLNMC; arc=none smtp.client-ip=209.85.219.45 Authentication-Results: smtp.subspace.kernel.org; dmarc=pass (p=none dis=none) header.from=gmail.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=gmail.com Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b="GishLNMC" Received: by mail-qv1-f45.google.com with SMTP id 6a1803df08f44-6b7a3b75133so6907936d6.2 for ; Fri, 26 Jul 2024 11:02:04 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20230601; t=1722016923; x=1722621723; darn=lists.linux.dev; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:from:to:cc:subject:date :message-id:reply-to; bh=HcftX5HJW/ffAWLqOplaweDSHLZMd/m9+wbktsfZfmo=; b=GishLNMCxZeI+DOgWCEUi0xjcGchnpsSCuV52bymojSyZ3MZPqyO7zEymYNKDxZs2J rtBGdmNWdQHRDxOVs03CjzdyR2kCSgSGqe00dpY7Fh9hoRc9pV5d6KhMeMCK/3Zf4QK+ pQzPBst0EJ8O3OW5H1GKjZMLbSkYXjSxncx56V1EC/YOX9CmRhWOkVK2ouDpbz2AyExh Jb4MeP56r4gU+sEZGfz0RtRRicRcbFOYM9s3/SUmEY+/Oa3j8GP66DvicMpfi8IXOGka jjdXOJsLrl6kpkq+VuRuHbldgLBdzAlY5ByrZcp097eJiAS1dp/oQH/XasSTktnS/mF9 TmYA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20230601; t=1722016923; x=1722621723; h=content-transfer-encoding:mime-version:references:in-reply-to :message-id:date:subject:cc:to:from:x-gm-message-state:from:to:cc :subject:date:message-id:reply-to; bh=HcftX5HJW/ffAWLqOplaweDSHLZMd/m9+wbktsfZfmo=; b=CD2s4reDgO3kMz2M/XMYWwSWXE4OWV8TR8omBwfnT4zRPIQM3ZUFziW3OdjGMdlSMl FW77DAqsPBF010XrMSvY1Rx2xnNNPUTpxZStE+obWM17md0cfIWEu2/ew7qOmTmRT7/O zQVKlhCMdPmC4Ekjvwm8r+vRuspGzRaz2vIPpIRB3jKrzecd7TRSX6lwOpsqZCUc95cR sE2Fc9ZNc4X0Gvg95WZeyLsMQzKjptUlMCMZ98h3fTmV/XwnEQ8wrMUxoNUprYTHInwZ uzgsanxvFZZ+ZY6Cs9x1zOI5YTSjet7znSuMCS4U5mxsINGSVL804Yi9fCSbeRLTUEr5 PB7A== X-Gm-Message-State: AOJu0YxjngLf7lVhO2JMipDK0/CaQab5FVVxacPMOoS54O9RGdeye8i9 k1tq61P+WWXjZfnAUp2QkiE7C+K98h+V8kEQBoOuZR+yAdP1NIw8GLd+Yw== X-Google-Smtp-Source: AGHT+IFjqSuCLrSFFBb53rHiFR3Eb2HFnD8Ry83GlB35X9lJCmXG4JrOHDt2QOOdPNuWveL5x1FMpQ== X-Received: by 2002:a05:6214:1247:b0:6b4:dd2a:aa44 with SMTP id 6a1803df08f44-6bb55a83e51mr8909176d6.37.1722016923315; Fri, 26 Jul 2024 11:02:03 -0700 (PDT) Received: from LOCLAP699.locus-rst-dev-locuspark.locus ([152.193.78.90]) by smtp.gmail.com with ESMTPSA id 6a1803df08f44-6bb3f8fccc8sm18877166d6.46.2024.07.26.11.02.02 (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); Fri, 26 Jul 2024 11:02:03 -0700 (PDT) From: James Prestwood To: iwd@lists.linux.dev Cc: James Prestwood Subject: [PATCH v2 3/3] netdev: fall back to RSSI polling if SET_CQM fails Date: Fri, 26 Jul 2024 11:01:57 -0700 Message-Id: <20240726180157.1654827-3-prestwoj@gmail.com> X-Mailer: git-send-email 2.34.1 In-Reply-To: <20240726180157.1654827-1-prestwoj@gmail.com> References: <20240726180157.1654827-1-prestwoj@gmail.com> Precedence: bulk X-Mailing-List: iwd@lists.linux.dev List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Some drivers fail to set a CQM threshold and report not supported. Its unclear exactly why but if this happens roaming is effectively broken. To work around this enable RSSI polling if -ENOTSUP is returned. The polling callback has been changed to emit the HIGH/LOW signal threshold events instead of just the RSSI level index, just as if a CQM event came from the kernel. --- src/netdev.c | 147 +++++++++++++++++++++++++++++++++------------------ 1 file changed, 96 insertions(+), 51 deletions(-) v2: * Removed some changes that shouldn't have been in this patch diff --git a/src/netdev.c b/src/netdev.c index 2d70fc38..47825c1c 100644 --- a/src/netdev.c +++ b/src/netdev.c @@ -187,6 +187,7 @@ struct netdev { bool retry_auth : 1; bool in_reassoc : 1; bool privacy : 1; + bool cqm_poll_fallback : 1; }; struct netdev_preauth_state { @@ -206,6 +207,9 @@ static struct l_genl_family *nl80211; static struct l_queue *netdev_list; static struct watchlist netdev_watches; static bool mac_per_ssid; +/* Threshold RSSI for roaming to trigger, configurable in main.conf */ +static int LOW_SIGNAL_THRESHOLD; +static int LOW_SIGNAL_THRESHOLD_5GHZ; static unsigned int iov_ie_append(struct iovec *iov, unsigned int n_iov, unsigned int c, @@ -642,6 +646,46 @@ static void netdev_set_rssi_level_idx(struct netdev *netdev) netdev->cur_rssi_level_idx = new_level; } +static void netdev_cqm_event_rssi_value(struct netdev *netdev, int rssi_val) +{ + bool new_rssi_low; + uint8_t prev_rssi_level_idx = netdev->cur_rssi_level_idx; + int threshold = netdev->frequency > 4000 ? LOW_SIGNAL_THRESHOLD_5GHZ : + LOW_SIGNAL_THRESHOLD; + + if (!netdev->connected) + return; + + if (rssi_val > 127) + rssi_val = 127; + else if (rssi_val < -127) + rssi_val = -127; + + netdev->cur_rssi = rssi_val; + + if (!netdev->event_filter) + return; + + new_rssi_low = rssi_val < threshold; + if (netdev->cur_rssi_low != new_rssi_low) { + int event = new_rssi_low ? + NETDEV_EVENT_RSSI_THRESHOLD_LOW : + NETDEV_EVENT_RSSI_THRESHOLD_HIGH; + + netdev->cur_rssi_low = new_rssi_low; + netdev->event_filter(netdev, event, NULL, netdev->user_data); + } + + if (!netdev->rssi_levels_num) + return; + + netdev_set_rssi_level_idx(netdev); + if (netdev->cur_rssi_level_idx != prev_rssi_level_idx) + netdev->event_filter(netdev, NETDEV_EVENT_RSSI_LEVEL_NOTIFY, + &netdev->cur_rssi_level_idx, + netdev->user_data); +} + static void netdev_rssi_poll_cb(struct l_genl_msg *msg, void *user_data) { struct netdev *netdev = user_data; @@ -678,11 +722,16 @@ static void netdev_rssi_poll_cb(struct l_genl_msg *msg, void *user_data) netdev->cur_rssi = info.cur_rssi; /* - * Note we don't have to handle LOW_SIGNAL_THRESHOLD here. The - * CQM single threshold RSSI monitoring should work even if the - * kernel driver doesn't support multiple thresholds. So the - * polling only handles the client-supplied threshold list. + * If the CMD_SET_CQM call failed RSSI polling was started. In this case + * we should behave just like its a CQM event and check both the RSSI + * level indexes and the HIGH/LOW thresholds. */ + if (netdev->cqm_poll_fallback) { + netdev_cqm_event_rssi_value(netdev, info.cur_rssi); + goto done; + } + + /* Otherwise just update the level notifications, CQM events work */ netdev_set_rssi_level_idx(netdev); if (netdev->cur_rssi_level_idx != prev_rssi_level_idx) netdev->event_filter(netdev, NETDEV_EVENT_RSSI_LEVEL_NOTIFY, @@ -1027,6 +1076,9 @@ static void netdev_free(void *data) netdev->get_link_cmd_id = 0; } + if (netdev->rssi_poll_timeout) + l_timeout_remove(netdev->rssi_poll_timeout); + scan_wdev_remove(netdev->wdev_id); watchlist_destroy(&netdev->station_watches); @@ -1058,10 +1110,6 @@ struct netdev *netdev_find(int ifindex) return l_queue_find(netdev_list, netdev_match, L_UINT_TO_PTR(ifindex)); } -/* Threshold RSSI for roaming to trigger, configurable in main.conf */ -static int LOW_SIGNAL_THRESHOLD; -static int LOW_SIGNAL_THRESHOLD_5GHZ; - static void netdev_cqm_event_rssi_threshold(struct netdev *netdev, uint32_t rssi_event) { @@ -1085,46 +1133,6 @@ static void netdev_cqm_event_rssi_threshold(struct netdev *netdev, netdev->event_filter(netdev, event, NULL, netdev->user_data); } -static void netdev_cqm_event_rssi_value(struct netdev *netdev, int rssi_val) -{ - bool new_rssi_low; - uint8_t prev_rssi_level_idx = netdev->cur_rssi_level_idx; - int threshold = netdev->frequency > 4000 ? LOW_SIGNAL_THRESHOLD_5GHZ : - LOW_SIGNAL_THRESHOLD; - - if (!netdev->connected) - return; - - if (rssi_val > 127) - rssi_val = 127; - else if (rssi_val < -127) - rssi_val = -127; - - netdev->cur_rssi = rssi_val; - - if (!netdev->event_filter) - return; - - new_rssi_low = rssi_val < threshold; - if (netdev->cur_rssi_low != new_rssi_low) { - int event = new_rssi_low ? - NETDEV_EVENT_RSSI_THRESHOLD_LOW : - NETDEV_EVENT_RSSI_THRESHOLD_HIGH; - - netdev->cur_rssi_low = new_rssi_low; - netdev->event_filter(netdev, event, NULL, netdev->user_data); - } - - if (!netdev->rssi_levels_num) - return; - - netdev_set_rssi_level_idx(netdev); - if (netdev->cur_rssi_level_idx != prev_rssi_level_idx) - netdev->event_filter(netdev, NETDEV_EVENT_RSSI_LEVEL_NOTIFY, - &netdev->cur_rssi_level_idx, - netdev->user_data); -} - static void netdev_cqm_event(struct l_genl_msg *msg, struct netdev *netdev) { struct l_genl_attr attr; @@ -3634,11 +3642,48 @@ static struct l_genl_msg *netdev_build_cmd_cqm_rssi_update( static void netdev_cmd_set_cqm_cb(struct l_genl_msg *msg, void *user_data) { + struct netdev *netdev = user_data; int err = l_genl_msg_get_error(msg); const char *ext_error; - if (err >= 0) + if (err >= 0) { + /* + * Looking at some driver code it appears that the -ENOTSUP CQM + * failure could be transient. Just in case, reset the fallback + * flag if CQM happens to start working again. + */ + if (netdev->cqm_poll_fallback) { + l_debug("CMD_SET_CQM succeeded, stop polling fallback"); + + if (netdev->rssi_poll_timeout) { + l_timeout_remove(netdev->rssi_poll_timeout); + netdev->rssi_poll_timeout = NULL; + } + + netdev->cqm_poll_fallback = false; + } + + return; + } + + /* + * Some drivers enable beacon filtering but also use software CQM which + * mac80211 detects and returns -ENOTSUP. There is no way to check this + * ahead of time so if we see this start polling in order to get RSSI + * updates. + */ + if (err == -ENOTSUP) { + l_debug("CMD_SET_CQM not supported, falling back to polling"); + netdev->cqm_poll_fallback = true; + + if (netdev->rssi_poll_timeout) + return; + + netdev->rssi_poll_timeout = l_timeout_create(1, + netdev_rssi_poll, netdev, NULL); + return; + } ext_error = l_genl_msg_get_extended_error(msg); l_error("CMD_SET_CQM failed: %s", @@ -3662,7 +3707,7 @@ static int netdev_cqm_rssi_update(struct netdev *netdev) return -EINVAL; if (!l_genl_family_send(nl80211, msg, netdev_cmd_set_cqm_cb, - NULL, NULL)) { + netdev, NULL)) { l_genl_msg_unref(msg); return -EIO; } @@ -5260,7 +5305,7 @@ int netdev_set_rssi_report_levels(struct netdev *netdev, const int8_t *levels, return -EINVAL; if (!l_genl_family_send(nl80211, cmd_set_cqm, netdev_cmd_set_cqm_cb, - NULL, NULL)) { + netdev, NULL)) { l_genl_msg_unref(cmd_set_cqm); return -EIO; }