- 01 Jul, 2022 11 commits
-
-
Felix Fietkau authored
In order to maintain fairness, the amount of queueing needs to be limited beyond the simple per-station AQL budget, otherwise the driver can simply repeatedly do scheduling rounds until all queues that have not used their AQL budget become eligble. To be conservative, use the high AQL limit for the first txq and add half of the low AQL for each subsequent queue. Signed-off-by:
Felix Fietkau <nbd@nbd.name> Acked-by:
Toke Høiland-Jørgensen <toke@toke.dk> Link: https://lore.kernel.org/r/20220625212411.36675-5-nbd@nbd.nameSigned-off-by:
Johannes Berg <johannes.berg@intel.com>
-
Felix Fietkau authored
This allows proper deficit accounting to ensure that they don't carry their deficit until the next time they become active Signed-off-by:
-
Felix Fietkau authored
When queueing packets for a station, deficit only gets added once the packets have been transmitted, which could be much later. During that time, a lot of temporary unfairness could happen, which could lead to bursty behavior. Fix this by subtracting the aql_tx_pending when checking the deficit in tx scheduling. Signed-off-by:
-
Felix Fietkau authored
32 bit is more than enough range for the airtime deficit Signed-off-by:
-
Felix Fietkau authored
This reverts commits 6a789ba6 and 2433647b. The virtual time scheduler code has a number of issues: - queues slowed down by hardware/firmware powersave handling were not properly handled. - on ath10k in push-pull mode, tx queues that the driver tries to pull from were starved, causing excessive latency - delay between tx enqueue and reported airtime use were causing excessively bursty tx behavior The bursty behavior may also be present on the round-robin scheduler, but there it is much easier to fix without introducing additional regressions Signed-off-by:
-
Johannes Berg authored
Somehow kernel-doc complains here about strong markup, but we really don't need the [] so just remove that. Signed-off-by: -
Johannes Berg authored
These struct members no longer exist, remove them from documentation. Signed-off-by: -
Mauro Carvalho Chehab authored
struct link_sta_info has now a cur_max_bandwidth data: net/mac80211/sta_info.h:569: warning: Function parameter or member 'cur_max_bandwidth' not described in 'link_sta_info' Copy the meaning from struct sta_info, documenting it. Signed-off-by:
Mauro Carvalho Chehab <mchehab@kernel.org> Link: https://lore.kernel.org/r/37d898634bb30776442a33833c48cbb21c90ecc6.1656409369.git.mchehab@kernel.orgSigned-off-by:
-
Mauro Carvalho Chehab authored
The lack of the colon makes it not parse the function parameter: include/net/mac80211.h:6250: warning: Function parameter or member 'vif' not described in 'ieee80211_channel_switch_disconnect' Fix it. Signed-off-by:
-
Mauro Carvalho Chehab authored
There are currently 17 kernel-doc warnings on this file: include/net/cfg80211.h:391: warning: Function parameter or member 'bw' not described in 'ieee80211_eht_mcs_nss_supp' include/net/cfg80211.h:437: warning: Function parameter or member 'eht_cap' not described in 'ieee80211_sband_iftype_data' include/net/cfg80211.h:507: warning: Function parameter or member 's1g' not described in 'ieee80211_sta_s1g_cap' include/net/cfg80211.h:1390: warning: Function parameter or member 'counter_offset_beacon' not described in 'cfg80211_color_change_settings' include/net/cfg80211.h:1390: warning: Function parameter or member 'counter_offset_presp' not described in 'cfg80211_color_change_settings' include/net/cfg80211.h:1430: warning: Enum value 'STATION_PARAM_APPLY_STA_TXPOWER' not described in enum 'station_parameters_apply_mask' include/net/cfg80211.h:2195: warning: Function parameter or member 'dot11MeshConnectedToAuthServer' not described in 'mesh_config' include/net/cfg80211.h:2341: warning: Function parameter or member 'short_ssid' not described in 'cfg80211_scan_6ghz_params' include/net/cfg80211.h:3328: warning: Function parameter or member 'kck_len' not described in 'cfg80211_gtk_rekey_data' include/net/cfg80211.h:3698: warning: Function parameter or member 'ftm' not described in 'cfg80211_pmsr_result' include/net/cfg80211.h:3828: warning: Function parameter or member 'global_mcast_stypes' not described in 'mgmt_frame_regs' include/net/cfg80211.h:4977: warning: Function parameter or member 'ftm' not described in 'cfg80211_pmsr_capabilities' include/net/cfg80211.h:5742: warning: Function parameter or member 'u' not described in 'wireless_dev' include/net/cfg80211.h:5742: warning: Function parameter or member 'links' not described in 'wireless_dev' include/net/cfg80211.h:5742: warning: Function parameter or member 'valid_links' not described in 'wireless_dev' include/net/cfg80211.h:6076: warning: Function parameter or member 'is_amsdu' not described in 'ieee80211_data_to_8023_exthdr' include/net/cfg80211.h:6949: warning: Function parameter or member 'sig_dbm' not described in 'cfg80211_notify_new_peer_candidate' Address them, in order to build a better documentation from this header. Signed-off-by:
-
Peter Chiu authored
S1g action frame with code 22 is not protected so update the robust action frame list. Signed-off-by:
Peter Chiu <chui-hao.chiu@mediatek.com> Link: https://lore.kernel.org/r/20220622010820.17522-1-chui-hao.chiu@mediatek.comSigned-off-by:
-
- 21 Jun, 2022 13 commits
-
-
Ping-Ke Shih authored
The DPK is a kind of RF calibration whose algorithm is to fine tune parameters and calibrate, and check the result. If the result isn't good enough, it could adjust parameters and try again. This issue is to read and show the result, but it could be a negative calibration result that causes divisor 0 and core dump. So, fix it by phy_div() that does division only if divisor isn't zero; otherwise, zero is adopted. divide error: 0000 [#1] PREEMPT SMP NOPTI CPU: 1 PID: 728 Comm: wpa_supplicant Not tainted 5.10.114-16019-g462a1661811a #1 <HASH:d024 28> RIP: 0010:rtw8852a_dpk+0x14ae/0x288f [rtw89_core] RSP: 0018:ffffa9bb412a7520 EFLAGS: 00010246 RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: 00000000000180fc RDI: ffffa141d01023c0 RBP: ffffa9bb412a76a0 R08: 0000000000001319 R09: 00000000ffffff92 R10: ffffffffc0292de3 R11: ffffffffc00d2f51 R12: 0000000000000000 R13: ffffa141d01023c0 R14: ffffffffc0290250 R15: ffffa141d0102638 FS: 00007fa99f5c2740(0000) GS:ffffa142e5e80000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000013e8e010 CR3: 0000000110d2c000 CR4: 0000000000750ee0 PKRU: 55555554 Call Trace: rtw89_core_sta_add+0x95/0x9c [rtw89_core <HASH:d239 29>] rtw89_ops_sta_state+0x5d/0x108 [rtw89_core <HASH:d239 29>] drv_sta_state+0x115/0x66f [mac80211 <HASH:81fe 30>] sta_info_insert_rcu+0x45c/0x713 [mac80211 <HASH:81fe 30>] sta_info_insert+0xf/0x1b [mac80211 <HASH:81fe 30>] ieee80211_prep_connection+0x9d6/0xb0c [mac80211 <HASH:81fe 30>] ieee80211_mgd_auth+0x2aa/0x352 [mac80211 <HASH:81fe 30>] cfg80211_mlme_auth+0x160/0x1f6 [cfg80211 <HASH:00cd 31>] nl80211_authenticate+0x2e5/0x306 [cfg80211 <HASH:00cd 31>] genl_rcv_msg+0x371/0x3a1 ? nl80211_stop_sched_scan+0xe5/0xe5 [cfg80211 <HASH:00cd 31>] ? genl_rcv+0x36/0x36 netlink_rcv_skb+0x8a/0xf9 genl_rcv+0x28/0x36 netlink_unicast+0x27b/0x3a0 netlink_sendmsg+0x2aa/0x469 sock_sendmsg_nosec+0x49/0x4d ____sys_sendmsg+0xe5/0x213 __sys_sendmsg+0xec/0x157 ? syscall_enter_from_user_mode+0xd7/0x116 do_syscall_64+0x43/0x55 entry_SYSCALL_64_after_hwframe+0x44/0xa9 RIP: 0033:0x7fa99f6e689b Fixes: e3ec7017 ("rtw89: add Realtek 802.11ax driver") Signed-off-by:
Ping-Ke Shih <pkshih@realtek.com> Signed-off-by:
Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20220613065144.15647-1-pkshih@realtek.com
-
Po-Hao Huang authored
Previously we only disable invalid reports for 52A, since we plan to support more ICs in the future, enable settings for those as well. Signed-off-by:
Po-Hao Huang <phhuang@realtek.com> Signed-off-by:
-
Ping-Ke Shih authored
TX BD (TX ring index) and TX WD (WiFi descriptor buffer) are freed asynchronously. With burst packets, we free TX WD, but the corresponding TX BD couldn't be freed yet. Then, TX can possibly get stuck due to no more TX BD. To avoid this, ignore reclaiming TX BD only if TX WD is no free space, because at this moment TX BD must have some spaces. Otherwise, we reclaim TX BD to resolve TX stuck issue. Signed-off-by:
-
Ping-Ke Shih authored
In low power mode, regular IO is power off, so we don't schedule napi to poll RX and TX completion. Therefore, calling ieee80211_rx_napi() with napi instance causes long RX latency. To fix this, use NULL as argument, and then it can use netif_receive_skb_list() to receive. Signed-off-by:
-
Ping-Ke Shih authored
Somehow, firmware could report invalid TX rate, and we consider the invalid rate as 0 that will make a wrong decision. So, drop invalid reports, and also suppress the warning message. Signed-off-by:
-
Ping-Ke Shih authored
Some warning messages could bother users. With proper handling, these situations don't really affect usage, but we still need to keep monitor these messages. If they happen frequently, we must review driver or hardware design to clarify. Signed-off-by:
-
Kuan-Chung Chen authored
To improve VO throughput, we enable VO TX AMPDU. We measure the latency of enable or disable VO TX AMPDU. The experimental results show that the difference between the two is insignificant only 300µs, so the little impact can be ignored for user experience. Moreover, we found some APs will have a group key handshake timeout issue when the EAPOL's TID is already setup BA session. Therefore, when transmitting EAPOL, if EAPOL's TID BA session is already setup, we need to delete it. Signed-off-by:
Kuan-Chung Chen <damon.chen@realtek.com> Signed-off-by:
-
Kuan-Chung Chen authored
The potential TX stuck occurs when there are lots of packets to be transmitted and the boundary of the tx_resource is hit. Add this patch to avoid the TX stuck when burst traffic. Signed-off-by:
-
Ping-Ke Shih authored
Since we can allocate MAC ID, BSSID and address CAM to TDLS peer, declare we can support TDLS now. Signed-off-by:
-
Ping-Ke Shih authored
In STA mode, if peer is TDLS. Allocate a BSSID CAM entry with peer's address to match address properly, and then hardware can ACK peer's packets and receive packets to driver. Signed-off-by:
-
Ping-Ke Shih authored
Normally, we allocate a BSSID CAM to a vif. By hardware design, we must allocate a BSSID CAM to each TDLS peer, so separate BSSID CAM operations that will be used by later patches. This patch doesn't change logic at all. Signed-off-by:
-
Ping-Ke Shih authored
Normally, we only allocate an address CAM and single one MAC ID to AP in STA mode. To support TDLS, we handle TDLS peers like AP handles stations. Signed-off-by:
-
Alexey Kodanev authored
As a result of the execution of the inner while loop, the value of 'idx' can be equal to LINK_QUAL_MAX_RETRY_NUM. However, this is not checked after the loop and 'idx' is used to write the LINK_QUAL_MAX_RETRY_NUM size array 'lq_cmd->rs_table[idx]' below in the outer loop. The fix is to check the new value of 'idx' inside the nested loop, and break both loops if index equals the size. Checking it at the start is now pointless, so let's remove it. Detected using the static analysis tool - Svace. Fixes: be663ab6 ("iwlwifi: split the drivers for agn and legacy devices 3945/4965") Signed-off-by:
Alexey Kodanev <aleksei.kodanev@bell-sw.com> Signed-off-by:
-
- 20 Jun, 2022 16 commits
-
-
Veerendranath Jakkam authored
The MLO links used for connection with an MLD AP are decided by the driver in case of SME offloaded to driver. Add support for the drivers to indicate the information of links used for MLO connection in connect and roam callbacks, update the connected links information in wdev from connect/roam result sent by driver. Also, send the connected links information to userspace. Add a netlink flag attribute to indicate that userspace supports handling of MLO connection. Drivers must not do MLO connection when this flag is not set. This is to maintain backwards compatibility with older supplicant versions which doesn't have support for MLO connection. Signed-off-by:
Veerendranath Jakkam <quic_vjakkam@quicinc.com> Signed-off-by:
-
Shaul Triebitz authored
Print the link ID in link_info_changed function. Signed-off-by:
Shaul Triebitz <shaul.triebitz@intel.com> Signed-off-by:
-
Johannes Berg authored
Add initial support to hwsim for creating MLO-capable radios dynamically (and statically via module parameter, but that's not really recommended.) Signed-off-by: -
Johannes Berg authored
When retrieving scan data, expose not just whether or not the interface (possibly an MLD) is associated to the BSS or not, but also on which link ID if it is an MLD. Signed-off-by: -
Johannes Berg authored
When getting/dumping an interface, expose information about valid links. Signed-off-by: -
Shaul Triebitz authored
Add to hwsim a link_data structure. For now, move the beacon interval and beacon timer to the link_data structure (making them per link). Set a beacon timer for each link and transmit a per link beacon (pass the link_id to ieee80211_beacon_get). Signed-off-by:
-
Johannes Berg authored
We should set the STA deflink addresses in case no link is really added. Fixes: 046d2e7c ("mac80211: prepare sta handling for MLO support") Signed-off-by:
-
Johannes Berg authored
Maintain a hash table of link-sta addresses so we can find them for management frames etc. where addresses haven't been replaced by the drivers to the MLD address yet. Signed-off-by: -
Johannes Berg authored
We need to be able to access these in a race-free way under traffic while adding/removing them, so RCU-ify the pointers. This requires passing a link_sta to a lot of functions so we don't have to do the RCU handling everywhere. Signed-off-by: -
Johannes Berg authored
Hardcoding link[0] is more confusing than using deflink, so use deflink for now, which requires making the macro a bit safer. Signed-off-by: -
Johannes Berg authored
This is only used in rx.c, so move it into the file. Signed-off-by: -
Shaul Triebitz authored
Pass the link id through to the get_beacon and return the beacon for a specific link id. Signed-off-by:
-
Shaul Triebitz authored
In start_ap and stop_ap mac80211 callbacks pass the link_id to the drivers. Signed-off-by:
-
Shaul Triebitz authored
Use link and link_conf according to the link_id provided by cfg in start_ap/stop_ap and change_beacon. Also use them in the functions called by them. Note that for a non MLD device, the link_id is 0, and link[0] and link_conf[0] equal to deflink and bss_conf respectively (what was there before). Also, call vif_info_change for BSS related changes (SSID), and link_info_change for LINK related changes (instead of the legacy bss_info_change). Signed-off-by:
-
Johannes Berg authored
When a link is added or removed via nl80211, these are called. Implement them so we don't have to check in all the different per-link commands whether we've already created the necessary datastructures. Signed-off-by: -
Johannes Berg authored
Add some optional callbacks for link add/remove so that drivers can react here. Initially, I thought it would be sufficient to just create the link in start_ap etc., but it turns out that's not so simple, since there are quite a few callbacks that can be called: if they're erroneously without start_ap, things might crash. Thus it might be easier for drivers to allocate all the necessary data structures immediately, to not have to worry about it in each callback, since cfg80211 checks that the link ID is valid (has been added.) Signed-off-by:
-
