Currently if a user enqueues a work item using schedule_delayed_work() the
used wq is "system_wq" (per-cpu wq) while queue_delayed_work() use
WORK_CPU_UNBOUND (used when a cpu is not specified). The same applies to
schedule_work() that is using system_wq and queue_work(), that makes use
again of WORK_CPU_UNBOUND.
This lack of consistency cannot be addressed without refactoring the API.
For more details see the Link tag below.
alloc_workqueue() treats all queues as per-CPU by default, while unbound
workqueues must opt-in via WQ_UNBOUND.
This default is suboptimal: most workloads benefit from unbound queues,
allowing the scheduler to place worker threads where they’re needed and
reducing noise when CPUs are isolated.
This continues the effort to refactor workqueue APIs, which began with
the introduction of new workqueues and a new alloc_workqueue flag in:
commit 128ea9f6cc ("workqueue: Add system_percpu_wq and system_dfl_wq")
commit 930c2ea566 ("workqueue: Add new WQ_PERCPU flag")
This change adds the WQ_UNBOUND flag to explicitly request
alloc_workqueue() to be unbound, because this specific workload has no
benefit being per-cpu.
With the introduction of the WQ_PERCPU flag (equivalent to !WQ_UNBOUND),
any alloc_workqueue() caller that doesn’t explicitly specify WQ_UNBOUND
must now use WQ_PERCPU.
Once migration is complete, WQ_UNBOUND can be removed and unbound will
become the implicit default.
Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
Link: https://lore.kernel.org/all/20250221112003.1dSuoGyc@linutronix.de/
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251118102032.54375-3-marco.crivellari@suse.com
Currently if a user enqueues a work item using schedule_delayed_work() the
used wq is "system_wq" (per-cpu wq) while queue_delayed_work() use
WORK_CPU_UNBOUND (used when a cpu is not specified). The same applies to
schedule_work() that is using system_wq and queue_work(), that makes use
again of WORK_CPU_UNBOUND.
This lack of consistency cannot be addressed without refactoring the API.
For more details see the Link tag below.
alloc_workqueue() treats all queues as per-CPU by default, while unbound
workqueues must opt-in via WQ_UNBOUND.
This default is suboptimal: most workloads benefit from unbound queues,
allowing the scheduler to place worker threads where they’re needed and
reducing noise when CPUs are isolated.
This continues the effort to refactor workqueue APIs, which began with
the introduction of new workqueues and a new alloc_workqueue flag in:
commit 128ea9f6cc ("workqueue: Add system_percpu_wq and system_dfl_wq")
commit 930c2ea566 ("workqueue: Add new WQ_PERCPU flag")
This change adds the WQ_UNBOUND flag to explicitly request
alloc_workqueue() to be unbound, because this specific workload has no
benefit being per-cpu.
With the introduction of the WQ_PERCPU flag (equivalent to !WQ_UNBOUND),
any alloc_workqueue() caller that doesn’t explicitly specify WQ_UNBOUND
must now use WQ_PERCPU.
Once migration is complete, WQ_UNBOUND can be removed and unbound will
become the implicit default.
Suggested-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Marco Crivellari <marco.crivellari@suse.com>
Link: https://lore.kernel.org/all/20250221112003.1dSuoGyc@linutronix.de/
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251118102032.54375-2-marco.crivellari@suse.com
The rtl8187_rx_cb() calculates the rx descriptor header address
by subtracting its size from the skb tail pointer.
However, it does not validate if the received packet
(skb->len from urb->actual_length) is large enough to contain this
header.
If a truncated packet is received, this will lead to a buffer
underflow, reading memory before the start of the skb data area,
and causing a kernel panic.
Add length checks for both rtl8187 and rtl8187b descriptor headers
before attempting to access them, dropping the packet cleanly if the
check fails.
Fixes: 6f7853f3cb ("rtl8187: change rtl8187_dev.c to support RTL8187B (part 2)")
Signed-off-by: Seungjin Bae <eeodqql09@gmail.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251118013258.1789949-2-eeodqql09@gmail.com
In rtl8180_init_rx_ring(), memory is allocated for skb packets and DMA
allocations in a loop. When an allocation fails, the previously
successful allocations are not freed on exit.
Fix that by jumping to err_free_rings label on error, which calls
rtl8180_free_rx_ring() to free the allocations. Remove the free of
rx_ring in rtl8180_init_rx_ring() error path, and set the freed
priv->rx_buf entry to null, to avoid double free.
Fixes: f653211197 ("Add rtl8180 wireless driver")
Signed-off-by: Abdun Nihaal <nihaal@cse.iitm.ac.in>
Reviewed-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251114094527.79842-1-nihaal@cse.iitm.ac.in
Normally use rtw89_write32_mask() with a consecutive bit mask, but mask of
RX filter B_AX_RX_FLTR_CFG_MASK is bits 31-22 and 15-0, which excludes bits
of B_AX_RX_MPDU_MAX_LEN_MASK (bits 21-16).
Though the original logic is well to set RX filter, change it to a separate
function to avoid the tricky design.
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://patch.msgid.link/20251113053459.34995-1-pkshih@realtek.com
The existing flow is doing BB pre-init before downloading BB MCU firmware,
because existing chip RTL8922A has BB MCU. However, the coming chips don't
have this, and BB pre-init configuring registers can affect downloading
WiFi-CPU firmware. Therefore, calling BB pre-init afterward for new coming
chips without BB MCU.
For existing WiFi 6 chips, no BB pre-init. For RTL8922A, don't change
the logic.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251114060128.35363-5-pkshih@realtek.com
Driver needs to initialize registers before downloading firmware, so
move pre-init part (power on) from original rtw89_mac_init(). The
consequence patches will add more pre-init codes before downloading
firmware.
Since rtw89_phy_init_bb_afe() is used by coming RTL8922D, don't change
logic at all for existing chips.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251114060128.35363-4-pkshih@realtek.com
In the wpa_supplicant rekey flow, it sends an EAPOL packet 4/4 through
nl80211_tx_control_port() and triggers wake_tx_queue() in the driver.
Then, it sends nl80211_new_key() to configure a new key in mac80211.
However, in wake_tx_queue(), a workqueue is used to process the TX packet,
which might cause the driver to process the EAPOL packet later than
nl80211_new_key(). This results in the EAPOL packet 4/4 being transmitted
with the new key and IV, causing it to be dropped by the AP. Therefore,
needs to flush TX queue before deleting the old key to ensure that the
EAPOL 4/4 packet is transmitted using the old key.
Signed-off-by: Chih-Kang Chang <gary.chang@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251114060128.35363-2-pkshih@realtek.com
To be able to capture and inject packets in monitor mode, do RF calibration
once setting certain channel. Since calibration costs time, do not change
behavior of normal usage, which do calibration only when starting as AP
or going to connect AP.
Since driver declares IEEE80211_HW_WANT_MONITOR_VIF, the pure monitor vif
must be only one when adding interface with type NL80211_IFTYPE_MONITOR.
Otherwise, monitor vif must be NULL.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251111022452.28093-8-pkshih@realtek.com
Coverity reported a potential out-of-bounds access when 'bw' exceeds the
valid range for the specified band. Add a helper `rtw89_bw_is_valid()`
to check bandwidth validity for each band before accessing limit tables.
Addresses-Coverity-ID: 1598844 ("Out-of-bounds access")
Addresses-Coverity-ID: 1598896 ("Out-of-bounds access")
Signed-off-by: Kuan-Chung Chen <damon.chen@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251111022452.28093-6-pkshih@realtek.com
The rules to diagnose MAC have a common header, and a cmd field is used
to know the exact command and its format. The rules with the same tuple of
fields {sheet, seq} can be seen as a set of compound rules, which treat
it as positive rule if just one of the rules is positive.
Take EQUALV rules as example, if value of {addr, mask} is equal to
predefined value as field val, a rule is positive. Fields addr_name_offset
and msg_offset are offsets related to textual messages for human readable.
Format of common rule header (8 bytes)
+-------+-----+--------+-----------+-----+---------+
| sheet | cmd | seq[2] | io / band | len | rsvd[2] |
+-------+-----+--------+-----------+-----+---------+
Format of rule command is EQUALV (equal value) (24 bytes):
+------+------------------+------+-----+------------+---------+
| addr | addr_name_offset | mask | val | msg_offset | rsvd[4] |
+------+------------------+------+-----+------------+---------+
Format of message:
+-----+----------+
| len | string[] |
+-----+----------+
An example of output:
Plain(Ignore)/Rules/Positive: 115(4)/86/81
Where, Plain is total rules written in firmware element.
Ignore is the ignored rules, such as USB IO, but current is PCIE.
Rules is number of set of compound rules.
Positive is number of positive Rules.
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251111022452.28093-4-pkshih@realtek.com
The firmware element ID 28 is a set of rules to diagnose if MAC get
abnormal. The latter patch will use these rules via debugfs to know
the status.
The element contains rules with their textual messages shown as below:
+------------------------------------+
| |
| +-----------+ |
| | rule_size |-------|----------+
+----------------+-----------+-------+ -- |
| rule[0] | \ |
| rule[1] | | <---+
| : | /
+------------------------------------+ --
| msg[0] msg[1] | each msg has variable length
| msg[2] msg[3] ... | (with address align 2)
| ... |
+------------------------------------+
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251111022452.28093-3-pkshih@realtek.com
TX wait skbs need to be completed when they are done. PCIe part does this
inside rtw89_pci_tx_status() during RPP processing. Other HCIs use a
mechanism based on C2H firmware messages.
Store TX wait skbs inside TX report queue so that it'll be possible to
identify completed items inside the C2H handler. Try to do this as
similar to PCIe path as possible. When the corresponding TX wait skb is
found inside TX report queue, unlink it from there and call
rtw89_core_tx_wait_complete() to mark the completion.
If the callee waiting for the completion has already timed out, the TX
wait skb is placed into TX wait list (like PCIe part does).
It's important that during HCI reset all pending TX wait frames should be
completed inside hci.ops->reset method before calling
rtw89_tx_wait_list_clear().
Found by Linux Verification Center (linuxtesting.org).
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251104135720.321110-11-pchelkin@ispras.ru
Frames flagged with IEEE80211_TX_CTL_REQ_TX_STATUS mean the driver has to
report to mac80211 stack whether AP sent ACK for the null frame/probe
request or not. It's not implemented in USB part of the driver yet.
PCIe HCI has its own way of getting TX status incorporated into RPP
feature, and it's always enabled there. Other HCIs need a different
scheme based on processing C2H messages.
Thus define a .tx_rpt_enabled flag indicating which HCIs need to enable a
TX report feature. Currently it is USB only.
Toggle a bit in the TX descriptor and place flagged skbs in a fix-sized
queue to wait for a message from the firmware. Firmware maintains a 4-bit
sequence number for required frames hence the queue can contain just 16
elements simultaneously. That's enough for normal driver / firmware
communication. If the firmware crashes for any reason and doesn't provide
TX reports in time, driver will handle this and report the obsolete frames
as dropped.
rtw89 also has a new feature providing a TX report for each transmission
attempt. Ignore a failed TX status reported by the firmware until retry
limit is reached or successful status appears. When there is no success
and the retry limit is reached, report the frame up to the wireless stack
as failed eventually.
HCI reset should stop all pending TX activity so forcefully flush the
queue there.
Found by Linux Verification Center (linuxtesting.org).
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251104135720.321110-9-pchelkin@ispras.ru
During HCI reset all pending TX URBs should be canceled. Use anchor to
keep track of them and have an ability to cancel them synchronously.
Note however that canceling RX URBs can't be done here in
rtw89_usb_ops_reset() as it breaks driver initialization.
Found by Linux Verification Center (linuxtesting.org).
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251104135720.321110-8-pchelkin@ispras.ru
rtw89 has several ways of handling TX status report events. The first one
is based on RPP feature which is used by PCIe HCI. The other one depends
on firmware sending a corresponding C2H message, quite similar to what
rtw88 has.
Toggle a bit in the TX descriptor to indicate to the firmware that TX
report for the frame is expected. This will allow handling TX wait skbs
and the ones flagged with IEEE80211_TX_CTL_REQ_TX_STATUS correctly.
Do the bulk of the patch according to the vendor driver for RTL8851BU.
However, there are slight differences in C2H message format between
different types of chips. RTL885xB ones follow format V0. RTL8852C has
format V1, and RTL8922AU has format V2.
Found by Linux Verification Center (linuxtesting.org).
Suggested-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251104135720.321110-6-pchelkin@ispras.ru
Pass TX status value directly into rtw89_core_tx_wait_complete(). This
will make it a bit in sync with further patches and will give flexibility
in future work. Also use scope based RCU locking which simplifies the
code of the function.
Found by Linux Verification Center (linuxtesting.org).
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251104135720.321110-5-pchelkin@ispras.ru
When there is an attempt to write data and RTW89_FLAG_UNPLUGGED is set,
this means device is disconnected and no urb is submitted. Return
appropriate error code to the caller to properly free the allocated
resources.
Found by Linux Verification Center (linuxtesting.org).
Fixes: 2135c28be6 ("wifi: rtw89: Add usb.{c,h}")
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Signed-off-by: Ping-Ke Shih <pkshih@realtek.com>
Link: https://patch.msgid.link/20251104135720.321110-3-pchelkin@ispras.ru
