v7:
1. rebased to ath.git ath-202202211030
2. add patch "ath11k: fix the warning of dev_wake in mhi_pm_disable_transition()"
3. remove below 3 patch which has upstream.
ath11k: add ath11k_qmi_free_resource() for recovery
ath11k: configure RDDM size to mhi for recovery by firmware
ath11k: add support for device recovery for QCA6390/WCN6855
v6: add patch "ath11k: Add hw-restart option to simulate_fw_crash"
v5:
1. rebased to ath.git ath-202202030905
2. change a few commit message
3. fix count set sequence of ath11k_core_reset in "ath11k: add synchronization operation between reconfigure of mac80211 and ath11k_base"
4. add patch "ath11k: configure RDDM size to mhi for recovery by firmware" to support RDDM
5. move destroy_workqueue(ab->workqueue_aux) from ath11k_pci_remove() to ath11k_core_free()
v4: add patch "ath11k: fix invalid m3 buffer address"
recovery will fail when download firmware without this patch
v3: remove time_left set but not used in
"ath11k: add synchronization operation between reconfigure of mac80211 and ath11k_base"
v2: s/initilized/initialized in commit log of patch
"ath11k: add synchronization operation between reconfigure of mac80211 and ath11k_base"
Add support for device recovery.
Wen Gong (4):
ath11k: add support for device recovery for QCA6390/WCN6855
ath11k: add synchronization operation between reconfigure of mac80211
and ath11k_base
ath11k: Add hw-restart option to simulate_fw_crash
ath11k: fix the warning of dev_wake in mhi_pm_disable_transition()
drivers/net/wireless/ath/ath11k/core.c | 121 ++++++++++++++++++++--
drivers/net/wireless/ath/ath11k/core.h | 18 ++++
drivers/net/wireless/ath/ath11k/debugfs.c | 4 +
drivers/net/wireless/ath/ath11k/mac.c | 40 +++++++
drivers/net/wireless/ath/ath11k/mhi.c | 33 ++++++
drivers/net/wireless/ath/ath11k/pci.c | 12 ++-
6 files changed, 216 insertions(+), 12 deletions(-)
base-commit: 193dc0446985372a9b06b7572e1b98e1963c23a9
--
2.31.1
When test device recovery with below command, it has warning in message
as below.
echo assert > /sys/kernel/debug/ath11k/wcn6855\ hw2.0/simulate_fw_crash
echo assert > /sys/kernel/debug/ath11k/qca6390\ hw2.0/simulate_fw_crash
warning message:
[ 1965.642121] ath11k_pci 0000:06:00.0: simulating firmware assert crash
[ 1968.471364] ieee80211 phy0: Hardware restart was requested
[ 1968.511305] ------------[ cut here ]------------
[ 1968.511368] WARNING: CPU: 3 PID: 1546 at drivers/bus/mhi/core/pm.c:505 mhi_pm_disable_transition+0xb37/0xda0 [mhi]
[ 1968.511443] Modules linked in: ath11k_pci ath11k mac80211 libarc4 cfg80211 qmi_helpers qrtr_mhi mhi qrtr nvme nvme_core
[ 1968.511563] CPU: 3 PID: 1546 Comm: kworker/u17:0 Kdump: loaded Tainted: G W 5.17.0-rc3-wt-ath+ #579
[ 1968.511629] Hardware name: Intel(R) Client Systems NUC8i7HVK/NUC8i7HVB, BIOS HNKBLi70.86A.0067.2021.0528.1339 05/28/2021
[ 1968.511704] Workqueue: mhi_hiprio_wq mhi_pm_st_worker [mhi]
[ 1968.511787] RIP: 0010:mhi_pm_disable_transition+0xb37/0xda0 [mhi]
[ 1968.511870] Code: a9 fe ff ff 4c 89 ff 44 89 04 24 e8 03 46 f6 e5 44 8b 04 24 41 83 f8 01 0f 84 21 fe ff ff e9 4c fd ff ff 0f 0b e9 af f8 ff ff <0f> 0b e9 5c f8 ff ff 48 89 df e8 da 9e ee e3 e9 12 fd ff ff 4c 89
[ 1968.511923] RSP: 0018:ffffc900024efbf0 EFLAGS: 00010286
[ 1968.511969] RAX: 00000000ffffffff RBX: ffff88811d241250 RCX: ffffffffc0176922
[ 1968.512014] RDX: 0000000000000000 RSI: 0000000000000004 RDI: ffff888118a90a24
[ 1968.512059] RBP: ffff888118a90800 R08: 0000000000000000 R09: ffff888118a90a27
[ 1968.512102] R10: ffffed1023152144 R11: 0000000000000001 R12: ffff888118a908ac
[ 1968.512229] R13: ffff888118a90928 R14: dffffc0000000000 R15: ffff888118a90a24
[ 1968.512310] FS: 0000000000000000(0000) GS:ffff888234200000(0000) knlGS:0000000000000000
[ 1968.512405] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 1968.512493] CR2: 00007f5538f443a8 CR3: 000000016dc28001 CR4: 00000000003706e0
[ 1968.512587] Call Trace:
[ 1968.512672] <TASK>
[ 1968.512751] ? _raw_spin_unlock_irq+0x1f/0x40
[ 1968.512859] mhi_pm_st_worker+0x3ac/0x790 [mhi]
[ 1968.512959] ? mhi_pm_mission_mode_transition.isra.0+0x7d0/0x7d0 [mhi]
[ 1968.513063] process_one_work+0x86a/0x1400
[ 1968.513184] ? pwq_dec_nr_in_flight+0x230/0x230
[ 1968.513312] ? move_linked_works+0x125/0x290
[ 1968.513416] worker_thread+0x6db/0xf60
[ 1968.513536] ? process_one_work+0x1400/0x1400
[ 1968.513627] kthread+0x241/0x2d0
[ 1968.513733] ? kthread_complete_and_exit+0x20/0x20
[ 1968.513821] ret_from_fork+0x22/0x30
[ 1968.513924] </TASK>
Reason is mhi_deassert_dev_wake() from mhi_device_put() is called
but mhi_assert_dev_wake() from __mhi_device_get_sync() is not called
in progress of recovery. Commit 8e0559921f9a ("bus: mhi: core:
Skip device wake in error or shutdown state") add check for the
pm_state of mhi in __mhi_device_get_sync(), and the pm_state is not
the normal state untill recovery is completed, so it leads the
dev_wake is not 0 and above warning print in mhi_pm_disable_transition()
while checking mhi_cntrl->dev_wake.
Add check in ath11k_pci_write32()/ath11k_pci_read32() to skip call
mhi_device_put() if mhi_device_get_sync() does not really do wake,
then the warning gone.
Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03003-QCAHSPSWPL_V1_V2_SILICONZ_LITE-2
Signed-off-by: Wen Gong <[email protected]>
---
drivers/net/wireless/ath/ath11k/pci.c | 12 ++++++++----
1 file changed, 8 insertions(+), 4 deletions(-)
diff --git a/drivers/net/wireless/ath/ath11k/pci.c b/drivers/net/wireless/ath/ath11k/pci.c
index 903758751c99..8a3ff12057e8 100644
--- a/drivers/net/wireless/ath/ath11k/pci.c
+++ b/drivers/net/wireless/ath/ath11k/pci.c
@@ -191,6 +191,7 @@ void ath11k_pci_write32(struct ath11k_base *ab, u32 offset, u32 value)
{
struct ath11k_pci *ab_pci = ath11k_pci_priv(ab);
u32 window_start;
+ int ret = 0;
/* for offset beyond BAR + 4K - 32, may
* need to wakeup MHI to access.
@@ -198,7 +199,7 @@ void ath11k_pci_write32(struct ath11k_base *ab, u32 offset, u32 value)
if (ab->hw_params.wakeup_mhi &&
test_bit(ATH11K_PCI_FLAG_INIT_DONE, &ab_pci->flags) &&
offset >= ACCESS_ALWAYS_OFF)
- mhi_device_get_sync(ab_pci->mhi_ctrl->mhi_dev);
+ ret = mhi_device_get_sync(ab_pci->mhi_ctrl->mhi_dev);
if (offset < WINDOW_START) {
iowrite32(value, ab->mem + offset);
@@ -222,7 +223,8 @@ void ath11k_pci_write32(struct ath11k_base *ab, u32 offset, u32 value)
if (ab->hw_params.wakeup_mhi &&
test_bit(ATH11K_PCI_FLAG_INIT_DONE, &ab_pci->flags) &&
- offset >= ACCESS_ALWAYS_OFF)
+ offset >= ACCESS_ALWAYS_OFF &&
+ !ret)
mhi_device_put(ab_pci->mhi_ctrl->mhi_dev);
}
@@ -230,6 +232,7 @@ u32 ath11k_pci_read32(struct ath11k_base *ab, u32 offset)
{
struct ath11k_pci *ab_pci = ath11k_pci_priv(ab);
u32 val, window_start;
+ int ret = 0;
/* for offset beyond BAR + 4K - 32, may
* need to wakeup MHI to access.
@@ -237,7 +240,7 @@ u32 ath11k_pci_read32(struct ath11k_base *ab, u32 offset)
if (ab->hw_params.wakeup_mhi &&
test_bit(ATH11K_PCI_FLAG_INIT_DONE, &ab_pci->flags) &&
offset >= ACCESS_ALWAYS_OFF)
- mhi_device_get_sync(ab_pci->mhi_ctrl->mhi_dev);
+ ret = mhi_device_get_sync(ab_pci->mhi_ctrl->mhi_dev);
if (offset < WINDOW_START) {
val = ioread32(ab->mem + offset);
@@ -261,7 +264,8 @@ u32 ath11k_pci_read32(struct ath11k_base *ab, u32 offset)
if (ab->hw_params.wakeup_mhi &&
test_bit(ATH11K_PCI_FLAG_INIT_DONE, &ab_pci->flags) &&
- offset >= ACCESS_ALWAYS_OFF)
+ offset >= ACCESS_ALWAYS_OFF &&
+ !ret)
mhi_device_put(ab_pci->mhi_ctrl->mhi_dev);
return val;
--
2.31.1
Currently ath11k has device recovery logic, it is introduced by this
patch "ath11k: Add support for subsystem recovery" which is upstream
by https://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git/commit/?h=ath11k-bringup&id=3a7b4838b6f6f234239f263ef3dc02e612a083ad.
The patch is for AHB devices such as IPQ8074, it has remote proc module
which is used to download the firmware and boots the processor which
firmware is running on. If firmware crashed, remote proc module will
detect it and download and boot firmware again. Below command will
trigger a firmware crash, and then user can test feature of device
recovery.
Test command:
echo assert > /sys/kernel/debug/ath11k/qca6390\ hw2.0/simulate_fw_crash
echo assert > /sys/kernel/debug/ath11k/wcn6855\ hw2.0/simulate_fw_crash
Unfortunately, QCA6390 is PCIe bus, it does not have the remote proc
module, it use mhi module to communicate between firmware and ath11k.
So ath11k does not support device recovery for QCA6390 currently.
This patch is to add the extra logic which is different for QCA6390.
When firmware crashed, MHI_CB_EE_RDDM event will be indicate by
firmware and then ath11k_mhi_op_status_cb which is the callback of
mhi_controller will receive the MHI_CB_EE_RDDM event, then ath11k
will start to do recovery process, ath11k_core_reset() calls
ath11k_hif_power_down()/ath11k_hif_power_up(), then the mhi/ath11k
will start to download and boot firmware. There are some logic to
avoid deadloop recovery and two simultaneous recovery operations.
And because it has muti-radios for the soc, so it add some logic
in ath11k_mac_op_reconfig_complete() to make sure all radios has
reconfig complete and then complete the device recovery.
Also it add workqueue_aux, because ab->workqueue is used when receive
ATH11K_QMI_EVENT_FW_READY in recovery process(queue_work(ab->workqueue,
&ab->restart_work)), and ath11k_core_reset will wait for max
ATH11K_RESET_TIMEOUT_HZ for the previous restart_work finished, if
ath11k_core_reset also queued in ab->workqueue, then it will delay
restart_work of previous recovery and lead previous recovery fail.
ath11k recovery success for QCA6390/WCN6855 after apply this patch.
Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03003-QCAHSPSWPL_V1_V2_SILICONZ_LITE-2
Signed-off-by: Wen Gong <[email protected]>
---
drivers/net/wireless/ath/ath11k/core.c | 70 ++++++++++++++++++++++++++
drivers/net/wireless/ath/ath11k/core.h | 13 +++++
drivers/net/wireless/ath/ath11k/mac.c | 18 +++++++
drivers/net/wireless/ath/ath11k/mhi.c | 33 ++++++++++++
4 files changed, 134 insertions(+)
diff --git a/drivers/net/wireless/ath/ath11k/core.c b/drivers/net/wireless/ath/ath11k/core.c
index 7c508e9baa6d..00c83fdb0702 100644
--- a/drivers/net/wireless/ath/ath11k/core.c
+++ b/drivers/net/wireless/ath/ath11k/core.c
@@ -1342,6 +1342,65 @@ static void ath11k_core_restart(struct work_struct *work)
complete(&ab->driver_recovery);
}
+static void ath11k_core_reset(struct work_struct *work)
+{
+ struct ath11k_base *ab = container_of(work, struct ath11k_base, reset_work);
+ int reset_count, fail_cont_count;
+ long time_left;
+
+ if (!(test_bit(ATH11K_FLAG_REGISTERED, &ab->dev_flags))) {
+ ath11k_warn(ab, "ignore reset dev flags 0x%lx\n", ab->dev_flags);
+ return;
+ }
+
+ /* Sometimes the recovery will fail and then the next all recovery fail,
+ * this is to avoid infinite recovery since it can not recovery success.
+ */
+ fail_cont_count = atomic_read(&ab->fail_cont_count);
+
+ if (fail_cont_count >= ATH11K_RESET_MAX_FAIL_COUNT_FINAL)
+ return;
+
+ if (fail_cont_count >= ATH11K_RESET_MAX_FAIL_COUNT_FIRST &&
+ time_before(jiffies, ab->reset_fail_timeout))
+ return;
+
+ reset_count = atomic_inc_return(&ab->reset_count);
+
+ if (reset_count > 1) {
+ /* Sometimes it happened another reset worker before the previous one
+ * completed, then the second reset worker will destroy the previous one,
+ * thus below is to avoid that.
+ */
+ ath11k_warn(ab, "already reseting count %d\n", reset_count);
+
+ reinit_completion(&ab->reset_complete);
+ time_left = wait_for_completion_timeout(&ab->reset_complete,
+ ATH11K_RESET_TIMEOUT_HZ);
+
+ if (time_left) {
+ ath11k_dbg(ab, ATH11K_DBG_BOOT, "to skip reset\n");
+ atomic_dec(&ab->reset_count);
+ return;
+ }
+
+ ab->reset_fail_timeout = jiffies + ATH11K_RESET_FAIL_TIMEOUT_HZ;
+ /* Record the continuous recovery fail count when recovery failed*/
+ fail_cont_count = atomic_inc_return(&ab->fail_cont_count);
+ }
+
+ ath11k_dbg(ab, ATH11K_DBG_BOOT, "reset starting\n");
+
+ ab->is_reset = true;
+ atomic_set(&ab->recovery_count, 0);
+
+ ath11k_hif_power_down(ab);
+ ath11k_qmi_free_resource(ab);
+ ath11k_hif_power_up(ab);
+
+ ath11k_dbg(ab, ATH11K_DBG_BOOT, "reset started\n");
+}
+
static int ath11k_init_hw_params(struct ath11k_base *ab)
{
const struct ath11k_hw_params *hw_params = NULL;
@@ -1411,6 +1470,9 @@ EXPORT_SYMBOL(ath11k_core_deinit);
void ath11k_core_free(struct ath11k_base *ab)
{
+ flush_workqueue(ab->workqueue_aux);
+ destroy_workqueue(ab->workqueue_aux);
+
flush_workqueue(ab->workqueue);
destroy_workqueue(ab->workqueue);
@@ -1434,9 +1496,14 @@ struct ath11k_base *ath11k_core_alloc(struct device *dev, size_t priv_size,
if (!ab->workqueue)
goto err_sc_free;
+ ab->workqueue_aux = create_singlethread_workqueue("ath11k_aux_wq");
+ if (!ab->workqueue_aux)
+ goto err_free_wq;
+
mutex_init(&ab->core_lock);
spin_lock_init(&ab->base_lock);
mutex_init(&ab->vdev_id_11d_lock);
+ init_completion(&ab->reset_complete);
INIT_LIST_HEAD(&ab->peers);
init_waitqueue_head(&ab->peer_mapping_wq);
@@ -1445,6 +1512,7 @@ struct ath11k_base *ath11k_core_alloc(struct device *dev, size_t priv_size,
INIT_WORK(&ab->restart_work, ath11k_core_restart);
INIT_WORK(&ab->update_11d_work, ath11k_update_11d);
INIT_WORK(&ab->rfkill_work, ath11k_rfkill_work);
+ INIT_WORK(&ab->reset_work, ath11k_core_reset);
timer_setup(&ab->rx_replenish_retry, ath11k_ce_rx_replenish_retry, 0);
init_completion(&ab->htc_suspend);
init_completion(&ab->wow.wakeup_completed);
@@ -1455,6 +1523,8 @@ struct ath11k_base *ath11k_core_alloc(struct device *dev, size_t priv_size,
return ab;
+err_free_wq:
+ destroy_workqueue(ab->workqueue);
err_sc_free:
kfree(ab);
return NULL;
diff --git a/drivers/net/wireless/ath/ath11k/core.h b/drivers/net/wireless/ath/ath11k/core.h
index d2fc7a7a98f4..c85301e3609b 100644
--- a/drivers/net/wireless/ath/ath11k/core.h
+++ b/drivers/net/wireless/ath/ath11k/core.h
@@ -39,6 +39,10 @@
extern unsigned int ath11k_frame_mode;
#define ATH11K_MON_TIMER_INTERVAL 10
+#define ATH11K_RESET_TIMEOUT_HZ (20 * HZ)
+#define ATH11K_RESET_MAX_FAIL_COUNT_FIRST 3
+#define ATH11K_RESET_MAX_FAIL_COUNT_FINAL 5
+#define ATH11K_RESET_FAIL_TIMEOUT_HZ (20 * HZ)
enum ath11k_supported_bw {
ATH11K_BW_20 = 0,
@@ -787,6 +791,15 @@ struct ath11k_base {
struct work_struct restart_work;
struct work_struct update_11d_work;
u8 new_alpha2[3];
+ struct workqueue_struct *workqueue_aux;
+ struct work_struct reset_work;
+ atomic_t reset_count;
+ atomic_t recovery_count;
+ bool is_reset;
+ struct completion reset_complete;
+ /* continuous recovery fail count */
+ atomic_t fail_cont_count;
+ unsigned long reset_fail_timeout;
struct {
/* protected by data_lock */
u32 fw_crash_counter;
diff --git a/drivers/net/wireless/ath/ath11k/mac.c b/drivers/net/wireless/ath/ath11k/mac.c
index d5b83f90d27a..5c62faf359a9 100644
--- a/drivers/net/wireless/ath/ath11k/mac.c
+++ b/drivers/net/wireless/ath/ath11k/mac.c
@@ -7881,6 +7881,8 @@ ath11k_mac_op_reconfig_complete(struct ieee80211_hw *hw,
enum ieee80211_reconfig_type reconfig_type)
{
struct ath11k *ar = hw->priv;
+ struct ath11k_base *ab = ar->ab;
+ int recovery_count;
if (reconfig_type != IEEE80211_RECONFIG_TYPE_RESTART)
return;
@@ -7892,6 +7894,22 @@ ath11k_mac_op_reconfig_complete(struct ieee80211_hw *hw,
ar->pdev->pdev_id);
ar->state = ATH11K_STATE_ON;
ieee80211_wake_queues(ar->hw);
+
+ if (ab->is_reset) {
+ recovery_count = atomic_inc_return(&ab->recovery_count);
+ ath11k_dbg(ab, ATH11K_DBG_BOOT,
+ "recovery count %d\n", recovery_count);
+ /* When there are multiple radios in an SOC,
+ * the recovery has to be done for each radio
+ */
+ if (recovery_count == ab->num_radios) {
+ atomic_dec(&ab->reset_count);
+ complete(&ab->reset_complete);
+ ab->is_reset = false;
+ atomic_set(&ab->fail_cont_count, 0);
+ ath11k_dbg(ab, ATH11K_DBG_BOOT, "reset success\n");
+ }
+ }
}
mutex_unlock(&ar->conf_mutex);
diff --git a/drivers/net/wireless/ath/ath11k/mhi.c b/drivers/net/wireless/ath/ath11k/mhi.c
index fc3524e83e52..61d83be4841f 100644
--- a/drivers/net/wireless/ath/ath11k/mhi.c
+++ b/drivers/net/wireless/ath/ath11k/mhi.c
@@ -292,15 +292,48 @@ static void ath11k_mhi_op_runtime_put(struct mhi_controller *mhi_cntrl)
{
}
+static char *ath11k_mhi_op_callback_to_str(enum mhi_callback reason)
+{
+ switch (reason) {
+ case MHI_CB_IDLE:
+ return "MHI_CB_IDLE";
+ case MHI_CB_PENDING_DATA:
+ return "MHI_CB_PENDING_DATA";
+ case MHI_CB_LPM_ENTER:
+ return "MHI_CB_LPM_ENTER";
+ case MHI_CB_LPM_EXIT:
+ return "MHI_CB_LPM_EXIT";
+ case MHI_CB_EE_RDDM:
+ return "MHI_CB_EE_RDDM";
+ case MHI_CB_EE_MISSION_MODE:
+ return "MHI_CB_EE_MISSION_MODE";
+ case MHI_CB_SYS_ERROR:
+ return "MHI_CB_SYS_ERROR";
+ case MHI_CB_FATAL_ERROR:
+ return "MHI_CB_FATAL_ERROR";
+ case MHI_CB_BW_REQ:
+ return "MHI_CB_BW_REQ";
+ default:
+ return "UNKNOWN";
+ }
+};
+
static void ath11k_mhi_op_status_cb(struct mhi_controller *mhi_cntrl,
enum mhi_callback cb)
{
struct ath11k_base *ab = dev_get_drvdata(mhi_cntrl->cntrl_dev);
+ ath11k_dbg(ab, ATH11K_DBG_BOOT, "mhi notify status reason %s\n",
+ ath11k_mhi_op_callback_to_str(cb));
+
switch (cb) {
case MHI_CB_SYS_ERROR:
ath11k_warn(ab, "firmware crashed: MHI_CB_SYS_ERROR\n");
break;
+ case MHI_CB_EE_RDDM:
+ if (!(test_bit(ATH11K_FLAG_UNREGISTERING, &ab->dev_flags)))
+ queue_work(ab->workqueue_aux, &ab->reset_work);
+ break;
default:
break;
}
--
2.31.1
ieee80211_reconfig() of mac80211 is the main function for recovery of
each ieee80211_hw and ath11k, and ath11k_core_reconfigure_on_crash()
is the main function for recovery of ath11k_base, it has more than
one ieee80211_hw and ath11k for each ath11k_base, so it need to add
synchronization between them, otherwise it has many issue.
For example, when ath11k_core_reconfigure_on_crash() is not complete,
mac80211 send a hw scan request to ath11k, it leads firmware crash,
because firmware has not been initialized at that moment, firmware
is only finished downloaded and loaded, it can not receive scan
command.
Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03003-QCAHSPSWPL_V1_V2_SILICONZ_LITE-2
Signed-off-by: Wen Gong <[email protected]>
---
drivers/net/wireless/ath/ath11k/core.c | 51 ++++++++++++++++++++++----
drivers/net/wireless/ath/ath11k/core.h | 5 +++
drivers/net/wireless/ath/ath11k/mac.c | 22 +++++++++++
3 files changed, 70 insertions(+), 8 deletions(-)
diff --git a/drivers/net/wireless/ath/ath11k/core.c b/drivers/net/wireless/ath/ath11k/core.c
index 00c83fdb0702..8ce93fd9e1e6 100644
--- a/drivers/net/wireless/ath/ath11k/core.c
+++ b/drivers/net/wireless/ath/ath11k/core.c
@@ -1263,12 +1263,11 @@ static void ath11k_update_11d(struct work_struct *work)
}
}
-static void ath11k_core_restart(struct work_struct *work)
+static void ath11k_core_pre_reconfigure_recovery(struct ath11k_base *ab)
{
- struct ath11k_base *ab = container_of(work, struct ath11k_base, restart_work);
struct ath11k *ar;
struct ath11k_pdev *pdev;
- int i, ret = 0;
+ int i;
spin_lock_bh(&ab->base_lock);
ab->stats.fw_crash_counter++;
@@ -1301,12 +1300,13 @@ static void ath11k_core_restart(struct work_struct *work)
wake_up(&ab->wmi_ab.tx_credits_wq);
wake_up(&ab->peer_mapping_wq);
+}
- ret = ath11k_core_reconfigure_on_crash(ab);
- if (ret) {
- ath11k_err(ab, "failed to reconfigure driver on crash recovery\n");
- return;
- }
+static void ath11k_core_post_reconfigure_recovery(struct ath11k_base *ab)
+{
+ struct ath11k *ar;
+ struct ath11k_pdev *pdev;
+ int i;
for (i = 0; i < ab->num_radios; i++) {
pdev = &ab->pdevs[i];
@@ -1342,6 +1342,27 @@ static void ath11k_core_restart(struct work_struct *work)
complete(&ab->driver_recovery);
}
+static void ath11k_core_restart(struct work_struct *work)
+{
+ struct ath11k_base *ab = container_of(work, struct ath11k_base, restart_work);
+ int ret;
+
+ if (!ab->is_reset)
+ ath11k_core_pre_reconfigure_recovery(ab);
+
+ ret = ath11k_core_reconfigure_on_crash(ab);
+ if (ret) {
+ ath11k_err(ab, "failed to reconfigure driver on crash recovery\n");
+ return;
+ }
+
+ if (ab->is_reset)
+ complete_all(&ab->reconfigure_complete);
+
+ if (!ab->is_reset)
+ ath11k_core_post_reconfigure_recovery(ab);
+}
+
static void ath11k_core_reset(struct work_struct *work)
{
struct ath11k_base *ab = container_of(work, struct ath11k_base, reset_work);
@@ -1393,6 +1414,18 @@ static void ath11k_core_reset(struct work_struct *work)
ab->is_reset = true;
atomic_set(&ab->recovery_count, 0);
+ reinit_completion(&ab->recovery_start);
+ atomic_set(&ab->recovery_start_count, 0);
+
+ ath11k_core_pre_reconfigure_recovery(ab);
+
+ reinit_completion(&ab->reconfigure_complete);
+ ath11k_core_post_reconfigure_recovery(ab);
+
+ ath11k_dbg(ab, ATH11K_DBG_BOOT, "waiting recovery start...\n");
+
+ time_left = wait_for_completion_timeout(&ab->recovery_start,
+ ATH11K_RECOVER_START_TIMEOUT_HZ);
ath11k_hif_power_down(ab);
ath11k_qmi_free_resource(ab);
@@ -1504,6 +1537,8 @@ struct ath11k_base *ath11k_core_alloc(struct device *dev, size_t priv_size,
spin_lock_init(&ab->base_lock);
mutex_init(&ab->vdev_id_11d_lock);
init_completion(&ab->reset_complete);
+ init_completion(&ab->reconfigure_complete);
+ init_completion(&ab->recovery_start);
INIT_LIST_HEAD(&ab->peers);
init_waitqueue_head(&ab->peer_mapping_wq);
diff --git a/drivers/net/wireless/ath/ath11k/core.h b/drivers/net/wireless/ath/ath11k/core.h
index c85301e3609b..7aa498daf07b 100644
--- a/drivers/net/wireless/ath/ath11k/core.h
+++ b/drivers/net/wireless/ath/ath11k/core.h
@@ -43,6 +43,8 @@ extern unsigned int ath11k_frame_mode;
#define ATH11K_RESET_MAX_FAIL_COUNT_FIRST 3
#define ATH11K_RESET_MAX_FAIL_COUNT_FINAL 5
#define ATH11K_RESET_FAIL_TIMEOUT_HZ (20 * HZ)
+#define ATH11K_RECONFIGURE_TIMEOUT_HZ (10 * HZ)
+#define ATH11K_RECOVER_START_TIMEOUT_HZ (20 * HZ)
enum ath11k_supported_bw {
ATH11K_BW_20 = 0,
@@ -795,8 +797,11 @@ struct ath11k_base {
struct work_struct reset_work;
atomic_t reset_count;
atomic_t recovery_count;
+ atomic_t recovery_start_count;
bool is_reset;
struct completion reset_complete;
+ struct completion reconfigure_complete;
+ struct completion recovery_start;
/* continuous recovery fail count */
atomic_t fail_cont_count;
unsigned long reset_fail_timeout;
diff --git a/drivers/net/wireless/ath/ath11k/mac.c b/drivers/net/wireless/ath/ath11k/mac.c
index 5c62faf359a9..c9524e417cf6 100644
--- a/drivers/net/wireless/ath/ath11k/mac.c
+++ b/drivers/net/wireless/ath/ath11k/mac.c
@@ -5726,6 +5726,27 @@ static int ath11k_mac_config_mon_status_default(struct ath11k *ar, bool enable)
return ret;
}
+static void ath11k_mac_wait_reconfigure(struct ath11k_base *ab)
+{
+ int recovery_start_count;
+
+ if (!ab->is_reset)
+ return;
+
+ recovery_start_count = atomic_inc_return(&ab->recovery_start_count);
+ ath11k_dbg(ab, ATH11K_DBG_MAC, "recovery start count %d\n", recovery_start_count);
+
+ if (recovery_start_count == ab->num_radios) {
+ complete(&ab->recovery_start);
+ ath11k_dbg(ab, ATH11K_DBG_MAC, "recovery started success\n");
+ }
+
+ ath11k_dbg(ab, ATH11K_DBG_MAC, "waiting reconfigure...\n");
+
+ wait_for_completion_timeout(&ab->reconfigure_complete,
+ ATH11K_RECONFIGURE_TIMEOUT_HZ);
+}
+
static int ath11k_mac_op_start(struct ieee80211_hw *hw)
{
struct ath11k *ar = hw->priv;
@@ -5742,6 +5763,7 @@ static int ath11k_mac_op_start(struct ieee80211_hw *hw)
break;
case ATH11K_STATE_RESTARTING:
ar->state = ATH11K_STATE_RESTARTED;
+ ath11k_mac_wait_reconfigure(ab);
break;
case ATH11K_STATE_RESTARTED:
case ATH11K_STATE_WEDGED:
--
2.31.1
On Fri, Feb 25, 2022 at 03:45:45AM -0500, Wen Gong wrote:
> Currently ath11k has device recovery logic, it is introduced by this
> patch "ath11k: Add support for subsystem recovery" which is upstream
> by https://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git/commit/?h=ath11k-bringup&id=3a7b4838b6f6f234239f263ef3dc02e612a083ad.
>
> The patch is for AHB devices such as IPQ8074, it has remote proc module
> which is used to download the firmware and boots the processor which
> firmware is running on. If firmware crashed, remote proc module will
> detect it and download and boot firmware again. Below command will
> trigger a firmware crash, and then user can test feature of device
> recovery.
>
> Test command:
> echo assert > /sys/kernel/debug/ath11k/qca6390\ hw2.0/simulate_fw_crash
> echo assert > /sys/kernel/debug/ath11k/wcn6855\ hw2.0/simulate_fw_crash
>
> Unfortunately, QCA6390 is PCIe bus, it does not have the remote proc
> module, it use mhi module to communicate between firmware and ath11k.
> So ath11k does not support device recovery for QCA6390 currently.
>
> This patch is to add the extra logic which is different for QCA6390.
> When firmware crashed, MHI_CB_EE_RDDM event will be indicate by
> firmware and then ath11k_mhi_op_status_cb which is the callback of
> mhi_controller will receive the MHI_CB_EE_RDDM event, then ath11k
> will start to do recovery process, ath11k_core_reset() calls
> ath11k_hif_power_down()/ath11k_hif_power_up(), then the mhi/ath11k
> will start to download and boot firmware. There are some logic to
> avoid deadloop recovery and two simultaneous recovery operations.
> And because it has muti-radios for the soc, so it add some logic
> in ath11k_mac_op_reconfig_complete() to make sure all radios has
> reconfig complete and then complete the device recovery.
>
> Also it add workqueue_aux, because ab->workqueue is used when receive
> ATH11K_QMI_EVENT_FW_READY in recovery process(queue_work(ab->workqueue,
> &ab->restart_work)), and ath11k_core_reset will wait for max
> ATH11K_RESET_TIMEOUT_HZ for the previous restart_work finished, if
> ath11k_core_reset also queued in ab->workqueue, then it will delay
> restart_work of previous recovery and lead previous recovery fail.
>
> ath11k recovery success for QCA6390/WCN6855 after apply this patch.
>
> Tested-on: QCA6390 hw2.0 PCI WLAN.HST.1.0.1-01740-QCAHSTSWPLZ_V2_TO_X86-1
> Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03003-QCAHSPSWPL_V1_V2_SILICONZ_LITE-2
>
> Signed-off-by: Wen Gong <[email protected]>
> ---
> drivers/net/wireless/ath/ath11k/core.c | 70 ++++++++++++++++++++++++++
> drivers/net/wireless/ath/ath11k/core.h | 13 +++++
> drivers/net/wireless/ath/ath11k/mac.c | 18 +++++++
> drivers/net/wireless/ath/ath11k/mhi.c | 33 ++++++++++++
> 4 files changed, 134 insertions(+)
>
> diff --git a/drivers/net/wireless/ath/ath11k/core.c b/drivers/net/wireless/ath/ath11k/core.c
> index 7c508e9baa6d..00c83fdb0702 100644
> --- a/drivers/net/wireless/ath/ath11k/core.c
> +++ b/drivers/net/wireless/ath/ath11k/core.c
> @@ -1342,6 +1342,65 @@ static void ath11k_core_restart(struct work_struct *work)
> complete(&ab->driver_recovery);
> }
>
> +static void ath11k_core_reset(struct work_struct *work)
> +{
> + struct ath11k_base *ab = container_of(work, struct ath11k_base, reset_work);
> + int reset_count, fail_cont_count;
> + long time_left;
> +
> + if (!(test_bit(ATH11K_FLAG_REGISTERED, &ab->dev_flags))) {
> + ath11k_warn(ab, "ignore reset dev flags 0x%lx\n", ab->dev_flags);
> + return;
> + }
> +
> + /* Sometimes the recovery will fail and then the next all recovery fail,
> + * this is to avoid infinite recovery since it can not recovery success.
> + */
> + fail_cont_count = atomic_read(&ab->fail_cont_count);
> +
> + if (fail_cont_count >= ATH11K_RESET_MAX_FAIL_COUNT_FINAL)
> + return;
> +
> + if (fail_cont_count >= ATH11K_RESET_MAX_FAIL_COUNT_FIRST &&
> + time_before(jiffies, ab->reset_fail_timeout))
> + return;
> +
> + reset_count = atomic_inc_return(&ab->reset_count);
> +
> + if (reset_count > 1) {
> + /* Sometimes it happened another reset worker before the previous one
> + * completed, then the second reset worker will destroy the previous one,
> + * thus below is to avoid that.
> + */
> + ath11k_warn(ab, "already reseting count %d\n", reset_count);
> +
> + reinit_completion(&ab->reset_complete);
> + time_left = wait_for_completion_timeout(&ab->reset_complete,
> + ATH11K_RESET_TIMEOUT_HZ);
> +
> + if (time_left) {
> + ath11k_dbg(ab, ATH11K_DBG_BOOT, "to skip reset\n");
> + atomic_dec(&ab->reset_count);
> + return;
> + }
> +
> + ab->reset_fail_timeout = jiffies + ATH11K_RESET_FAIL_TIMEOUT_HZ;
> + /* Record the continuous recovery fail count when recovery failed*/
> + fail_cont_count = atomic_inc_return(&ab->fail_cont_count);
This results in a compiler warning because fail_cont_count is not used
afterwards. I would suggest to replace the statement with
atomic_inc(&ab->fail_cont_count);
Guenter
> + }
> +
> + ath11k_dbg(ab, ATH11K_DBG_BOOT, "reset starting\n");
> +
> + ab->is_reset = true;
> + atomic_set(&ab->recovery_count, 0);
> +
> + ath11k_hif_power_down(ab);
> + ath11k_qmi_free_resource(ab);
> + ath11k_hif_power_up(ab);
> +
> + ath11k_dbg(ab, ATH11K_DBG_BOOT, "reset started\n");
> +}
> +
> static int ath11k_init_hw_params(struct ath11k_base *ab)
> {
> const struct ath11k_hw_params *hw_params = NULL;
> @@ -1411,6 +1470,9 @@ EXPORT_SYMBOL(ath11k_core_deinit);
>
> void ath11k_core_free(struct ath11k_base *ab)
> {
> + flush_workqueue(ab->workqueue_aux);
> + destroy_workqueue(ab->workqueue_aux);
> +
> flush_workqueue(ab->workqueue);
> destroy_workqueue(ab->workqueue);
>
> @@ -1434,9 +1496,14 @@ struct ath11k_base *ath11k_core_alloc(struct device *dev, size_t priv_size,
> if (!ab->workqueue)
> goto err_sc_free;
>
> + ab->workqueue_aux = create_singlethread_workqueue("ath11k_aux_wq");
> + if (!ab->workqueue_aux)
> + goto err_free_wq;
> +
> mutex_init(&ab->core_lock);
> spin_lock_init(&ab->base_lock);
> mutex_init(&ab->vdev_id_11d_lock);
> + init_completion(&ab->reset_complete);
>
> INIT_LIST_HEAD(&ab->peers);
> init_waitqueue_head(&ab->peer_mapping_wq);
> @@ -1445,6 +1512,7 @@ struct ath11k_base *ath11k_core_alloc(struct device *dev, size_t priv_size,
> INIT_WORK(&ab->restart_work, ath11k_core_restart);
> INIT_WORK(&ab->update_11d_work, ath11k_update_11d);
> INIT_WORK(&ab->rfkill_work, ath11k_rfkill_work);
> + INIT_WORK(&ab->reset_work, ath11k_core_reset);
> timer_setup(&ab->rx_replenish_retry, ath11k_ce_rx_replenish_retry, 0);
> init_completion(&ab->htc_suspend);
> init_completion(&ab->wow.wakeup_completed);
> @@ -1455,6 +1523,8 @@ struct ath11k_base *ath11k_core_alloc(struct device *dev, size_t priv_size,
>
> return ab;
>
> +err_free_wq:
> + destroy_workqueue(ab->workqueue);
> err_sc_free:
> kfree(ab);
> return NULL;
> diff --git a/drivers/net/wireless/ath/ath11k/core.h b/drivers/net/wireless/ath/ath11k/core.h
> index d2fc7a7a98f4..c85301e3609b 100644
> --- a/drivers/net/wireless/ath/ath11k/core.h
> +++ b/drivers/net/wireless/ath/ath11k/core.h
> @@ -39,6 +39,10 @@
> extern unsigned int ath11k_frame_mode;
>
> #define ATH11K_MON_TIMER_INTERVAL 10
> +#define ATH11K_RESET_TIMEOUT_HZ (20 * HZ)
> +#define ATH11K_RESET_MAX_FAIL_COUNT_FIRST 3
> +#define ATH11K_RESET_MAX_FAIL_COUNT_FINAL 5
> +#define ATH11K_RESET_FAIL_TIMEOUT_HZ (20 * HZ)
>
> enum ath11k_supported_bw {
> ATH11K_BW_20 = 0,
> @@ -787,6 +791,15 @@ struct ath11k_base {
> struct work_struct restart_work;
> struct work_struct update_11d_work;
> u8 new_alpha2[3];
> + struct workqueue_struct *workqueue_aux;
> + struct work_struct reset_work;
> + atomic_t reset_count;
> + atomic_t recovery_count;
> + bool is_reset;
> + struct completion reset_complete;
> + /* continuous recovery fail count */
> + atomic_t fail_cont_count;
> + unsigned long reset_fail_timeout;
> struct {
> /* protected by data_lock */
> u32 fw_crash_counter;
> diff --git a/drivers/net/wireless/ath/ath11k/mac.c b/drivers/net/wireless/ath/ath11k/mac.c
> index d5b83f90d27a..5c62faf359a9 100644
> --- a/drivers/net/wireless/ath/ath11k/mac.c
> +++ b/drivers/net/wireless/ath/ath11k/mac.c
> @@ -7881,6 +7881,8 @@ ath11k_mac_op_reconfig_complete(struct ieee80211_hw *hw,
> enum ieee80211_reconfig_type reconfig_type)
> {
> struct ath11k *ar = hw->priv;
> + struct ath11k_base *ab = ar->ab;
> + int recovery_count;
>
> if (reconfig_type != IEEE80211_RECONFIG_TYPE_RESTART)
> return;
> @@ -7892,6 +7894,22 @@ ath11k_mac_op_reconfig_complete(struct ieee80211_hw *hw,
> ar->pdev->pdev_id);
> ar->state = ATH11K_STATE_ON;
> ieee80211_wake_queues(ar->hw);
> +
> + if (ab->is_reset) {
> + recovery_count = atomic_inc_return(&ab->recovery_count);
> + ath11k_dbg(ab, ATH11K_DBG_BOOT,
> + "recovery count %d\n", recovery_count);
> + /* When there are multiple radios in an SOC,
> + * the recovery has to be done for each radio
> + */
> + if (recovery_count == ab->num_radios) {
> + atomic_dec(&ab->reset_count);
> + complete(&ab->reset_complete);
> + ab->is_reset = false;
> + atomic_set(&ab->fail_cont_count, 0);
> + ath11k_dbg(ab, ATH11K_DBG_BOOT, "reset success\n");
> + }
> + }
> }
>
> mutex_unlock(&ar->conf_mutex);
> diff --git a/drivers/net/wireless/ath/ath11k/mhi.c b/drivers/net/wireless/ath/ath11k/mhi.c
> index fc3524e83e52..61d83be4841f 100644
> --- a/drivers/net/wireless/ath/ath11k/mhi.c
> +++ b/drivers/net/wireless/ath/ath11k/mhi.c
> @@ -292,15 +292,48 @@ static void ath11k_mhi_op_runtime_put(struct mhi_controller *mhi_cntrl)
> {
> }
>
> +static char *ath11k_mhi_op_callback_to_str(enum mhi_callback reason)
> +{
> + switch (reason) {
> + case MHI_CB_IDLE:
> + return "MHI_CB_IDLE";
> + case MHI_CB_PENDING_DATA:
> + return "MHI_CB_PENDING_DATA";
> + case MHI_CB_LPM_ENTER:
> + return "MHI_CB_LPM_ENTER";
> + case MHI_CB_LPM_EXIT:
> + return "MHI_CB_LPM_EXIT";
> + case MHI_CB_EE_RDDM:
> + return "MHI_CB_EE_RDDM";
> + case MHI_CB_EE_MISSION_MODE:
> + return "MHI_CB_EE_MISSION_MODE";
> + case MHI_CB_SYS_ERROR:
> + return "MHI_CB_SYS_ERROR";
> + case MHI_CB_FATAL_ERROR:
> + return "MHI_CB_FATAL_ERROR";
> + case MHI_CB_BW_REQ:
> + return "MHI_CB_BW_REQ";
> + default:
> + return "UNKNOWN";
> + }
> +};
> +
> static void ath11k_mhi_op_status_cb(struct mhi_controller *mhi_cntrl,
> enum mhi_callback cb)
> {
> struct ath11k_base *ab = dev_get_drvdata(mhi_cntrl->cntrl_dev);
>
> + ath11k_dbg(ab, ATH11K_DBG_BOOT, "mhi notify status reason %s\n",
> + ath11k_mhi_op_callback_to_str(cb));
> +
> switch (cb) {
> case MHI_CB_SYS_ERROR:
> ath11k_warn(ab, "firmware crashed: MHI_CB_SYS_ERROR\n");
> break;
> + case MHI_CB_EE_RDDM:
> + if (!(test_bit(ATH11K_FLAG_UNREGISTERING, &ab->dev_flags)))
> + queue_work(ab->workqueue_aux, &ab->reset_work);
> + break;
> default:
> break;
> }
Add hw-restart to directly restart wlan. Like UTF mode start it will
restart hardware and download firmware again.
Usage:
1. Run command:
echo hw-restart > /sys/kernel/debug/ath11k/qca6390\ hw2.0/simulate_fw_crash
echo hw-restart > /sys/kernel/debug/ath11k/wcn6855\ hw2.0/simulate_fw_crash
2. wlan will be restart and do recovery process and success.
Tested-on: WCN6855 hw2.0 PCI WLAN.HSP.1.1-03003-QCAHSPSWPL_V1_V2_SILICONZ_LITE-2
Signed-off-by: Wen Gong <[email protected]>
---
drivers/net/wireless/ath/ath11k/debugfs.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/net/wireless/ath/ath11k/debugfs.c b/drivers/net/wireless/ath/ath11k/debugfs.c
index c0ebdb1262f0..7375defb0f23 100644
--- a/drivers/net/wireless/ath/ath11k/debugfs.c
+++ b/drivers/net/wireless/ath/ath11k/debugfs.c
@@ -557,6 +557,10 @@ static ssize_t ath11k_write_simulate_fw_crash(struct file *file,
ret = ath11k_wmi_force_fw_hang_cmd(ar,
ATH11K_WMI_FW_HANG_ASSERT_TYPE,
ATH11K_WMI_FW_HANG_DELAY);
+ } else if (!strcmp(buf, "hw-restart")) {
+ ath11k_info(ab, "user requested hw restart\n");
+ queue_work(ab->workqueue_aux, &ab->reset_work);
+ ret = 0;
} else {
ret = -EINVAL;
goto exit;
--
2.31.1