ath10k: testmode support
Subsystems (2)
v4 cover ath10k: testmode support
Hi,
latest version of my testmode patches. Getting closer, more or less cosmetic changes this time :)
v4:
* fix comment style (Michal)
* add extra checks for at->testmode.utf (Michal)
* add __ath10k_tm_cmd_utf_stop() and call it from
ath10k_testmode_destroy() (Michal)
v3:
* remove false "fall through" comment in ath10k_start() (Michal)
* fix multiline comment style (Michal)
* put testmode interface definitions to testmode_uapi.h (or similar) (Johannes)
* document __set_bit(ar->fw_features) better (Michal)
* ath10k_tm_cmd_utf_start(): return ar->fw_feature value if there's an error
* free firmware in ath10k_tm_cmd_utf_start() when an error happens (Michal)
* ath10k_tm_cmd_utf_start(): move hif_power_down() to err_power_down label
v2:
* check error from ath10k_wmi_alloc_skb() (Michal)
* int err -> int ret (Michal)
* fix conflicts in core.c
* add ar->testmode.utf_monitor to check if ath10k_tm_event_wmi() should deliver WMI events or not
* print mode in ath10k_download_fw() log messages
* set ar->testmode.utf to NULL after it's released (Michal)
* release ar->testmode.utf also in ath10k_tm_cmd_utf_stop()
* rename ath10k_testmode_unregister() to ath10k_testmode_destroy()
* call ath10k_core_stop() in ath10k_testmode_destroy() (Michal)
* document wmi packet format in testmode commands (Michal)
* consume WMI packets so that mac80211 calls are not made in UTF mode (Michal)
* force to use 10x WMI interface with UTF firmware
* API change: remove ATH10K_TM_CMD_EVENT_WMI, use ATH10K_TM_CMD_WMI also for events
* API change: add major and minor version numbers to ATH10K_TM_CMD_GET_VERSION for the user space more easily to detect interface changes
* use conf_mutex in ath10k_tm_cmd_wmi()
* print "UTF firmware started" and "UTF firmware stopped"
* print "testmode %d" in ath10k_print_driver_info()
* don't call ath10k_wait_for_suspend() in UTF mode
---
Kalle Valo (2):
ath10k: make ath10k_wmi_cmd_send() public
ath10k: add testmode
drivers/net/wireless/ath/ath10k/Makefile | 1 drivers/net/wireless/ath/ath10k/core.c | 88 ++++-- drivers/net/wireless/ath/ath10k/core.h | 22 +- drivers/net/wireless/ath/ath10k/debug.c | 5 drivers/net/wireless/ath/ath10k/debug.h | 1 drivers/net/wireless/ath/ath10k/hw.h | 2 drivers/net/wireless/ath/ath10k/mac.c | 9 + drivers/net/wireless/ath/ath10k/testmode.c | 381 ++++++++++++++++++++++++++ drivers/net/wireless/ath/ath10k/testmode.h | 46 +++ drivers/net/wireless/ath/ath10k/testmode_i.h | 70 +++++ drivers/net/wireless/ath/ath10k/wmi.c | 23 +- drivers/net/wireless/ath/ath10k/wmi.h | 4 12 files changed, 619 insertions(+), 33 deletions(-) create mode 100644 drivers/net/wireless/ath/ath10k/testmode.c create mode 100644 drivers/net/wireless/ath/ath10k/testmode.h create mode 100644 drivers/net/wireless/ath/ath10k/testmode_i.h
PATCH v4 1/2 ath10k: make ath10k_wmi_cmd_send() public
We need this function to send wmi packets from testmode.c.
--- drivers/net/wireless/ath/ath10k/wmi.c | 5 ++--- drivers/net/wireless/ath/ath10k/wmi.h | 4 ++++ 2 files changed, 6 insertions(+), 3 deletions(-)
drivers/net/wireless/ath/ath10k/wmi.c +2 -3
@@ -624,7 +624,7 @@ int ath10k_wmi_wait_for_unified_ready(struct ath10k *ar)return ret; } -static struct sk_buff *ath10k_wmi_alloc_skb (struct ath10k *ar,u32 len) +struct sk_buff *ath10k_wmi_alloc_skb (struct ath10k *ar,u32 len) {struct sk_buff *skb;u32 round_len =roundup (len,4 );
@@ -725,8 +725,7 @@ static void ath10k_wmi_op_ep_tx_credits(struct ath10k *ar)wake_up (&ar->wmi.tx_credits_wq); } -static int ath10k_wmi_cmd_send (struct ath10k *ar,struct sk_buff *skb, -u32 cmd_id) +int ath10k_wmi_cmd_send (struct ath10k *ar,struct sk_buff *skb,u32 cmd_id) {int ret = -EOPNOTSUPP ;
drivers/net/wireless/ath/ath10k/wmi.h +4 -0
@@ -4737,6 +4737,10 @@ int ath10k_wmi_wait_for_service_ready(struct ath10k *ar);int ath10k_wmi_wait_for_unified_ready (struct ath10k *ar);int ath10k_wmi_connect (struct ath10k *ar); + +struct sk_buff *ath10k_wmi_alloc_skb (struct ath10k *ar,u32 len); +int ath10k_wmi_cmd_send (struct ath10k *ar,struct sk_buff *skb,u32 cmd_id); +int ath10k_wmi_pdev_set_channel (struct ath10k *ar,const struct wmi_channel_arg *);int ath10k_wmi_pdev_suspend_target (struct ath10k *ar,u32 suspend_opt);
PATCH v4 2/2 ath10k: add testmode
In reply to Kalle ValoAdd testmode interface for starting and using UTF firmware which is used to run factory tests. This is implemented by adding new state ATH10K_STATE_UTF and user space can enable this state with ATH10K_TM_CMD_UTF_START command. To go back to normal mode user space can send ATH10K_TM_CMD_UTF_STOP.
--- drivers/net/wireless/ath/ath10k/Makefile | 1 drivers/net/wireless/ath/ath10k/core.c | 88 ++++-- drivers/net/wireless/ath/ath10k/core.h | 22 +- drivers/net/wireless/ath/ath10k/debug.c | 5 drivers/net/wireless/ath/ath10k/debug.h | 1 drivers/net/wireless/ath/ath10k/hw.h | 2 drivers/net/wireless/ath/ath10k/mac.c | 9 + drivers/net/wireless/ath/ath10k/testmode.c | 381 ++++++++++++++++++++++++++ drivers/net/wireless/ath/ath10k/testmode.h | 46 +++ drivers/net/wireless/ath/ath10k/testmode_i.h | 70 +++++ drivers/net/wireless/ath/ath10k/wmi.c | 18 + 11 files changed, 613 insertions(+), 30 deletions(-) create mode 100644 drivers/net/wireless/ath/ath10k/testmode.c create mode 100644 drivers/net/wireless/ath/ath10k/testmode.h create mode 100644 drivers/net/wireless/ath/ath10k/testmode_i.h
drivers/net/wireless/ath/ath10k/Makefile +1 -0
@@ -11,6 +11,7 @@ ath10k_core-y += mac.o \ bmi.o ath10k_core-$(CONFIG_ATH10K_DEBUGFS) += spectral.o +ath10k_core-$(CONFIG_NL80211_TESTMODE) += testmode.o ath10k_core-$(CONFIG_ATH10K_TRACING) += trace.o obj-$(CONFIG_ATH10K_PCI) += ath10k_pci.o
drivers/net/wireless/ath/ath10k/core.c +62 -26
173 lines not on this page · raw message
drivers/net/wireless/ath/ath10k/core.h +21 -1
@@ -330,6 +330,17 @@ enum ath10k_state {* prevents completion timeouts and makes the driver more responsive to * userspace commands. This is also prevents recursive recovery. */ ATH10K_STATE_WEDGED , + +/* factory tests */ + ATH10K_STATE_UTF, +}; + +enum ath10k_firmware_mode { +/* the default mode, standard 802.11 functionality */ +ATH10K_FIRMWARE_MODE_NORMAL , + +/* factory tests etc */ +ATH10K_FIRMWARE_MODE_UTF , }; enum ath10k_fw_features {
@@ -543,6 +554,15 @@ struct ath10k {struct ath10k_spec_scan config; } spectral; +struct { +/* protected by conf_mutex */ +const struct firmware *utf; + DECLARE_BITMAP(orig_fw_features, ATH10K_FW_FEATURE_COUNT); + +/* protected by data_lock */ +bool utf_monitor; + } testmode; +/* must be last */ u8 drv_priv[0 ]__aligned(sizeof(void *)) ; };
@@ -551,7 +571,7 @@ struct ath10k *ath10k_core_create(size_t priv_size, struct device *dev, const struct ath10k_hif_ops *hif_ops);void ath10k_core_destroy (struct ath10k *ar); -int ath10k_core_start (struct ath10k *ar); +int ath10k_core_start (struct ath10k *ar,enum ath10k_firmware_mode mode);int ath10k_wait_for_suspend (struct ath10k *ar,u32 suspend_opt);void ath10k_core_stop (struct ath10k *ar);int ath10k_core_register (struct ath10k *ar,u32 chip_id);
drivers/net/wireless/ath/ath10k/debug.c +3 -2
@@ -134,11 +134,12 @@ void ath10k_print_driver_info(struct ath10k *ar) ar->fw_api, ar->htt.target_version_major, ar->htt.target_version_minor); -ath10k_info (ar,"debug %d debugfs %d tracing %d dfs %d\n" , +ath10k_info (ar,"debug %d debugfs %d tracing %d dfs %d testmode %d\n" ,config_enabled (CONFIG_ATH10K_DEBUG ),config_enabled (CONFIG_ATH10K_DEBUGFS ),config_enabled (CONFIG_ATH10K_TRACING ), -config_enabled (CONFIG_ATH10K_DFS_CERTIFIED )); +config_enabled (CONFIG_ATH10K_DFS_CERTIFIED ), +config_enabled (CONFIG_NL80211_TESTMODE )); }EXPORT_SYMBOL (ath10k_print_driver_info);
drivers/net/wireless/ath/ath10k/debug.h +1 -0
@@ -34,6 +34,7 @@ enum ath10k_debug_mask {ATH10K_DBG_DATA =0x00000200 ,ATH10K_DBG_BMI =0x00000400 ,ATH10K_DBG_REGULATORY =0x00000800 , +ATH10K_DBG_TESTMODE =0x00001000 , ATH10K_DBG_ANY = 0xffffffff, };
drivers/net/wireless/ath/ath10k/hw.h +2 -0
@@ -36,6 +36,8 @@#define ATH10K_FW_API2_FILE "firmware-2.bin"#define ATH10K_FW_API3_FILE "firmware-3.bin" +#define ATH10K_FW_UTF_FILE "utf.bin" +/* includes also the null byte */ #define ATH10K_FIRMWARE_MAGIC "QCA-ATH10K"
drivers/net/wireless/ath/ath10k/mac.c +8 -1
@@ -26,6 +26,7 @@#include "wmi.h" #include "htt.h" #include "txrx.h" +#include "testmode.h" /**********/ /* Crypto */
@@ -2480,6 +2481,9 @@ static int ath10k_start(struct ieee80211_hw *hw)WARN_ON (1 ); ret = -EINVAL ; goto err; +case ATH10K_STATE_UTF : + ret = -EBUSY ; + goto err; } ret =ath10k_hif_power_up (ar);
@@ -2488,7 +2492,7 @@ static int ath10k_start(struct ieee80211_hw *hw) goto err_off; } - ret =ath10k_core_start (ar); + ret =ath10k_core_start (ar,ATH10K_FIRMWARE_MODE_NORMAL );if (ret) {ath10k_err (ar,"Could not init core: %d\n" , ret); goto err_power_down;
@@ -4456,6 +4460,9 @@ static const struct ieee80211_ops ath10k_ops = { .sta_rc_update = ath10k_sta_rc_update, .get_tsf = ath10k_get_tsf, .ampdu_action = ath10k_ampdu_action, + + CFG80211_TESTMODE_CMD(ath10k_tm_cmd) +#ifdef CONFIG_PM .suspend = ath10k_suspend, .resume = ath10k_resume,
drivers/net/wireless/ath/ath10k/testmode.c new file +381 -0
381 lines not on this page · raw message
drivers/net/wireless/ath/ath10k/testmode.h new file +46 -0
46 lines not on this page · raw message
drivers/net/wireless/ath/ath10k/testmode_i.h new file +70 -0
70 lines not on this page · raw message
drivers/net/wireless/ath/ath10k/wmi.c +18 -0
@@ -23,6 +23,7 @@#include "debug.h" #include "wmi.h" #include "mac.h" +#include "testmode.h" /* MAIN WMI cmd track */ static struct wmi_cmd_map wmi_cmd_map = {
@@ -2479,6 +2480,7 @@ static void ath10k_wmi_10x_process_rx(struct ath10k *ar, struct sk_buff *skb) {struct wmi_cmd_hdr *cmd_hdr;enum wmi_10x_event_id id; +bool consumed; cmd_hdr = (struct wmi_cmd_hdr *)skb->data; id =MS (__le32_to_cpu (cmd_hdr->cmd_id),WMI_CMD_HDR_CMD_ID );
@@ -2488,6 +2490,18 @@ static void ath10k_wmi_10x_process_rx(struct ath10k *ar, struct sk_buff *skb)trace_ath10k_wmi_event (ar, id, skb->data, skb->len); + consumed =ath10k_tm_event_wmi (ar, id, skb); + +/* Ready event must be handled normally also in UTF mode so that we +* know the UTF firmware has booted, others we are just bypass WMI +* events to testmode. +*/ +if (consumed && id !=WMI_10X_READY_EVENTID ) { +ath10k_dbg (ar,ATH10K_DBG_WMI , +"wmi testmode consumed 0x%x\n" , id); + goto out; + } +switch (id) {case WMI_10X_MGMT_RX_EVENTID :ath10k_wmi_event_mgmt_rx (ar, skb);
@@ -2574,11 +2588,15 @@ static void ath10k_wmi_10x_process_rx(struct ath10k *ar, struct sk_buff *skb)case WMI_10X_READY_EVENTID :ath10k_wmi_ready_event_rx (ar, skb);break ; +case WMI_10X_PDEV_UTF_EVENTID : +/* ignore utf events */ +break ;default :ath10k_warn (ar,"Unknown eventid: %d\n" , id);break ; } +out:dev_kfree_skb (skb); }
Re: [PATCH v4 2/2] ath10k: add testmode
On 10 September 2014 17:23, Kalle Valo <[email protected]> wrote:
[...]
Just a small nitpick:
Quoted 6 lines
drivers/net/wireless/ath/ath10k/testmode.c new file +6
-0+static int ath10k_tm_cmd_utf_stop (struct ath10k *ar,struct nlattr *tb[]) +{ +int ret; + +ath10k_dbg (ar,ATH10K_DBG_TESTMODE ,"testmode cmd utf stop\n" ); +
drivers/net/wireless/ath/ath10k/testmode.c
Quoted 10 lines
drivers/net/wireless/ath/ath10k/testmode.c new file +10
-0+mutex_lock (&ar->conf_mutex); + +if (ar->state !=ATH10K_STATE_UTF ) { + ret = -ENETDOWN ; + goto out; + } + +__ath10k_tm_cmd_utf_stop (ar); + + ar->state =ATH10K_STATE_OFF ;// <-- this
I would move it to __ath10k_tm_cmd_utf_stop() itself so that the state is reset in ath10k_testmode_destroy() as well.
Michał
Re: [PATCH v4 2/2] ath10k: add testmode
Quoted 16 lines · Michal Kazior
On 10 September 2014 17:23, Kalle Valo <[email protected]> wrote:
[...]Just a small nitpick:
Quoted 11 lines
(quoted diff) +11
-0+static int ath10k_tm_cmd_utf_stop(struct ath10k *ar, struct nlattr *tb[]) +{ + int ret; + + ath10k_dbg(ar, ATH10K_DBG_TESTMODE, "testmode cmd utf stop\n"); + + mutex_lock(&ar->conf_mutex); + + if (ar->state != ATH10K_STATE_UTF) { + ret = -ENETDOWN; + goto out;
Quoted 8 lines
Review
Quoted 5 lines
(quoted diff) +5
-0+ } + + __ath10k_tm_cmd_utf_stop(ar); + + ar->state = ATH10K_STATE_OFF; // <-- thisI would move it to
__ath10k_tm_cmd_utf_stop()itself so that the state is reset inath10k_testmode_destroy()as well.
Ok. It doesn't make any difference on the functionality, but I'll change it anyway. I folded this patch to the commit:
drivers/net/wireless/ath/ath10k/testmode.c +2 -1
@@ -253,6 +253,8 @@ static void __ath10k_tm_cmd_utf_stop(struct ath10k *ar)release_firmware (ar->testmode.utf); ar->testmode.utf =NULL ; + + ar->state =ATH10K_STATE_OFF ; }static int ath10k_tm_cmd_utf_stop (struct ath10k *ar,struct nlattr *tb[])
@@ -270,7 +272,6 @@ static int ath10k_tm_cmd_utf_stop(struct ath10k *ar, struct nlattr *tb[])__ath10k_tm_cmd_utf_stop (ar); - ar->state =ATH10K_STATE_OFF ; ret =0 ;ath10k_info (ar,"UTF firmware stopped\n" );
Signature · 1 line
Kalle Valo
Kalle Valo
AppliedQuoted 3 lines · Kalle Valo
latest version of my testmode patches. Getting closer, more or less cosmetic changes this time :)
[...]
Quoted 4 lines
Kalle Valo (2):
ath10k: makeath10k_wmi_cmd_send()public
ath10k: add testmode
Applied, with an addition of the small change in patch 2 suggested by Michal.
Signature · 1 line
Kalle Valo
Quoted 4 lines · Kalle Valo
Quoted 2 lines · Kalle Valo
latest version of my testmode patches. Getting closer, more or less cosmetic changes this time :)
While looking through this patchset (sorry it took two months), I have to wonder: what's this utf.bin blob? It seems in order to use testmode, we need this binary blob that I can't seem to find (I looked in linux-firmware and https://github.com/kvalo/ath10k-firmware). If, for whatever reason, this binary blob is not releasable, what's the point of adding in testmode to this driver?
- Pushpal
Quoted 6 lines · Pushpal Sidhu
Quoted 4 lines · Kalle Valo
Quoted 2 lines · Kalle Valo
latest version of my testmode patches. Getting closer, more or less cosmetic changes this time :)
Quoted 6 lines
While looking through this patchset (sorry it took two months), I have to wonder: what's this utf.bin blob? It seems in order to use testmode, we need this binary blob that I can't seem to find (I looked in linux-firmware and https://github.com/kvalo/ath10k-firmware). If, for whatever reason, this binary blob is not releasable, what's the point of adding in testmode to this driver?
testmode is for functionality needed in the factory, it's not meant for normal users. That's why all distros should have CONFIG_NL80211_TESTMODE disabled.
Signature · 1 line
Kalle Valo
Quoted 15 lines · Kalle Valo
Quoted 6 lines · Pushpal Sidhu
Quoted 4 lines · Kalle Valo
Quoted 2 lines · Kalle Valo
latest version of my testmode patches. Getting closer, more or less cosmetic changes this time :)
Quoted 6 lines
While looking through this patchset (sorry it took two months), I have to wonder: what's this utf.bin blob? It seems in order to use testmode, we need this binary blob that I can't seem to find (I looked in linux-firmware and https://github.com/kvalo/ath10k-firmware). If, for whatever reason, this binary blob is not releasable, what's the point of adding in testmode to this driver?
Quoted 3 lines
testmode is for functionality needed in the factory, it's not meant for normal users. That's why all distros should have
CONFIG_NL80211_TESTMODEdisabled.
Kalle,
So what your saying is that 'the factory' (Quallcomm I assume) is using the ath10k driver with some firmware it deems not publicly releasable and the rest of us are paying for the extra lines of code and thrash in the open source driver? This sounds kind of dirty. It seems to me that QCA should either release the firmware or we should remove any special handling in the driver for it.
Is this practice used in other Linux wireless drivers? I've cc'd John to see if he has any comments on that.
Personally, I would rather see QCA document and release the uts firmware which is likely useful for others rather than removing the support for it. We have users often as for a continuous transmit mode switch so that they can get through certification testing. But really, deep down I would much rather QCA simply open up the firmware so that developers that are trying really hard to make this a quality driver can succeed and we can all benefit.
Tim
Quoted 7 lines
--
Kalle Valo
--
To unsubscribe from this list: send the line "unsubscribe linux-wireless" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html