Merge pull request #18822 from jue89/feature/at86rf2xx_ztimer

drivers/at86rf2xx: migrate to ztimer
RIOT-OS · Oct 31, 2022 · d704a1a · d704a1a
2 parents a21f05a + 782409e
commit d704a1a
Show file tree

Hide file tree

Showing 7 changed files with 13 additions and 11 deletions.
diff --git a/drivers/at86rf2xx/Kconfig b/drivers/at86rf2xx/Kconfig
@@ -14,7 +14,8 @@ menuconfig MODULE_AT86RF2XX
     select MODULE_NETDEV
     select MODULE_NETDEV_IEEE802154
     select MODULE_NETDEV_LEGACY_API
-    select MODULE_XTIMER
+    select MODULE_ZTIMER
+    select MODULE_ZTIMER_USEC
 
 if MODULE_AT86RF2XX
 

diff --git a/drivers/at86rf2xx/Makefile.dep b/drivers/at86rf2xx/Makefile.dep
@@ -1,7 +1,8 @@
 DEFAULT_MODULE += auto_init_at86rf2xx
 DEFAULT_MODULE += netdev_ieee802154_oqpsk
 
-USEMODULE += xtimer
+USEMODULE += ztimer
+USEMODULE += ztimer_usec
 USEMODULE += ieee802154
 USEMODULE += netdev_ieee802154
 USEMODULE += netdev_legacy_api

diff --git a/drivers/at86rf2xx/aes_spi.c b/drivers/at86rf2xx/aes_spi.c
@@ -18,7 +18,7 @@
 
 #include <assert.h>
 
-#include "xtimer.h"
+#include "ztimer.h"
 #include "periph/spi.h"
 #include "at86rf2xx_aes.h"
 
@@ -55,7 +55,7 @@ uint8_t _aes_status(at86rf2xx_t *dev)
 static inline
 void _aes_wait_for_result(at86rf2xx_t *dev)
 {
-    xtimer_usleep(AT86RF2XX_AES_DELAY_US);
+    ztimer_sleep(ZTIMER_USEC, AT86RF2XX_AES_DELAY_US);
     uint8_t status = _aes_status(dev);
     /*
         If this assert fires, there probably is an implementation error.

diff --git a/drivers/at86rf2xx/at86rf2xx_internal.c b/drivers/at86rf2xx/at86rf2xx_internal.c
@@ -21,7 +21,7 @@
  * @}
  */
 
-#include "xtimer.h"
+#include "ztimer.h"
 #include "at86rf2xx_internal.h"
 #include "at86rf2xx_registers.h"
 
@@ -127,9 +127,9 @@ void at86rf2xx_assert_awake(at86rf2xx_t *dev)
 #else
         gpio_clear(dev->params.sleep_pin);
 #endif
-        xtimer_usleep(AT86RF2XX_WAKEUP_DELAY);
+        ztimer_sleep(ZTIMER_USEC, AT86RF2XX_WAKEUP_DELAY);
 
-        /* update state: on some platforms, the timer behind xtimer
+        /* update state: on some platforms, the timer behind ztimer
          * may be inaccurate or the radio itself may take longer
          * to wake up due to extra capacitance on the oscillator.
          * Spin until we are actually awake
@@ -149,10 +149,10 @@ void at86rf2xx_hardware_reset(at86rf2xx_t *dev)
     *(AT86RF2XX_REG__TRXPR) |= AT86RF2XX_TRXPR_TRXRST;
 #else
     gpio_clear(dev->params.reset_pin);
-    xtimer_usleep(AT86RF2XX_RESET_PULSE_WIDTH);
+    ztimer_sleep(ZTIMER_USEC, AT86RF2XX_RESET_PULSE_WIDTH);
     gpio_set(dev->params.reset_pin);
 #endif
-    xtimer_usleep(AT86RF2XX_RESET_DELAY);
+    ztimer_sleep(ZTIMER_USEC, AT86RF2XX_RESET_DELAY);
 
     /* update state: if the radio state was P_ON (initialization phase),
      * it remains P_ON. Otherwise, it should go to TRX_OFF

diff --git a/drivers/at86rf2xx/include/at86rf2xx_internal.h b/drivers/at86rf2xx/include/at86rf2xx_internal.h
@@ -65,7 +65,7 @@ extern "C" {
 
 /**
  * @brief   Minimum reset pulse width, refer p.190. We use 62us so
- *          that it is at least one tick on platforms with coarse xtimers
+ *          that it is at least one tick on platforms with coarse ztimers
  */
 #define AT86RF2XX_RESET_PULSE_WIDTH     (62U)
 

diff --git a/sys/net/gnrc/Makefile.dep b/sys/net/gnrc/Makefile.dep
@@ -493,5 +493,6 @@ ifneq (,$(filter gnrc_lwmac,$(USEMODULE)))
   USEMODULE += gnrc_netif
   USEMODULE += gnrc_nettype_lwmac
   USEMODULE += gnrc_mac
+  USEMODULE += xtimer
   FEATURES_REQUIRED += periph_rtt
 endif
diff --git a/tests/driver_at86rf2xx_aes/main.c b/tests/driver_at86rf2xx_aes/main.c
@@ -19,7 +19,6 @@
 #include <stdio.h>
 #include <string.h>
 
-#include "xtimer.h"
 #include "at86rf2xx.h"
 #include "at86rf2xx_params.h"
 #include "at86rf2xx_aes.h"