diff --git a/artiq/coredevice/spi.py b/artiq/coredevice/spi.py
index b75de123c..d6ebeb7db 100644
--- a/artiq/coredevice/spi.py
+++ b/artiq/coredevice/spi.py
@@ -1,6 +1,6 @@
 from artiq.language.core import *
-from artiq.language.types import *
-from artiq.coredevice.rt2wb import *
+from artiq.language.units import MHz
+from artiq.coredevice.rt2wb import rt2wb_write, rt2wb_read_sync
 
 
 SPI_DATA_ADDR, SPI_XFER_ADDR, SPI_CONFIG_ADDR = range(3)
@@ -26,56 +26,63 @@ class SPIMaster:
     """
     def __init__(self, dmgr, ref_period, channel):
         self.core = dmgr.get("core")
+        self.ref_period = ref_period
         self.ref_period_mu = seconds_to_mu(ref_period, self.core)
         self.channel = channel
-        self.write_div = 0
-        self.read_div = 0
-        # a full transfer takes prep_mu + xfer_mu
-        self.prep_mu = 0
-        # chained transfers can happen every xfer_mu
-        self.xfer_mu = 0
-        # The second transfer of a chain be written ref_period_mu
-        # after the first. Read data is available every xfer_mu starting
-        # a bit before prep_mu + xfer_mu.
-
-    @portable
-    def predict_xfer_mu(self, write_length, read_length):
-        # this is only the intrinsic bit cycle duration
-        return int(self.ref_period_mu*(
-            write_length*self.write_div +
-            read_length*self.read_div))
-
-    @portable
-    def predict_prep_mu(self, write_div):
-        return int(self.ref_period_mu*(
-            2 +  # intermediate transfers
-            # one write_div for the wait+idle cycle
-            self.write_div))
+        self.write_period_mu = 0
+        self.read_period_mu = 0
+        self.xfer_period_mu = 0
+        # A full transfer takes write_period_mu + xfer_period_mu.
+        # Chained transfers can happen every xfer_period_mu.
+        # The second transfer of a chain can be written 2*ref_period_mu
+        # after the first. Read data is available every xfer_period_mu starting
+        # a bit after xfer_period_mu (depending on clk_phase).
+        # To chain transfers together, new data must be written before
+        # pending transfer's read data becomes available.
 
     @kernel
-    def set_config(self, flags=0, write_div=6, read_div=6):
-        self.write_div = write_div
-        self.read_div = read_div
-        self.prep_mu = self.predict_prep_mu(write_div)
+    def set_config(self, flags=0, write_freq=20*MHz, read_freq=20*MHz):
+        write_div = round(1/(write_freq*self.ref_period))
+        read_div = round(1/(read_freq*self.ref_period))
+        self.set_config_mu(flags, write_div, read_div)
+
+    @kernel
+    def set_config_mu(self, flags=0, write_div=6, read_div=6):
         rt2wb_write(now_mu(), self.channel, SPI_CONFIG_ADDR, flags |
-                    ((write_div - 2) << 8) | ((read_div - 2) << 20))
-        delay_mu(self.ref_period_mu)
+                    ((write_div - 2) << 16) | ((read_div - 2) << 24))
+        self.write_period_mu = int(write_div*self.ref_period_mu)
+        self.read_period_mu = int(read_div*self.ref_period_mu)
+        delay_mu(2*self.ref_period_mu)
+
+    @portable
+    def get_xfer_period_mu(self, write_length, read_length):
+        return int(write_length*self.write_period_mu +
+                   read_length*self.read_period_mu)
 
     @kernel
     def set_xfer(self, chip_select=0, write_length=0, read_length=0):
-        self.xfer_mu = self.predict_xfer_mu(write_length, read_length)
         rt2wb_write(now_mu(), self.channel, SPI_XFER_ADDR,
                     chip_select | (write_length << 16) | (read_length << 24))
-        delay_mu(self.ref_period_mu)
+        self.xfer_period_mu = self.get_xfer_period_mu(
+            write_length, read_length)
+        delay_mu(int(2*self.ref_period_mu))
 
     @kernel
     def write(self, data):
         rt2wb_write(now_mu(), self.channel, SPI_DATA_ADDR, data)
-        delay_mu(int(self.prep_mu + self.xfer_mu))
+        delay_mu(int(self.write_period_mu + self.xfer_period_mu))
+
+    @kernel
+    def read_async(self):
+        rt2wb_write(now_mu(), self.channel, SPI_DATA_ADDR | SPI_RT2WB_READ, 0)
+        delay_mu(int(2*self.ref_period_mu))
 
     @kernel
     def read_sync(self):
-        r = rt2wb_read_sync(now_mu(), self.channel, SPI_DATA_ADDR |
-                            SPI_RT2WB_READ, int(self.ref_period_mu))
-        delay_mu(self.ref_period_mu)
-        return r
+        return rt2wb_read_sync(now_mu(), self.channel, SPI_DATA_ADDR |
+                               SPI_RT2WB_READ, int(2*self.ref_period_mu))
+
+    @kernel
+    def get_config_sync(self):
+        return rt2wb_read_sync(now_mu(), self.channel, SPI_CONFIG_ADDR |
+                               SPI_RT2WB_READ, int(2*self.ref_period_mu))