diff --git a/src/main/scala/devices/spi/SPIArbiter.scala b/src/main/scala/devices/spi/SPIArbiter.scala
index c9650ef..4f401af 100644
--- a/src/main/scala/devices/spi/SPIArbiter.scala
+++ b/src/main/scala/devices/spi/SPIArbiter.scala
@@ -7,9 +7,21 @@ class SPIInnerIO(c: SPIParamsBase) extends SPILinkIO(c) {
   val lock = Output(Bool())
 }
 
+
+/** Arbitor for SPI FlashMap and SPI FIFO
+  *
+  * {{{
+  * Input 0: SPI FlashMap
+  * Input 1: SPI FIFO
+  * Output: to SPI Media
+  * Select signal: come from SoC
+  * }}}
+  */
 class SPIArbiter(c: SPIParamsBase, n: Int) extends Module {
   val io = IO(new Bundle {
+    /** from fifo and SPIFlashMap */
     val inner = Flipped(Vec(n, new SPIInnerIO(c)))
+    /** to SPIMedia */
     val outer = new SPILinkIO(c)
     val sel = Input(UInt(log2Up(n).W))
   })
diff --git a/src/main/scala/devices/spi/SPIConsts.scala b/src/main/scala/devices/spi/SPIConsts.scala
index 0681635..2460ddc 100644
--- a/src/main/scala/devices/spi/SPIConsts.scala
+++ b/src/main/scala/devices/spi/SPIConsts.scala
@@ -9,6 +9,13 @@ object SPIProtocol {
   def Quad   = 2.U(width.W)
 
   def cases = Seq(Single, Dual, Quad)
+  /** protocol select status
+    * {{{
+    * Single -> (1,0,0)
+    * Dual   -> (0,1,0)
+    * Quad   -> (0,0,1)
+    * }}}
+    */
   def decode(x: UInt): Seq[Bool] = cases.map(_ === x)
 }
 
diff --git a/src/main/scala/devices/spi/SPIFlash.scala b/src/main/scala/devices/spi/SPIFlash.scala
index a7f0868..5df8226 100644
--- a/src/main/scala/devices/spi/SPIFlash.scala
+++ b/src/main/scala/devices/spi/SPIFlash.scala
@@ -8,14 +8,19 @@ import chisel3.util._
 
 class SPIFlashInsn(c: SPIFlashParamsBase) extends SPIBundle(c) {
   val cmd = new Bundle with HasSPIProtocol {
+    /** value of command byte */
     val code = Bits(c.insnCmdBits.W)
+    /** enable sending of command */
     val en = Bool()
   }
   val addr = new Bundle with HasSPIProtocol {
+    /** number of address bytes */
     val len = UInt(c.insnAddrLenBits.W)
   }
   val pad = new Bundle {
+    /** first 8 bits to transmit during dummy cycles */
     val code = Bits(c.frameBits.W)
+    /** number of dummy cycles */
     val cnt = Bits(c.insnPadLenBits.W)
   }
   val data = new Bundle with HasSPIProtocol
@@ -46,12 +51,25 @@ class SPIFlashAddr(c: SPIFlashParamsBase) extends SPIBundle(c) {
   val hold = UInt(c.insnAddrBits.W)
 }
 
+/** SPI SoC Controller
+  *
+  * This module processes the operation information from SoC in proper format and order
+  * before transmits them to SPI Media.
+  */
 class SPIFlashMap(c: SPIFlashParamsBase) extends Module {
   val io = IO(new Bundle {
+    /** enable signal from SPI Top */
     val en = Input(Bool())
+    /** control signals from SPI Top
+      *
+      * insn from top TL bus CRs
+      */
     val ctrl = Input(new SPIFlashControl(c))
+    /** falsh address info from TL bus */
     val addr = Flipped(Decoupled(new SPIFlashAddr(c)))
+    /** transmits Flash response to TL bus */
     val data = Decoupled(UInt(c.frameBits.W))
+    /** connected to SPI Aribitor */
     val link = new SPIInnerIO(c)
   })
 
diff --git a/src/main/scala/devices/spi/SPIMedia.scala b/src/main/scala/devices/spi/SPIMedia.scala
index 9ca309d..9e9d46b 100644
--- a/src/main/scala/devices/spi/SPIMedia.scala
+++ b/src/main/scala/devices/spi/SPIMedia.scala
@@ -19,9 +19,36 @@ class SPILinkIO(c: SPIParamsBase) extends SPIBundle(c) {
   val disableOE = c.oeDisableDummy.option(Output(Bool())) // disable oe during dummy cycles in flash mode
 }
 
+/** Low level SPI protocol controller
+  *
+  * It recives operations from SPI Fifo(Tx) and outputs them in SPI Port(out).
+  * In the meantime, it recives response from slave device and uploads them to SPI Fifo(Rx).
+  * It maintain a statemachine for inter-frame delay and chip select function(control signals from TL bus).
+  *
+  * ==Component==
+  *  - SPI Physical
+  *  - SPI Physical control logic
+  *
+  * ==Datapass==
+  * {{{
+  * TL bus <-> SPI Fifo <-> SPI Media <-> SPI Port
+  * }}}
+  *
+  * ==ControlPass==
+  * {{{
+  *   S_main : 3 possible job
+  *            transfer op to SPIPhysical /(drive op.valid)
+  *            assert CS
+  *            idle
+  *   s_interxfr: process inter-frame delay cycles
+  *   s_intercs: Deassert CS
+  * }}}
+  */
 class SPIMedia(c: SPIParamsBase) extends Module {
   val io = IO(new Bundle {
+    /** top level port */
     val port = new SPIPortIO(c)
+    /** recives ctrl signals from SPITopModule */
     val ctrl = new Bundle {
       val sck = Input(new SPIClocking(c))
       val dla = Input(new SPIDelay(c))
@@ -29,6 +56,7 @@ class SPIMedia(c: SPIParamsBase) extends Module {
       val extradel = Input(new SPIExtraDelay(c))
       val sampledel = Input(new SPISampleDelay(c))
     }
+    /** connected to SPI Fifo */
     val link = Flipped(new SPILinkIO(c))
   })
 
diff --git a/src/main/scala/devices/spi/SPIPhysical.scala b/src/main/scala/devices/spi/SPIPhysical.scala
index 389879f..e0a29e7 100644
--- a/src/main/scala/devices/spi/SPIPhysical.scala
+++ b/src/main/scala/devices/spi/SPIPhysical.scala
@@ -39,11 +39,36 @@ class SPIPhyControl(c: SPIParamsBase) extends SPIBundle(c) {
   val sampledel = new SPISampleDelay (c)
 }
 
+/** Basic SPI component to deal with data transfer.
+  *
+  * It recives operations from SPI Fifo(Tx) and outputs them in SPI Port(out).
+  * In the meantime, it recives feedback from slave device and uploads them to TL bus.
+  *
+  * ==Structure==
+  *  - baud rate generator: counter to implement the target baud rate
+  *  - scnt counter: counter of operation transfer
+  *  - sample counter: delay counter for 'sample'
+  *  - last counter: delay counter for 'last'
+  *  - buffer control logic
+  *   - rxd logic: transmit slave device response(dq.i) to buffer
+  *   - txd logic: construct the data to be transmited on dq.o from buffer
+  *
+  * ==datapath==
+  * {{{
+  * dq.i -> rxd -> samples -> buffer -> txd -> dq.o
+  * operation data(parellel in):
+  * TL Interface -> buffer
+  * }}}
+  */
 class SPIPhysical(c: SPIParamsBase) extends Module {
   val io = IO(new SPIBundle(c) {
+    /** top-level SPI port */
     val port = new SPIPortIO(c)
+    /** recives ctrol signals from SPI TopModule */
     val ctrl = Input(new SPIPhyControl(c))
+    /** recives operation from TL bus through SPI Fifo(Tx) */
     val op = Flipped(Decoupled(new SPIMicroOp(c)))
+    /** transmits slave device response to SPI Fifo(Rx) */
     val rx = Valid(Bits(c.frameBits.W))
   })
 
@@ -51,20 +76,41 @@ class SPIPhysical(c: SPIParamsBase) extends Module {
   val ctrl = Reg(new SPIPhyControl(c))
   val proto = SPIProtocol.decode(ctrl.fmt.proto)
 
+  /** indicates operations have been all accepted */
   val accept = WireDefault(false.B)
+  /** sample counter enable signal
+    *
+    * follows [[cref]]
+    */
   val sample = WireDefault(false.B)
+  /** setup txd
+    *
+    * follows [[cref]]
+    * drives [[setup_d]]
+    */
   val setup = WireDefault(false.B)
+  /** enable for last delay counter */
   val last = WireDefault(false.B)
 
   val setup_d = RegNext(setup)
 
+  /** counter of operation transfer
+    *
+    * init = op.cnt when op.data comes in
+    * related ctrl signals: [[beat]]
+    */
   val scnt = RegInit(0.U(c.countBits.W))
+  /** Baud rate generator
+    *
+    * reset = io.ctrl.sck.div(3) when
+    * related ctrl signals: [[stop]]
+    */
   val tcnt = Reg(UInt(c.divisorBits.W))
 
   val stop = (scnt === 0.U)
   val beat = (tcnt === 0.U)
 
-  //Making a delay counter for 'sample'
+  // Making a delay counter for 'sample'
   val totalCoarseDel = (io.ctrl.extradel.coarse + io.ctrl.sampledel.sd)
   val sample_d = RegInit(false.B) 
   val del_cntr = RegInit(UInt(c.divisorBits.W), (c.defaultSampleDel).U)
@@ -87,7 +133,8 @@ class SPIPhysical(c: SPIParamsBase) extends Module {
   }.otherwise {
     sample_d := false.B
   }
-  //Making a delay counter for 'last'
+  // Making a delay counter for 'last'
+  /** indicates the last set of data */
   val last_d = RegInit(false.B) 
   val del_cntr_last = RegInit(UInt(c.divisorBits.W), (c.defaultSampleDel).U)
   when (beat && last) {
@@ -110,16 +157,27 @@ class SPIPhysical(c: SPIParamsBase) extends Module {
   }
   val decr = Mux(beat, scnt, tcnt) - 1.U
   val sched = WireDefault(beat)
+  /** tcnt counter reset signal */
   tcnt := Mux(sched, ctrl.sck.div, decr)
 
+  /** sck output */
   val sck = Reg(Bool())
+  /** actual clock for
+    *
+    * flips in [[beat]]
+    *
+    * drives [[sample]] and [[setup]]
+    */
   val cref = RegInit(true.B)
+  /** spi mode */
   val cinv = ctrl.sck.pha ^ ctrl.sck.pol
-
+  /** converts data to matched endian */
   private def convert(data: UInt, fmt: SPIFormat) =
     Mux(fmt.endian === SPIEndian.MSB, data, Cat(data.asBools))
 
+  // recives reversed dq.i
   val rxd = Cat(io.port.dq.reverse.map(_.i))
+  // rxd after dalay added
   val rxd_delayed = VecInit(Seq.fill(io.port.dq.size)(false.B))
 
   //Adding fine-granularity delay buffers on the received data
@@ -134,21 +192,29 @@ class SPIPhysical(c: SPIParamsBase) extends Module {
     rxd_delayed := rxd.asBools
   }
 
+  // buffer logic
   val rxd_fin = rxd_delayed.asUInt
+  /** @todo why */
   val samples = Seq(rxd_fin(1), rxd_fin(1, 0), rxd_fin)
 
+  // assuming quad
   val buffer = Reg(UInt(c.frameBits.W))
   val buffer_in = convert(io.op.bits.data, io.ctrl.fmt)
   val shift = Mux ((totalCoarseDel > 0.U), setup_d || (sample_d && stop), sample_d)
   buffer := Mux1H(proto, samples.zipWithIndex.map { case (data, i) =>
     val n = 1 << i
     val m = c.frameBits -1
+    // shift: buffer[3:0], sample[3:0]
+    // stay
+    // buffer(3,0) or buffer(8,4)
     Cat(Mux(shift, buffer(m-n, 0), buffer(m, n)),
+        // sample[3:0], or buffer (3:0)
         Mux(sample_d, data, buffer(n-1, 0)))
   })
 
   private def upper(x: UInt, n: Int) = x(c.frameBits-1, c.frameBits-n)
 
+  /** the data to be transmited in port.dq.o */
   val txd = RegInit(0.U(io.port.dq.size.W))
   val txd_in = Mux(accept, upper(buffer_in, 4), upper(buffer, 4))
   val txd_sel = SPIProtocol.decode(Mux(accept, io.ctrl.fmt.proto, ctrl.fmt.proto))
@@ -156,7 +222,7 @@ class SPIPhysical(c: SPIParamsBase) extends Module {
   when (setup) {
     txd := Mux1H(txd_sel, txd_shf)
   }
-
+  // txd enable
   val tx = (ctrl.fmt.iodir === SPIDirection.Tx)
   val txen_in = (proto.head +: proto.tail.map(_ && tx)).scanRight(false.B)(_ || _).init
   val txen = txen_in :+ txen_in.last
@@ -171,13 +237,13 @@ class SPIPhysical(c: SPIParamsBase) extends Module {
       dq.ie := ~(dq.oe)
   }
   io.op.ready := false.B
-
+  /** when true, transmits buffer to SPI fifo */
   val done = RegInit(true.B)
   done := done || last_d
 
   io.rx.valid := done
   io.rx.bits := convert(buffer, ctrl.fmt)
-
+  /** indicates if the op requires data transfer */
   val xfr = Reg(Bool())
 
   when (stop) {
@@ -205,9 +271,10 @@ class SPIPhysical(c: SPIParamsBase) extends Module {
       sck := ctrl.sck.pol
     }
   }
-
+  // data and buffer transfer complete
   when (accept && done) {
     io.op.ready := true.B 
+    // op data comes in
     when (io.op.valid) {
       scnt := op.cnt
       rdisableOE := io.op.bits.disableOE.getOrElse(false.B)
diff --git a/src/main/scala/devices/spi/TLSPI.scala b/src/main/scala/devices/spi/TLSPI.scala
index ef23fc4..c24d413 100644
--- a/src/main/scala/devices/spi/TLSPI.scala
+++ b/src/main/scala/devices/spi/TLSPI.scala
@@ -38,7 +38,21 @@ trait SPIParamsBase {
   lazy val rxDepthBits = log2Floor(rxDepth) + 1
 
 }
-
+/** SPI Parameters
+  *
+  * @param rAddress spi device base address
+  * @param rSize SPI control registers address space size
+  * @param rxDepth rx fifo depth
+  * @param txDepth tx fifo depth
+  * @param csWidth width of chip select signal
+  * @param frameBits number of bits in a frame
+  * @param delayBits width of delay control registers
+  * @param divisorBits baud rate divisor
+  * @param fineDelayBits width of fine delay control register
+  * @param sampleDelayBits width of sample delay control register
+  * @param defaultSampleDel default sample delay number
+  * @param oeDisableDummy disable outout enable during dummy cycles in flash mode
+  */
 case class SPIParams(
     rAddress: BigInt,
     rSize: BigInt = 0x1000,
@@ -62,6 +76,31 @@ case class SPIParams(
   require(!oeDisableDummy)
 }
 
+/** Controller for Quad Serial Peripheral Interface.
+  *
+  * It's a memory-mapped device using data queue to transfer data across the SPI bus.
+  *
+  * ==Features==
+  *  i. Full-duplex operation, simultaneous receive and transmit
+  *  i. Operates in master mode
+  *  i. FIFO depth and width is configurable
+  *  i. Serial clock with programmable polarity, programmable baud rate generator, configurable external device selects number
+  *  i. Separate transmit and receive FIFOs
+  *  i. TL slave interface to Soc
+  *  i. Programmable master mode clock frequencies
+  *
+  * ==Component==
+  *  - SPI FIFO
+  *  - SPI Media
+  *  - SPI control registers
+  *
+  * ==Datapass==
+  * {{{
+  * TL bus <--> SPITopModule <--> IO Port
+  * }}}
+  * ==Interruption==
+  * send interrupt if fifo watermark triggers
+  */
 class SPITopModule(c: SPIParamsBase, outer: TLSPIBase)
     extends LazyModuleImp(outer) {
 
diff --git a/src/main/scala/devices/spi/TLSPIFlash.scala b/src/main/scala/devices/spi/TLSPIFlash.scala
index 89121d3..4d709d8 100644
--- a/src/main/scala/devices/spi/TLSPIFlash.scala
+++ b/src/main/scala/devices/spi/TLSPIFlash.scala
@@ -25,6 +25,22 @@ trait SPIFlashParamsBase extends SPIParamsBase {
   lazy val insnAddrLenBits = log2Floor(insnAddrBytes) + 1
 }
 
+/** SPI Flash Parameters
+  *
+  * @param rAddress control registers base address in TL
+  * @param fAddress flash base address in TL
+  * @param rSize SPI control registers address space size
+  * @param fSize SPI Flash address space size
+  * @param rxDepth rx fifo depth
+  * @param txDepth tx fifo depth
+  * @param csWidth width of chip select signal
+  * @param delayBits width of delay control registers
+  * @param divisorBits baud rate divisor
+  * @param fineDelayBits width of fine delay control register
+  * @param sampleDelayBits width of sample delay control register
+  * @param defaultSampleDel default sample delay number
+  * @param oeDisableDummy disable outout enable during dummy cycles in flash mode
+  */
 case class SPIFlashParams(
     rAddress: BigInt,
     fAddress: BigInt,
@@ -51,6 +67,15 @@ case class SPIFlashParams(
   require(defaultSampleDel >= 0)
 }
 
+/** QSPI which contains Serial Memory Mode to act as a Serial Flash memory controller
+  *
+  * ==Features==
+  *  1. Support two mode:
+  *   i. SPI mode as descrriped in TLSPI
+  *   i. Serial Memory Mode: In this Mode, the QSPI acts as a serial Flash memory controller.
+  *  1. Set of TL accessible FLASH control registers to perform any Flash command
+  *  1. Command, address and data can be sent independently using different modes.(Single, Dual, Quad).
+  */
 class SPIFlashTopModule(c: SPIFlashParamsBase, outer: TLSPIFlashBase)
   extends SPITopModule(c, outer) {