--- /dev/null
+/*
+ * Copyright 2013 Red Hat Inc.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: Ben Skeggs
+ */
+
+#ifdef INCLUDE_PROC
+process(PROC_MEMX, #memx_init, #memx_recv)
+#endif
+
+/******************************************************************************
+ * MEMX data segment
+ *****************************************************************************/
+#ifdef INCLUDE_DATA
+.equ #memx_opcode 0
+.equ #memx_header 2
+.equ #memx_length 4
+.equ #memx_func 8
+
+#define handler(cmd,hdr,len,func) /*
+*/ .b16 MEMX_##cmd /*
+*/ .b16 hdr /*
+*/ .b16 len /*
+*/ .b16 0 /*
+*/ .b32 func
+
+memx_func_head:
+handler(ENTER , 0x0000, 0x0000, #memx_func_enter)
+memx_func_next:
+handler(LEAVE , 0x0000, 0x0000, #memx_func_leave)
+handler(WR32 , 0x0000, 0x0002, #memx_func_wr32)
+handler(WAIT , 0x0004, 0x0000, #memx_func_wait)
+handler(DELAY , 0x0001, 0x0000, #memx_func_delay)
+handler(VBLANK, 0x0001, 0x0000, #memx_func_wait_vblank)
+handler(TRAIN , 0x0000, 0x0000, #memx_func_train)
+memx_func_tail:
+
+.equ #memx_func_size #memx_func_next - #memx_func_head
+.equ #memx_func_num (#memx_func_tail - #memx_func_head) / #memx_func_size
+
+memx_ts_start:
+.b32 0
+memx_ts_end:
+.b32 0
+
+memx_data_head:
+.skip 0x0800
+memx_data_tail:
+
+memx_train_head:
+.skip 0x0100
+memx_train_tail:
+#endif
+
+/******************************************************************************
+ * MEMX code segment
+ *****************************************************************************/
+#ifdef INCLUDE_CODE
+// description
+//
+// $r15 - current (memx)
+// $r4 - packet length
+// $r3 - opcode desciption
+// $r0 - zero
+memx_func_enter:
+#if NVKM_PPWR_CHIPSET == GT215
+ movw $r8 0x1610
+ nv_rd32($r7, $r8)
+ imm32($r6, 0xfffffffc)
+ and $r7 $r6
+ movw $r6 0x2
+ or $r7 $r6
+ nv_wr32($r8, $r7)
+#else
+ movw $r6 0x001620
+ imm32($r7, ~0x00000aa2);
+ nv_rd32($r8, $r6)
+ and $r8 $r7
+ nv_wr32($r6, $r8)
+
+ imm32($r7, ~0x00000001)
+ nv_rd32($r8, $r6)
+ and $r8 $r7
+ nv_wr32($r6, $r8)
+
+ movw $r6 0x0026f0
+ nv_rd32($r8, $r6)
+ and $r8 $r7
+ nv_wr32($r6, $r8)
+#endif
+
+ mov $r6 NV_PPWR_OUTPUT_SET_FB_PAUSE
+ nv_iowr(NV_PPWR_OUTPUT_SET, $r6)
+ memx_func_enter_wait:
+ nv_iord($r6, NV_PPWR_OUTPUT)
+ and $r6 NV_PPWR_OUTPUT_FB_PAUSE
+ bra z #memx_func_enter_wait
+
+ nv_iord($r6, NV_PPWR_TIMER_LOW)
+ st b32 D[$r0 + #memx_ts_start] $r6
+ ret
+
+// description
+//
+// $r15 - current (memx)
+// $r4 - packet length
+// $r3 - opcode desciption
+// $r0 - zero
+memx_func_leave:
+ nv_iord($r6, NV_PPWR_TIMER_LOW)
+ st b32 D[$r0 + #memx_ts_end] $r6
+
+ mov $r6 NV_PPWR_OUTPUT_CLR_FB_PAUSE
+ nv_iowr(NV_PPWR_OUTPUT_CLR, $r6)
+ memx_func_leave_wait:
+ nv_iord($r6, NV_PPWR_OUTPUT)
+ and $r6 NV_PPWR_OUTPUT_FB_PAUSE
+ bra nz #memx_func_leave_wait
+
+#if NVKM_PPWR_CHIPSET == GT215
+ movw $r8 0x1610
+ nv_rd32($r7, $r8)
+ imm32($r6, 0xffffffcc)
+ and $r7 $r6
+ nv_wr32($r8, $r7)
+#else
+ movw $r6 0x0026f0
+ imm32($r7, 0x00000001)
+ nv_rd32($r8, $r6)
+ or $r8 $r7
+ nv_wr32($r6, $r8)
+
+ movw $r6 0x001620
+ nv_rd32($r8, $r6)
+ or $r8 $r7
+ nv_wr32($r6, $r8)
+
+ imm32($r7, 0x00000aa2);
+ nv_rd32($r8, $r6)
+ or $r8 $r7
+ nv_wr32($r6, $r8)
+#endif
+ ret
+
+#if NVKM_PPWR_CHIPSET < GF119
+// description
+//
+// $r15 - current (memx)
+// $r4 - packet length
+// +00: head to wait for vblank on
+// $r3 - opcode desciption
+// $r0 - zero
+memx_func_wait_vblank:
+ ld b32 $r6 D[$r1 + 0x00]
+ cmp b32 $r6 0x0
+ bra z #memx_func_wait_vblank_head0
+ cmp b32 $r6 0x1
+ bra z #memx_func_wait_vblank_head1
+ bra #memx_func_wait_vblank_fini
+
+ memx_func_wait_vblank_head1:
+ movw $r7 0x20
+ bra #memx_func_wait_vblank_0
+
+ memx_func_wait_vblank_head0:
+ movw $r7 0x8
+
+ memx_func_wait_vblank_0:
+ nv_iord($r6, NV_PPWR_INPUT)
+ and $r6 $r7
+ bra nz #memx_func_wait_vblank_0
+
+ memx_func_wait_vblank_1:
+ nv_iord($r6, NV_PPWR_INPUT)
+ and $r6 $r7
+ bra z #memx_func_wait_vblank_1
+
+ memx_func_wait_vblank_fini:
+ add b32 $r1 0x4
+ ret
+
+#else
+
+// XXX: currently no-op
+//
+// $r15 - current (memx)
+// $r4 - packet length
+// +00: head to wait for vblank on
+// $r3 - opcode desciption
+// $r0 - zero
+memx_func_wait_vblank:
+ add b32 $r1 0x4
+ ret
+
+#endif
+
+// description
+//
+// $r15 - current (memx)
+// $r4 - packet length
+// +00*n: addr
+// +04*n: data
+// $r3 - opcode desciption
+// $r0 - zero
+memx_func_wr32:
+ ld b32 $r6 D[$r1 + 0x00]
+ ld b32 $r5 D[$r1 + 0x04]
+ add b32 $r1 0x08
+ nv_wr32($r6, $r5)
+ sub b32 $r4 0x02
+ bra nz #memx_func_wr32
+ ret
+
+// description
+//
+// $r15 - current (memx)
+// $r4 - packet length
+// +00: addr
+// +04: mask
+// +08: data
+// +0c: timeout (ns)
+// $r3 - opcode desciption
+// $r0 - zero
+memx_func_wait:
+ nv_iord($r8, NV_PPWR_TIMER_LOW)
+ ld b32 $r14 D[$r1 + 0x00]
+ ld b32 $r13 D[$r1 + 0x04]
+ ld b32 $r12 D[$r1 + 0x08]
+ ld b32 $r11 D[$r1 + 0x0c]
+ add b32 $r1 0x10
+ call(wait)
+ ret
+
+// description
+//
+// $r15 - current (memx)
+// $r4 - packet length
+// +00: time (ns)
+// $r3 - opcode desciption
+// $r0 - zero
+memx_func_delay:
+ ld b32 $r14 D[$r1 + 0x00]
+ add b32 $r1 0x04
+ call(nsec)
+ ret
+
+// description
+//
+// $r15 - current (memx)
+// $r4 - packet length
+// $r3 - opcode desciption
+// $r0 - zero
+memx_func_train:
+#if NVKM_PPWR_CHIPSET == GT215
+// $r5 - outer loop counter
+// $r6 - inner loop counter
+// $r7 - entry counter (#memx_train_head + $r7)
+ movw $r5 0x3
+ movw $r7 0x0
+
+// Read random memory to wake up... things
+ imm32($r9, 0x700000)
+ nv_rd32($r8,$r9)
+ movw $r14 0x2710
+ call(nsec)
+
+ memx_func_train_loop_outer:
+ mulu $r8 $r5 0x101
+ sethi $r8 0x02000000
+ imm32($r9, 0x1111e0)
+ nv_wr32($r9, $r8)
+ push $r5
+
+ movw $r6 0x0
+ memx_func_train_loop_inner:
+ movw $r8 0x1111
+ mulu $r9 $r6 $r8
+ shl b32 $r8 $r9 0x10
+ or $r8 $r9
+ imm32($r9, 0x100720)
+ nv_wr32($r9, $r8)
+
+ imm32($r9, 0x100080)
+ nv_rd32($r8, $r9)
+ or $r8 $r8 0x20
+ nv_wr32($r9, $r8)
+
+ imm32($r9, 0x10053c)
+ imm32($r8, 0x80003002)
+ nv_wr32($r9, $r8)
+
+ imm32($r14, 0x100560)
+ imm32($r13, 0x80000000)
+ add b32 $r12 $r13 0
+ imm32($r11, 0x001e8480)
+ call(wait)
+
+ // $r5 - inner inner loop counter
+ // $r9 - result
+ movw $r5 0
+ imm32($r9, 0x8300ffff)
+ memx_func_train_loop_4x:
+ imm32($r10, 0x100080)
+ nv_rd32($r8, $r10)
+ imm32($r11, 0xffffffdf)
+ and $r8 $r11
+ nv_wr32($r10, $r8)
+
+ imm32($r10, 0x10053c)
+ imm32($r8, 0x80003002)
+ nv_wr32($r10, $r8)
+
+ imm32($r14, 0x100560)
+ imm32($r13, 0x80000000)
+ mov b32 $r12 $r13
+ imm32($r11, 0x00002710)
+ call(wait)
+
+ nv_rd32($r13, $r14)
+ and $r9 $r9 $r13
+
+ add b32 $r5 1
+ cmp b16 $r5 0x4
+ bra l #memx_func_train_loop_4x
+
+ add b32 $r10 $r7 #memx_train_head
+ st b32 D[$r10 + 0] $r9
+ add b32 $r6 1
+ add b32 $r7 4
+
+ cmp b16 $r6 0x10
+ bra l #memx_func_train_loop_inner
+
+ pop $r5
+ add b32 $r5 1
+ cmp b16 $r5 7
+ bra l #memx_func_train_loop_outer
+
+#endif
+ ret
+
+// description
+//
+// $r15 - current (memx)
+// $r14 - sender process name
+// $r13 - message (exec)
+// $r12 - head of script
+// $r11 - tail of script
+// $r0 - zero
+memx_exec:
+ push $r14
+ push $r13
+ mov b32 $r1 $r12
+ mov b32 $r2 $r11
+
+ memx_exec_next:
+ // fetch the packet header
+ ld b32 $r3 D[$r1]
+ add b32 $r1 4
+ extr $r4 $r3 16:31
+ extr $r3 $r3 0:15
+
+ // execute the opcode handler
+ sub b32 $r3 1
+ mulu $r3 #memx_func_size
+ ld b32 $r5 D[$r3 + #memx_func_head + #memx_func]
+ call $r5
+
+ // keep going, if we haven't reached the end
+ cmp b32 $r1 $r2
+ bra l #memx_exec_next
+
+ // send completion reply
+ ld b32 $r11 D[$r0 + #memx_ts_start]
+ ld b32 $r12 D[$r0 + #memx_ts_end]
+ sub b32 $r12 $r11
+ nv_iord($r11, NV_PPWR_INPUT)
+ pop $r13
+ pop $r14
+ call(send)
+ ret
+
+// description
+//
+// $r15 - current (memx)
+// $r14 - sender process name
+// $r13 - message
+// $r12 - data0
+// $r11 - data1
+// $r0 - zero
+memx_info:
+ cmp b16 $r12 0x1
+ bra e #memx_info_train
+
+ memx_info_data:
+ mov $r12 #memx_data_head
+ mov $r11 #memx_data_tail - #memx_data_head
+ bra #memx_info_send
+
+ memx_info_train:
+ mov $r12 #memx_train_head
+ mov $r11 #memx_train_tail - #memx_train_head
+
+ memx_info_send:
+ call(send)
+ ret
+
+// description
+//
+// $r15 - current (memx)
+// $r14 - sender process name
+// $r13 - message
+// $r12 - data0
+// $r11 - data1
+// $r0 - zero
+memx_recv:
+ cmp b32 $r13 MEMX_MSG_EXEC
+ bra e #memx_exec
+ cmp b32 $r13 MEMX_MSG_INFO
+ bra e #memx_info
+ ret
+
+// description
+//
+// $r15 - current (memx)
+// $r0 - zero
+memx_init:
+ ret
+#endif
Code Review / kvmfornfv.git / blobdiff