Commit 370678c5 authored by Dave Airlie's avatar Dave Airlie
Browse files

Merge tag 'drm-intel-fixes-2020-07-01' of... 

Merge tag 'drm-intel-fixes-2020-07-01' of git://anongit.freedesktop.org/drm/drm-intel

 into drm-fixes

drm/i915 fixes for v5.8-rc4:
- GVT fixes
- Include asm sources for render cache clear batches

Signed-off-by: default avatarDave Airlie <airlied@redhat.com>
From: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/87imf7l6ee.fsf@intel.com
parents a0d9dc02 55fd7e02

drivers/gpu/drm/i915/gt/shaders/README

0 → 100644
+46 −0
Original line number Diff line number Diff line 
ASM sources for auto generated shaders

======================================



The i915/gt/hsw_clear_kernel.c and i915/gt/ivb_clear_kernel.c files contain

pre-compiled batch chunks that will clear any residual render cache during

context switch.



They are generated from their respective platform ASM files present on

i915/gt/shaders/clear_kernel directory.



The generated .c files should never be modified directly. Instead, any modification

needs to be done on the on their respective ASM files and build instructions below

needes to be followed.



Building

========



Environment

-----------



IGT GPU tool scripts and the Mesa's i965 instruction assembler tool are used

on building.



Please make sure your Mesa tool is compiled with "-Dtools=intel" and

"-Ddri-drivers=i965", and run this script from IGT source root directory"



The instructions bellow assume:

    *  IGT gpu tools source code is located on your home directory (~) as ~/igt

    *  Mesa source code is located on your home directory (~) as ~/mesa

       and built under the ~/mesa/build directory

    *  Linux kernel source code is under your home directory (~) as ~/linux



Instructions

------------



~ $ cp ~/linux/drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm \

       ~/igt/lib/i915/shaders/clear_kernel/ivb.asm

~ $ cd ~/igt

igt $ ./scripts/generate_clear_kernel.sh -g ivb \

      -m ~/mesa/build/src/intel/tools/i965_asm



~ $ cp ~/linux/drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm \

    ~/igt/lib/i915/shaders/clear_kernel/hsw.asm

~ $ cd ~/igt

igt $ ./scripts/generate_clear_kernel.sh -g hsw \

      -m ~/mesa/build/src/intel/tools/i965_asm
 
 No newline at end of file

drivers/gpu/drm/i915/gt/shaders/clear_kernel/hsw.asm

0 → 100644
+119 −0
Original line number Diff line number Diff line 
// SPDX-License-Identifier: MIT

/*

 * Copyright © 2020 Intel Corporation

 */



/*

 * Kernel for PAVP buffer clear.

 *

 *	1. Clear all 64 GRF registers assigned to the kernel with designated value;

 *	2. Write 32x16 block of all "0" to render target buffer which indirectly clears

 *	   512 bytes of Render Cache.

 */



/* Store designated "clear GRF" value */

mov(1)          f0.1<1>UW       g1.2<0,1,0>UW                   { align1 1N };



/**

 * Curbe Format

 *

 * DW 1.0 - Block Offset to write Render Cache

 * DW 1.1 [15:0] - Clear Word

 * DW 1.2 - Delay iterations

 * DW 1.3 - Enable Instrumentation (only for debug)

 * DW 1.4 - Rsvd (intended for context ID)

 * DW 1.5 - [31:16]:SliceCount, [15:0]:SubSlicePerSliceCount

 * DW 1.6 - Rsvd MBZ (intended for Enable Wait on Total Thread Count)

 * DW 1.7 - Rsvd MBZ (inteded for Total Thread Count)

 *

 * Binding Table

 *

 * BTI 0: 2D Surface to help clear L3 (Render/Data Cache)

 * BTI 1: Wait/Instrumentation Buffer

 *  Size : (SliceCount * SubSliceCount  * 16 EUs/SubSlice) rows * (16 threads/EU) cols (Format R32_UINT)

 *         Expected to be initialized to 0 by driver/another kernel

 *  Layout:

 *          RowN: Histogram for EU-N: (SliceID*SubSlicePerSliceCount + SSID)*16 + EUID [assume max 16 EUs / SS]

 *          Col-k[DW-k]: Threads Executed on ThreadID-k for EU-N

 */

add(1)          g1.2<1>UD       g1.2<0,1,0>UD   0x00000001UD    { align1 1N }; /* Loop count to delay kernel: Init to (g1.2 + 1) */

cmp.z.f0.0(1)   null<1>UD       g1.3<0,1,0>UD   0x00000000UD    { align1 1N };

(+f0.0) jmpi(1) 352D                                            { align1 WE_all 1N };



/**

 * State Register has info on where this thread is running

 *	IVB: sr0.0 :: [15:13]: MBZ, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID

 *	HSW: sr0.0 :: 15: MBZ, [14:13]: SliceID, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID

 */

mov(8)          g3<1>UD         0x00000000UD                    { align1 1Q };

shr(1)          g3<1>D          sr0<0,1,0>D     12D             { align1 1N };

and(1)          g3<1>D          g3<0,1,0>D      1D              { align1 1N }; /* g3 has HSID */

shr(1)          g3.1<1>D        sr0<0,1,0>D     13D             { align1 1N };

and(1)          g3.1<1>D        g3.1<0,1,0>D    3D              { align1 1N }; /* g3.1 has sliceID */

mul(1)          g3.5<1>D        g3.1<0,1,0>D    g1.10<0,1,0>UW  { align1 1N };

add(1)          g3<1>D          g3<0,1,0>D      g3.5<0,1,0>D    { align1 1N }; /* g3 = sliceID * SubSlicePerSliceCount + HSID */

shr(1)          g3.2<1>D        sr0<0,1,0>D     8D              { align1 1N };

and(1)          g3.2<1>D        g3.2<0,1,0>D    15D             { align1 1N }; /* g3.2 = EUID */

mul(1)          g3.4<1>D        g3<0,1,0>D      16D             { align1 1N };

add(1)          g3.2<1>D        g3.2<0,1,0>D    g3.4<0,1,0>D    { align1 1N }; /* g3.2 now points to EU row number (Y-pixel = V address )  in instrumentation surf */



mov(8)          g5<1>UD         0x00000000UD                    { align1 1Q };

and(1)          g3.3<1>D        sr0<0,1,0>D     7D              { align1 1N };

mul(1)          g3.3<1>D        g3.3<0,1,0>D    4D              { align1 1N };



mov(8)          g4<1>UD         g0<8,8,1>UD                     { align1 1Q }; /* Initialize message header with g0 */

mov(1)          g4<1>UD         g3.3<0,1,0>UD                   { align1 1N }; /* Block offset */

mov(1)          g4.1<1>UD       g3.2<0,1,0>UD                   { align1 1N }; /* Block offset */

mov(1)          g4.2<1>UD       0x00000003UD                    { align1 1N }; /* Block size (1 row x 4 bytes) */

and(1)          g4.3<1>UD       g4.3<0,1,0>UW   0xffffffffUD    { align1 1N };



/* Media block read to fetch current value at specified location in instrumentation buffer */

sendc(8)        g5<1>UD         g4<8,8,1>F      0x02190001



                            render MsgDesc: media block read MsgCtrl = 0x0 Surface = 1 mlen 1 rlen 1 { align1 1Q };

add(1)          g5<1>D          g5<0,1,0>D      1D              { align1 1N };



/* Media block write for updated value at specified location in instrumentation buffer */

sendc(8)        g5<1>UD         g4<8,8,1>F      0x040a8001

                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 1 mlen 2 rlen 0 { align1 1Q };



/* Delay thread for specified parameter */

add.nz.f0.0(1)  g1.2<1>UD       g1.2<0,1,0>UD   -1D             { align1 1N };

(+f0.0) jmpi(1) -32D                                            { align1 WE_all 1N };



/* Store designated "clear GRF" value */

mov(1)          f0.1<1>UW       g1.2<0,1,0>UW                   { align1 1N };



/* Initialize looping parameters */

mov(1)          a0<1>D          0D                              { align1 1N }; /* Initialize a0.0:w=0 */

mov(1)          a0.4<1>W        127W                            { align1 1N }; /* Loop count. Each loop contains 16 GRF's */



/* Write 32x16 all "0" block */

mov(8)          g2<1>UD         g0<8,8,1>UD                     { align1 1Q };

mov(8)          g127<1>UD       g0<8,8,1>UD                     { align1 1Q };

mov(2)          g2<1>UD         g1<2,2,1>UW                     { align1 1N };

mov(1)          g2.2<1>UD       0x000f000fUD                    { align1 1N }; /* Block size (16x16) */

and(1)          g2.3<1>UD       g2.3<0,1,0>UW   0xffffffefUD    { align1 1N };

mov(16)         g3<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g4<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g5<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g6<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g7<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g8<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g9<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g10<1>UD        0x00000000UD                    { align1 1H };

sendc(8)        null<1>UD       g2<8,8,1>F      0x120a8000

                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q };

add(1)          g2<1>UD         g1<0,1,0>UW     0x0010UW        { align1 1N };

sendc(8)        null<1>UD       g2<8,8,1>F      0x120a8000

                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q };



/* Now, clear all GRF registers */

add.nz.f0.0(1)  a0.4<1>W        a0.4<0,1,0>W    -1W             { align1 1N };

mov(16)         g[a0]<1>UW      f0.1<0,1,0>UW                   { align1 1H };

add(1)          a0<1>D          a0<0,1,0>D      32D             { align1 1N };

(+f0.0) jmpi(1) -64D                                            { align1 WE_all 1N };



/* Terminante the thread */

sendc(8)        null<1>UD       g127<8,8,1>F    0x82000010

                            thread_spawner MsgDesc: mlen 1 rlen 0           { align1 1Q EOT };

drivers/gpu/drm/i915/gt/shaders/clear_kernel/ivb.asm

0 → 100644
+117 −0
Original line number Diff line number Diff line 
// SPDX-License-Identifier: MIT

/*

 * Copyright © 2020 Intel Corporation

 */



/*

 * Kernel for PAVP buffer clear.

 *

 *	1. Clear all 64 GRF registers assigned to the kernel with designated value;

 *	2. Write 32x16 block of all "0" to render target buffer which indirectly clears

 *	   512 bytes of Render Cache.

 */



/* Store designated "clear GRF" value */

mov(1)          f0.1<1>UW       g1.2<0,1,0>UW                   { align1 1N };



/**

 * Curbe Format

 *

 * DW 1.0 - Block Offset to write Render Cache

 * DW 1.1 [15:0] - Clear Word

 * DW 1.2 - Delay iterations

 * DW 1.3 - Enable Instrumentation (only for debug)

 * DW 1.4 - Rsvd (intended for context ID)

 * DW 1.5 - [31:16]:SliceCount, [15:0]:SubSlicePerSliceCount

 * DW 1.6 - Rsvd MBZ (intended for Enable Wait on Total Thread Count)

 * DW 1.7 - Rsvd MBZ (inteded for Total Thread Count)

 *

 * Binding Table

 *

 * BTI 0: 2D Surface to help clear L3 (Render/Data Cache)

 * BTI 1: Wait/Instrumentation Buffer

 *  Size : (SliceCount * SubSliceCount  * 16 EUs/SubSlice) rows * (16 threads/EU) cols (Format R32_UINT)

 *         Expected to be initialized to 0 by driver/another kernel

 *  Layout :

 *           RowN: Histogram for EU-N: (SliceID*SubSlicePerSliceCount + SSID)*16 + EUID [assume max 16 EUs / SS]

 *           Col-k[DW-k]: Threads Executed on ThreadID-k for EU-N

 */

add(1)          g1.2<1>UD       g1.2<0,1,0>UD   0x00000001UD    { align1 1N }; /* Loop count to delay kernel: Init to (g1.2 + 1) */

cmp.z.f0.0(1)   null<1>UD       g1.3<0,1,0>UD   0x00000000UD    { align1 1N };

(+f0.0) jmpi(1) 44D                                             { align1 WE_all 1N };



/**

 * State Register has info on where this thread is running

 *	IVB: sr0.0 :: [15:13]: MBZ, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID

 *	HSW: sr0.0 :: 15: MBZ, [14:13]: SliceID, 12: HSID (Half-Slice ID), [11:8]EUID, [2:0] ThreadSlotID

 */

mov(8)          g3<1>UD         0x00000000UD                    { align1 1Q };

shr(1)          g3<1>D          sr0<0,1,0>D     12D             { align1 1N };

and(1)          g3<1>D          g3<0,1,0>D      1D              { align1 1N }; /* g3 has HSID */

shr(1)          g3.1<1>D        sr0<0,1,0>D     13D             { align1 1N };

and(1)          g3.1<1>D        g3.1<0,1,0>D    3D              { align1 1N }; /* g3.1 has sliceID */

mul(1)          g3.5<1>D        g3.1<0,1,0>D    g1.10<0,1,0>UW  { align1 1N };

add(1)          g3<1>D          g3<0,1,0>D      g3.5<0,1,0>D    { align1 1N }; /* g3 = sliceID * SubSlicePerSliceCount + HSID */

shr(1)          g3.2<1>D        sr0<0,1,0>D     8D              { align1 1N };

and(1)          g3.2<1>D        g3.2<0,1,0>D    15D             { align1 1N }; /* g3.2 = EUID */

mul(1)          g3.4<1>D        g3<0,1,0>D      16D             { align1 1N };

add(1)          g3.2<1>D        g3.2<0,1,0>D    g3.4<0,1,0>D    { align1 1N }; /* g3.2 now points to EU row number (Y-pixel = V address )  in instrumentation surf */



mov(8)          g5<1>UD         0x00000000UD                    { align1 1Q };

and(1)          g3.3<1>D        sr0<0,1,0>D     7D              { align1 1N };

mul(1)          g3.3<1>D        g3.3<0,1,0>D    4D              { align1 1N };



mov(8)          g4<1>UD         g0<8,8,1>UD                     { align1 1Q }; /* Initialize message header with g0 */

mov(1)          g4<1>UD         g3.3<0,1,0>UD                   { align1 1N }; /* Block offset */

mov(1)          g4.1<1>UD       g3.2<0,1,0>UD                   { align1 1N }; /* Block offset */

mov(1)          g4.2<1>UD       0x00000003UD                    { align1 1N }; /* Block size (1 row x 4 bytes) */

and(1)          g4.3<1>UD       g4.3<0,1,0>UW   0xffffffffUD    { align1 1N };



/* Media block read to fetch current value at specified location in instrumentation buffer */

sendc(8)        g5<1>UD         g4<8,8,1>F      0x02190001

                            render MsgDesc: media block read MsgCtrl = 0x0 Surface = 1 mlen 1 rlen 1 { align1 1Q };

add(1)          g5<1>D          g5<0,1,0>D      1D              { align1 1N };



/* Media block write for updated value at specified location in instrumentation buffer */

sendc(8)        g5<1>UD         g4<8,8,1>F      0x040a8001

                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 1 mlen 2 rlen 0 { align1 1Q };

/* Delay thread for specified parameter */

add.nz.f0.0(1)  g1.2<1>UD       g1.2<0,1,0>UD   -1D             { align1 1N };

(+f0.0) jmpi(1) -4D                                             { align1 WE_all 1N };



/* Store designated "clear GRF" value */

mov(1)          f0.1<1>UW       g1.2<0,1,0>UW                   { align1 1N };



/* Initialize looping parameters */

mov(1)          a0<1>D          0D                              { align1 1N }; /* Initialize a0.0:w=0 */

mov(1)          a0.4<1>W        127W                            { align1 1N }; /* Loop count. Each loop contains 16 GRF's */



/* Write 32x16 all "0" block */

mov(8)          g2<1>UD         g0<8,8,1>UD                     { align1 1Q };

mov(8)          g127<1>UD       g0<8,8,1>UD                     { align1 1Q };

mov(2)          g2<1>UD         g1<2,2,1>UW                     { align1 1N };

mov(1)          g2.2<1>UD       0x000f000fUD                    { align1 1N }; /* Block size (16x16) */

and(1)          g2.3<1>UD       g2.3<0,1,0>UW   0xffffffefUD    { align1 1N };

mov(16)         g3<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g4<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g5<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g6<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g7<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g8<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g9<1>UD         0x00000000UD                    { align1 1H };

mov(16)         g10<1>UD        0x00000000UD                    { align1 1H };

sendc(8)        null<1>UD       g2<8,8,1>F      0x120a8000

                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q };

add(1)          g2<1>UD         g1<0,1,0>UW     0x0010UW        { align1 1N };

sendc(8)        null<1>UD       g2<8,8,1>F      0x120a8000

                            render MsgDesc: media block write MsgCtrl = 0x0 Surface = 0 mlen 9 rlen 0 { align1 1Q };



/* Now, clear all GRF registers */

add.nz.f0.0(1)  a0.4<1>W        a0.4<0,1,0>W    -1W             { align1 1N };

mov(16)         g[a0]<1>UW      f0.1<0,1,0>UW                   { align1 1H };

add(1)          a0<1>D          a0<0,1,0>D      32D             { align1 1N };

(+f0.0) jmpi(1) -8D                                             { align1 WE_all 1N };



/* Terminante the thread */

sendc(8)        null<1>UD       g127<8,8,1>F    0x82000010

                            thread_spawner MsgDesc: mlen 1 rlen 0           { align1 1Q EOT };

drivers/gpu/drm/i915/gvt/debugfs.c

+1 −1
Original line number Diff line number Diff line
@@ -66,7 +66,7 @@ static inline int mmio_diff_handler(struct intel_gvt *gvt, 
	vreg = vgpu_vreg(param->vgpu, offset);



	if (preg != vreg) {

		node = kmalloc(sizeof(*node), GFP_KERNEL);

		node = kmalloc(sizeof(*node), GFP_ATOMIC);

		if (!node)

			return -ENOMEM;

drivers/gpu/drm/i915/gvt/handlers.c

+13 −11
Original line number Diff line number Diff line
@@ -1726,13 +1726,13 @@ static int ring_mode_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, 
		(*(u32 *)p_data) &= ~_MASKED_BIT_ENABLE(2);

	write_vreg(vgpu, offset, p_data, bytes);



	if (data & _MASKED_BIT_ENABLE(1)) {

	if (IS_MASKED_BITS_ENABLED(data, 1)) {

		enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST);

		return 0;

	}



	if (IS_COFFEELAKE(vgpu->gvt->gt->i915) &&

	    data & _MASKED_BIT_ENABLE(2)) {

	    IS_MASKED_BITS_ENABLED(data, 2)) {

		enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST);

		return 0;

	}
@@ -1741,14 +1741,14 @@ static int ring_mode_mmio_write(struct intel_vgpu *vgpu, unsigned int offset, 
	 * pvinfo, if not, we will treat this guest as non-gvtg-aware

	 * guest, and stop emulating its cfg space, mmio, gtt, etc.

	 */

	if (((data & _MASKED_BIT_ENABLE(GFX_PPGTT_ENABLE)) ||

			(data & _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE)))

			&& !vgpu->pv_notified) {

	if ((IS_MASKED_BITS_ENABLED(data, GFX_PPGTT_ENABLE) ||

	    IS_MASKED_BITS_ENABLED(data, GFX_RUN_LIST_ENABLE)) &&

	    !vgpu->pv_notified) {

		enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST);

		return 0;

	}

	if ((data & _MASKED_BIT_ENABLE(GFX_RUN_LIST_ENABLE))

			|| (data & _MASKED_BIT_DISABLE(GFX_RUN_LIST_ENABLE))) {

	if (IS_MASKED_BITS_ENABLED(data, GFX_RUN_LIST_ENABLE) ||

	    IS_MASKED_BITS_DISABLED(data, GFX_RUN_LIST_ENABLE)) {

		enable_execlist = !!(data & GFX_RUN_LIST_ENABLE);



		gvt_dbg_core("EXECLIST %s on ring %s\n",
@@ -1809,7 +1809,7 @@ static int ring_reset_ctl_write(struct intel_vgpu *vgpu, 
	write_vreg(vgpu, offset, p_data, bytes);

	data = vgpu_vreg(vgpu, offset);



	if (data & _MASKED_BIT_ENABLE(RESET_CTL_REQUEST_RESET))

	if (IS_MASKED_BITS_ENABLED(data, RESET_CTL_REQUEST_RESET))

		data |= RESET_CTL_READY_TO_RESET;

	else if (data & _MASKED_BIT_DISABLE(RESET_CTL_REQUEST_RESET))

		data &= ~RESET_CTL_READY_TO_RESET;
@@ -1827,7 +1827,8 @@ static int csfe_chicken1_mmio_write(struct intel_vgpu *vgpu, 
	(*(u32 *)p_data) &= ~_MASKED_BIT_ENABLE(0x18);

	write_vreg(vgpu, offset, p_data, bytes);



	if (data & _MASKED_BIT_ENABLE(0x10) || data & _MASKED_BIT_ENABLE(0x8))

	if (IS_MASKED_BITS_ENABLED(data, 0x10) ||

	    IS_MASKED_BITS_ENABLED(data, 0x8))

		enter_failsafe_mode(vgpu, GVT_FAILSAFE_UNSUPPORTED_GUEST);



	return 0;
@@ -3055,6 +3056,7 @@ static int init_skl_mmio_info(struct intel_gvt *gvt) 
	MMIO_D(_MMIO(0x72380), D_SKL_PLUS);

	MMIO_D(_MMIO(0x7239c), D_SKL_PLUS);

	MMIO_D(_MMIO(_PLANE_SURF_3_A), D_SKL_PLUS);

	MMIO_D(_MMIO(_PLANE_SURF_3_B), D_SKL_PLUS);



	MMIO_D(CSR_SSP_BASE, D_SKL_PLUS);

	MMIO_D(CSR_HTP_SKL, D_SKL_PLUS);
@@ -3131,8 +3133,8 @@ static int init_skl_mmio_info(struct intel_gvt *gvt) 
	MMIO_DFH(GEN9_WM_CHICKEN3, D_SKL_PLUS, F_MODE_MASK | F_CMD_ACCESS,

		 NULL, NULL);



	MMIO_D(GAMT_CHKN_BIT_REG, D_KBL);

	MMIO_D(GEN9_CTX_PREEMPT_REG, D_KBL | D_SKL);

	MMIO_D(GAMT_CHKN_BIT_REG, D_KBL | D_CFL);

	MMIO_D(GEN9_CTX_PREEMPT_REG, D_SKL_PLUS);



	return 0;

}