nvc0: fix offsets of MP perf counters input parameters

This fixes a regression introduced in
1da704a94c57aa0b0cf8faaa3236fe47dfb8f88c because the offset has moved
from 0x600 to 0x620, and the kernels used for reading MP perf counters
have to be re-assembled.

This also fixes amd_performance_monitor_measure piglit.

Fixes: 1da704a ("nvc0: increase the tex handles area size in the driver")
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>

diff --git a/src/gallium/drivers/nouveau/nvc0/nvc0_query_hw_sm.c b/src/gallium/drivers/nouveau/nvc0/nvc0_query_hw_sm.c
index 944349d9e8a0a5c4d42d22b60d0ab0057bb96978..d4721201bbaa84d59d8463e54ac04a3a45f1a676 100644 (file)
--- a/src/gallium/drivers/nouveau/nvc0/nvc0_query_hw_sm.c
+++ b/src/gallium/drivers/nouveau/nvc0/nvc0_query_hw_sm.c
@@ -372,9 +372,9 @@ static const uint64_t nve4_read_hw_sm_counters_code[] =
     * mov b32 $r6 $pm6
     * mov b32 $r7 $pm7
     * set $p0 0x1 eq u32 $r8 0x0
-    * mov b32 $r10 c7[0x600]
+    * mov b32 $r10 c7[0x620]
     * ext u32 $r8 $r12 0x414
-    * mov b32 $r11 c7[0x604]
+    * mov b32 $r11 c7[0x624]
     * sched 0x04 0x2e 0x04 0x20 0x20 0x28 0x04
     * ext u32 $r9 $r12 0x208
     * (not $p0) exit
@@ -392,7 +392,7 @@ static const uint64_t nve4_read_hw_sm_counters_code[] =
     * add b32 $r12 $c $r12 $r9
     * st b128 wt g[$r10d] $r0q
     * sched 0x4 0x2c 0x20 0x04 0x2e 0x00 0x00
-    * mov b32 $r0 c7[0x608]
+    * mov b32 $r0 c7[0x628]
     * add b32 $r13 $r13 0x0 $c
     * $p1 st b128 wt g[$r12d+0x40] $r4q
     * st b32 wt g[$r12d+0x50] $r0
@@ -410,9 +410,9 @@ static const uint64_t nve4_read_hw_sm_counters_code[] =
    0x2c00000028019c04ULL,
    0x2c0000002c01dc04ULL,
    0x190e0000fc81dc03ULL,
-   0x28005c1800029de4ULL,
+   0x28005c1880029de4ULL,
    0x7000c01050c21c03ULL,
-   0x28005c181002dde4ULL,
+   0x28005c189002dde4ULL,
    0x204282020042e047ULL,
    0x7000c00820c25c03ULL,
    0x80000000000021e7ULL,
@@ -430,7 +430,7 @@ static const uint64_t nve4_read_hw_sm_counters_code[] =
    0x4801000024c31c03ULL,
    0x9400000000a01fc5ULL,
    0x200002e04202c047ULL,
-   0x28005c1820001de4ULL,
+   0x28005c18a0001de4ULL,
    0x0800000000d35c42ULL,
    0x9400000100c107c5ULL,
    0x9400000140c01f85ULL,
@@ -453,9 +453,9 @@ static const uint64_t nvf0_read_hw_sm_counters_code[] =
    0x86400000051c001aULL,
    0x86400000059c001eULL,
    0xdb201c007f9c201eULL,
-   0x64c03ce0c01c002aULL,
+   0x64c03ce0c41c002aULL,
    0xc00000020a1c3021ULL,
-   0x64c03ce0c09c002eULL,
+   0x64c03ce0c49c002eULL,
    0x0810a0808010b810ULL,
    0xc0000001041c3025ULL,
    0x180000000020003cULL,
@@ -473,7 +473,7 @@ static const uint64_t nvf0_read_hw_sm_counters_code[] =
    0xe0840000049c3032ULL,
    0xfe800000001c2800ULL,
    0x080000b81080b010ULL,
-   0x64c03ce0c11c0002ULL,
+   0x64c03ce0c51c0002ULL,
    0xe08040007f9c3436ULL,
    0xfe80000020043010ULL,
    0xfc800000281c3000ULL,
@@ -1105,14 +1105,14 @@ static const uint64_t nvc0_read_hw_sm_counters_code[] =
     * mov b32 $r6 $pm6
     * mov b32 $r7 $pm7
     * set $p0 0x1 eq u32 $r8 0x0
-    * mov b32 $r10 c15[0x600]
-    * mov b32 $r11 c15[0x604]
+    * mov b32 $r10 c15[0x620]
+    * mov b32 $r11 c15[0x624]
     * ext u32 $r8 $r9 0x414
     * (not $p0) exit
     * mul $r8 u32 $r8 u32 48
     * add b32 $r10 $c $r10 $r8
     * add b32 $r11 $r11 0x0 $c
-    * mov b32 $r8 c15[0x608]
+    * mov b32 $r8 c15[0x628]
     * st b128 wt g[$r10d+0x00] $r0q
     * st b128 wt g[$r10d+0x10] $r4q
     * st b32 wt g[$r10d+0x20] $r8
@@ -1128,14 +1128,14 @@ static const uint64_t nvc0_read_hw_sm_counters_code[] =
    0x2c00000028019c04ULL,
    0x2c0000002c01dc04ULL,
    0x190e0000fc81dc03ULL,
-   0x28007c1800029de4ULL,
-   0x28007c181002dde4ULL,
+   0x28007c1880029de4ULL,
+   0x28007c189002dde4ULL,
    0x7000c01050921c03ULL,
    0x80000000000021e7ULL,
    0x10000000c0821c02ULL,
    0x4801000020a29c03ULL,
    0x0800000000b2dc42ULL,
-   0x28007c1820021de4ULL,
+   0x28007c18a0021de4ULL,
    0x9400000000a01fc5ULL,
    0x9400000040a11fc5ULL,
    0x9400000080a21f85ULL,