experiments/dmp.py: DMP/GoFetch bug test experiment

Signed-off-by: Hector Martin <marcan@marcan.st>

Author: marcan

proxyclient/experiments/dmp.py

@@ -0,0 +1,219 @@
+#!/usr/bin/env python3
+# SPDX-License-Identifier: MIT
+import sys, pathlib, time
+sys.path.append(str(pathlib.Path(__file__).resolve().parents[1]))
+
+from m1n1.setup import *
+from m1n1 import asm
+
+REPETITIONS = 64
+
+PAGE_SIZE = 16384
+
+TEST_ECORE = 1
+TEST_PCORE = 4
+
+L2_LINE_SIZE = 128
+PNRG_a = 75
+PRNG_m = 31337
+rnd_idx = 8
+
+def prng(x):
+    return (PNRG_a * x) % PRNG_m
+
+SIZE_DATA_ARRAY = (PRNG_m * L2_LINE_SIZE)
+
+data_buf_addr = u.memalign(PAGE_SIZE, SIZE_DATA_ARRAY)
+p.memset64(data_buf_addr, 0x5555555555555555, SIZE_DATA_ARRAY)
+aop_addr = u.memalign(PAGE_SIZE, PAGE_SIZE)
+p.memset64(aop_addr, 0x5555555555555555, PAGE_SIZE)
+
+freq = u.mrs(CNTFRQ_EL0)
+code = u.malloc(0x1000)
+
+util = asm.ARMAsm("""
+test:
+    dc civac, x0
+    dc civac, x1
+    isb sy
+
+    mov x7, #0x8000
+1:
+    add x2, x2, #1
+    mul x2, x2, x2
+    sub x7, x7, #1
+    cbnz x7, 1b
+    and x2, x2, #(15 << 60)
+
+    add x1, x1, x2
+    ldrb w2, [x1, #512]
+    and x2, x2, #(15 << 60)
+
+    add x0, x0, x2
+
+    dsb sy
+    isb
+    mrs x9, S3_2_c15_c0_0 // PMC0_EL1
+    isb
+    ldr x2, [x0, x2]
+    isb
+    mrs x10, S3_2_c15_c0_0
+    sub x5, x10, x9
+
+    and x2, x2, #(15 << 60)
+    mov x7, #0x4000
+1:
+    add x2, x2, #1
+    mul x2, x2, x2
+    sub x7, x7, #1
+    cbnz x7, 1b
+
+    and x2, x2, #(15 << 60)
+
+    dsb sy
+    isb
+    mrs x9, S3_2_c15_c0_0
+    isb
+    ldr x2, [x1, x2]
+    isb
+    mrs x10, S3_2_c15_c0_0
+    sub x0, x10, x9
+
+    isb sy
+
+    lsl x5, x5, #32
+    orr x0, x0, x5
+    ret
+""", code)
+for i in util.disassemble():
+    print(i)
+iface.writemem(code, util.data)
+p.dc_cvau(code, len(util.data))
+p.ic_ivau(code, len(util.data))
+
+# Set higher cpufreq pstate on all clusters
+p.cpufreq_init()
+p.smp_start_secondaries()
+p.smp_set_wfe_mode(True);
+
+def cpu_call(cpu, x, *args):
+    return p.smp_call_sync(cpu, x | REGION_RX_EL1, *args)
+
+def init_core(cpu):
+    p.mmu_init_secondary(cpu)
+
+    def mrs(x):
+        return u.mrs(x, call=lambda x, *args: cpu_call(cpu, x, *args))
+    def msr(x, v):
+        u.msr(x, v, call=lambda x, *args: cpu_call(cpu, x, *args))
+
+    is_ecore = not (mrs(MPIDR_EL1) & (1 << 16))
+    # Enable DC MVA ops
+    v = mrs(EHID4_EL1 if is_ecore else HID4_EL1)
+    v &= ~(1 << 11)
+    msr(EHID4_EL1 if is_ecore else HID4_EL1, v)
+
+    # Enable PMU
+    v = mrs(PMCR0_EL1)
+    v |= 1 | (1<<30)
+    msr(PMCR0_EL1, v)
+    msr(PMCR1_EL1, 0xffffffffffffffff)
+
+    # Enable TBI
+    v = mrs(TCR_EL1)
+    v |= (1 << 37)
+    msr(TCR_EL1, v)
+
+    # Enable user cache ops
+    v = mrs(SCTLR_EL1)
+    v |= (1 << 26)
+    msr(SCTLR_EL1, v)
+
+init_core(TEST_ECORE)
+init_core(TEST_PCORE)
+
+# Enable DC MVA ops
+v = u.mrs(EHID4_EL1)
+v &= ~(1 << 11)
+u.msr(EHID4_EL1, v)
+
+def test_cpu(cpu, mask):
+    global rnd_idx
+
+    total_aop = total_ptr = 0
+    p.memset64(data_buf_addr, 0x5555555555555555, SIZE_DATA_ARRAY)
+    p.memset64(aop_addr, 0x5555555555555555, PAGE_SIZE)
+    for i in range(REPETITIONS):
+        test_offset = L2_LINE_SIZE * rnd_idx
+        test_addr = data_buf_addr + test_offset
+
+        p.write64(aop_addr, test_addr | mask | REGION_RWX_EL0)
+        p.dc_civac(aop_addr, L2_LINE_SIZE)
+        # p.dc_civac(data_buf_addr, SIZE_DATA_ARRAY)
+
+        elapsed = p.smp_call_sync_el0(cpu, util.test | REGION_RWX_EL0, aop_addr | REGION_RWX_EL0, test_addr | REGION_RWX_EL0, 7 << 60)
+        time_aop = elapsed >> 32
+        time_ptr = elapsed & 0xffffffff
+        total_aop += time_aop
+        total_ptr += time_ptr
+
+        rnd_idx = prng(rnd_idx)
+
+    return total_aop, total_ptr
+
+
+print("ECore plain:", test_cpu(TEST_ECORE, 0))
+print("ECore mask: ", test_cpu(TEST_ECORE, 0xaaaaaaaa00000000))
+print("PCore plain:", test_cpu(TEST_PCORE, 0))
+print("PCore mask: ", test_cpu(TEST_PCORE, 0xaaaaaaaa00000000))
+
+for reg in (
+    # "HID0_EL1",
+    # "HID1_EL1",
+    # "HID2_EL1",
+    # "HID3_EL1",
+    "HID4_EL1",
+    # "HID5_EL1",
+    # "HID6_EL1",
+    # "HID7_EL1",
+    # "HID8_EL1",
+    # "HID9_EL1",
+    # "HID10_EL1",
+    "HID11_EL1",
+    # "HID13_EL1",
+    # "HID14_EL1",
+    # "HID16_EL1",
+    # "HID17_EL1",
+    # "HID18_EL1",
+    "HID21_EL1",
+    # "HID26_EL1",
+    # "HID27_EL1",
+):
+
+    cpu = TEST_PCORE
+    hid = u.mrs(reg, call=lambda x, *args: cpu_call(cpu, x, *args))
+
+    for i in range(64):
+        if (reg, i) not in (
+            ("HID4_EL1", 4),
+            ("HID11_EL1", 30),
+            ("HID21_EL1", 40),
+            ):
+            continue
+
+        bit = (1 << i)
+        print(f"Test {reg} bit {i}:", end=" ")
+
+        u.msr(reg, hid ^ bit, call=lambda x, *args: cpu_call(cpu, x, *args))
+
+        tval = test_cpu(cpu, 0)[1]
+        control = test_cpu(cpu, 0xaaaaaaaa00000000)[1]
+
+
+        if tval < (0.75 * control):
+            print(f"DMP active {tval} {control}")
+        else:
+            print(f"DMP INACTIVE {tval} {control}")
+
+    u.msr(reg, hid, call=lambda x, *args: cpu_call(cpu, x, *args))
+