[Fix] (MG_Remote, P11 T0): the T0 POST also checks what the CPU writes through the held lock after the import and its first GPU use, in both directions - Mali-G1-Ultra passed the old single round (the import cleans the CPU cache) while its GPU saw 0/1024 post-import words, so T0 was granted and create-indirect-in-world-align1024 x Magma lost every Flywheel object on inproc/spawn (0.867/0.848); it now fails there and runs T2 (0.999984), and the Redmi passes both rounds and keeps T0

This commit is contained in:
2026-09-30 10:21:46 -04:00
parent 028cf4eb62
commit 378020adf3
7 changed files with 137 additions and 18 deletions
@@ -786,7 +786,9 @@ namespace MobileGL::MG_Backend::DirectVulkan {
// read - this server's own host map of the import reads what the CPU wrote through the lock;
// gpu - one submission copies that pattern out of the import (the GPU read) and fills another
// region of it (the GPU write), then a transfer->host barrier and a fence wait;
// back - the fill is read through the HELD pointer, the copy from a host-visible staging buffer.
// back - the fill is read through the HELD pointer, the copy from a host-visible staging buffer;
// steady - the same again with bytes the CPU wrote AFTER the import and its first GPU use (the
// block at the end says why the first round alone is not enough).
Bool VkBufferManager::SelfTestWireImport(char* whyOut, Uint64 whyBytes) {
const auto answer = [&](Bool ok, const String& text) {
std::snprintf(whyOut, static_cast<SizeT>(whyBytes), "%s", text.c_str());
@@ -909,6 +911,55 @@ namespace MobileGL::MG_Backend::DirectVulkan {
String(imported.coherent ? "coherent" : "non-coherent") + " memory), GPU read " +
std::to_string(gpuRead) + "/1024 words of the CPU's pattern, GPU write seen through the "
"held lock " + std::to_string(heldSaw) + "/1024 words";
if (!ok) break;
// STEADY STATE. The round above reads bytes the CPU wrote BEFORE the import, and an import
// can clean the CPU's caches on its way in, so it passes on a device whose GPU never sees a
// write made through the held lock afterwards - which is all of T0's real traffic (Mali-G1
// on a Dimensity: the first round passed, then Flywheel's staging scatter read stale bytes
// and the create-indirect trace lost every instanced object). So the CPU now writes a fresh
// pattern into region 1 and reads region 4 back into its cache, and one more submission
// copies region 1 out and overwrites region 4.
constexpr Uint64 kSteadyAt = 1 * 4096; // region 1
constexpr Uint32 kSteadyFillWord = 0x3D5EA11Eu;
auto* heldWrite = static_cast<volatile Uint32*>(held.ptr);
Uint32 cached = 0;
for (Uint32 i = 0; i < kWords; ++i) {
heldWrite[kSteadyAt / 4 + i] = 0xA5000000u ^ (i * 40503u);
cached ^= heldWrite[kFillAt / 4 + i];
}
(void)cached;
std::memset(staging.mapped, 0, kWords * 4);
if (vkResetFences(ctx.device, 1, &fence) != VK_SUCCESS ||
vkResetCommandPool(ctx.device, pool, 0) != VK_SUCCESS ||
vkBeginCommandBuffer(commands, &begin) != VK_SUCCESS) {
ok = false;
verdict = "steady: the second submission could not be recorded";
break;
}
vkCmdPipelineBarrier(commands, VK_PIPELINE_STAGE_HOST_BIT, VK_PIPELINE_STAGE_TRANSFER_BIT, 0, 1,
&hostWrites, 0, nullptr, 0, nullptr);
VkBufferCopy steadyRegion{kSteadyAt, 0, kWords * 4};
vkCmdCopyBuffer(commands, imported.buffer, staging.buffer, 1, &steadyRegion);
vkCmdFillBuffer(commands, imported.buffer, kFillAt, kWords * 4, kSteadyFillWord);
vkCmdPipelineBarrier(commands, VK_PIPELINE_STAGE_TRANSFER_BIT, VK_PIPELINE_STAGE_HOST_BIT, 0, 1, &toHost,
0, nullptr, 0, nullptr);
vkEndCommandBuffer(commands);
if (vkQueueSubmit(ctx.queue, 1, &submit, fence) != VK_SUCCESS ||
vkWaitForFences(ctx.device, 1, &fence, VK_TRUE, 2'000'000'000ull) != VK_SUCCESS) {
ok = false;
verdict = "steady: the second submission did not complete within 2 s";
break;
}
Uint32 steadyRead = 0, steadySaw = 0;
for (Uint32 i = 0; i < kWords; ++i) {
if (copied[i] == (0xA5000000u ^ (i * 40503u))) ++steadyRead;
if (heldWords[kFillAt / 4 + i] == kSteadyFillWord) ++steadySaw;
}
ok = steadyRead == kWords && steadySaw == kWords;
verdict += "; with the import in use, GPU read " + std::to_string(steadyRead) +
"/1024 words the CPU wrote after it, GPU write seen through the held lock " +
std::to_string(steadySaw) + "/1024 words";
} while (false);
if (fence != VK_NULL_HANDLE) vkDestroyFence(ctx.device, fence, nullptr);
if (pool != VK_NULL_HANDLE) vkDestroyCommandPool(ctx.device, pool, nullptr);