derek/gem5
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

dev-amdgpu: Add SDMA device translation helper

Browse Source 
Adding a helper function to remove duplicate code in the copy packet
methods. Adds more comments on that code to explain what it is doing.
This could in theory also be used in other packets in the future.

Change-Id: Id0ed50c87260a2f12f53cb14e927f8c49bb99072
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/62718
Reviewed-by: Jason Lowe-Power <power.jg@gmail.com>
Tested-by: kokoro <noreply+kokoro@google.com>
Reviewed-by: Matt Sinclair <mattdsinclair@gmail.com>
Maintainer: Matt Sinclair <mattdsinclair@gmail.com>
Maintainer: Jason Lowe-Power <power.jg@gmail.com>
This commit is contained in:
Matthew Poremba
2022-08-27 08:28:29 -07:00
parent f20d12656a
commit e0e2806fc4
2 changed files with 52 additions and 39 deletions
Download Patch File Download Diff File Expand all files Collapse all files

84
src/dev/amdgpu/sdma_engine.cc
View File

@@ -96,6 +96,39 @@ SDMAEngine::getGARTAddr(Addr addr) const
return addr;
}
Addr
SDMAEngine::getDeviceAddress(Addr raw_addr)
{
// SDMA packets can access both host and device memory as either a source
// or destination address. We don't know which until it is translated, so
// we do a dummy functional translation to determine if the address
// resides in system memory or not.
auto tgen = translate(raw_addr, 64);
auto addr_range = *(tgen->begin());
Addr tmp_addr = addr_range.paddr;
DPRINTF(SDMAEngine, "getDeviceAddress raw_addr %#lx -> %#lx\n",
raw_addr, tmp_addr);
// SDMA packets will access device memory through the MMHUB aperture in
// supervisor mode (vmid == 0) and in user mode (vmid > 0). In the case
// of vmid == 0 the address is already an MMHUB address in the packet,
// so simply subtract the MMHUB base. For vmid > 0 the address is a
// virtual address that must first be translated. The translation will
// return an MMHUB address, then we can similarly subtract the base to
// get the device address. Otherwise, for host, device address is 0.
Addr device_addr = 0;
if ((gpuDevice->getVM().inMMHUB(raw_addr) && cur_vmid == 0) ||
(gpuDevice->getVM().inMMHUB(tmp_addr) && cur_vmid != 0)) {
if (cur_vmid == 0) {
device_addr = raw_addr - gpuDevice->getVM().getMMHUBBase();
} else {
device_addr = tmp_addr - gpuDevice->getVM().getMMHUBBase();
}
}
return device_addr;
}
/**
* GPUController will perform DMA operations on VAs, and because
* page faults are not currently supported for GPUController, we
@@ -533,31 +566,16 @@ SDMAEngine::copy(SDMAQueue *q, sdmaCopy *pkt)
pkt->source = getGARTAddr(pkt->source);
DPRINTF(SDMAEngine, "GART addr %lx\n", pkt->source);
// Check if the source is MMHUB. If it is we need to use the GpuMemMgr
// to read memory and not dmaReadVirt.
auto tgen = translate(pkt->source, 64);
auto addr_range = *(tgen->begin());
Addr tmp_addr = addr_range.paddr;
DPRINTF(SDMAEngine, "Tmp addr %#lx -> %#lx\n", pkt->source, tmp_addr);
if ((gpuDevice->getVM().inMMHUB(pkt->source) && cur_vmid == 0) ||
(gpuDevice->getVM().inMMHUB(tmp_addr) && cur_vmid != 0)) {
Addr mmhubAddr = 0;
if (cur_vmid == 0) {
mmhubAddr = pkt->source - gpuDevice->getVM().getMMHUBBase();
} else {
mmhubAddr = tmp_addr - gpuDevice->getVM().getMMHUBBase();
}
uint8_t *dmaBuffer = new uint8_t[pkt->count];
DPRINTF(SDMAEngine, "Copying from device address %#lx\n", mmhubAddr);
// Read data from the source first, then call the copyReadData method
uint8_t *dmaBuffer = new uint8_t[pkt->count];
Addr device_addr = getDeviceAddress(pkt->source);
if (device_addr) {
DPRINTF(SDMAEngine, "Copying from device address %#lx\n", device_addr);
auto cb = new EventFunctionWrapper(
[ = ]{ copyReadData(q, pkt, dmaBuffer); }, name());
gpuDevice->getMemMgr()->readRequest(mmhubAddr, dmaBuffer, pkt->count,
gpuDevice->getMemMgr()->readRequest(device_addr, dmaBuffer, pkt->count,
0, cb);
} else {
// first we have to read needed data from the source address
uint8_t *dmaBuffer = new uint8_t[pkt->count];
auto cb = new DmaVirtCallback<uint64_t>(
[ = ] (const uint64_t &) { copyReadData(q, pkt, dmaBuffer); });
dmaReadVirt(pkt->source, pkt->count, cb, (void *)dmaBuffer);
@@ -577,26 +595,14 @@ SDMAEngine::copyReadData(SDMAQueue *q, sdmaCopy *pkt, uint8_t *dmaBuffer)
}
delete [] dmaBuffer64;
// Aperture is unknown until translating. Do a dummy translation.
auto tgen = translate(pkt->dest, 64);
auto addr_range = *(tgen->begin());
Addr tmp_addr = addr_range.paddr;
DPRINTF(SDMAEngine, "Tmp addr %#lx -> %#lx\n", pkt->dest, tmp_addr);
// Writing generated data to the destination address.
if ((gpuDevice->getVM().inMMHUB(pkt->dest) && cur_vmid == 0) ||
(gpuDevice->getVM().inMMHUB(tmp_addr) && cur_vmid != 0)) {
Addr mmhubAddr = 0;
if (cur_vmid == 0) {
mmhubAddr = pkt->dest - gpuDevice->getVM().getMMHUBBase();
} else {
mmhubAddr = tmp_addr - gpuDevice->getVM().getMMHUBBase();
}
DPRINTF(SDMAEngine, "Copying to device address %#lx\n", mmhubAddr);
Addr device_addr = getDeviceAddress(pkt->dest);
// Write read data to the destination address then call the copyDone method
if (device_addr) {
DPRINTF(SDMAEngine, "Copying to device address %#lx\n", device_addr);
auto cb = new EventFunctionWrapper(
[ = ]{ copyDone(q, pkt, dmaBuffer); }, name());
gpuDevice->getMemMgr()->writeRequest(mmhubAddr, dmaBuffer, pkt->count,
0, cb);
gpuDevice->getMemMgr()->writeRequest(device_addr, dmaBuffer,
pkt->count, 0, cb);
} else {
auto cb = new DmaVirtCallback<uint64_t>(
[ = ] (const uint64_t &) { copyDone(q, pkt, dmaBuffer); });

7
src/dev/amdgpu/sdma_engine.hh
View File

@@ -155,6 +155,13 @@ class SDMAEngine : public DmaVirtDevice
Addr getGARTAddr(Addr addr) const;
TranslationGenPtr translate(Addr vaddr, Addr size) override;
/**
* Translate an address in an SDMA packet. Return the device address if
* address in the packet is on the device and 0 if the the address in the
* packet is on the host/system memory.
*/
Addr getDeviceAddress(Addr raw_addr);
/**
* Inherited methods.
*/