Matthew Poremba has submitted this change. (
https://gem5-review.googlesource.com/c/public/gem5/+/64451?usp=email )
Change subject: arch-vega: Implement PDE2 and PDE1 as PTE
......................................................................
arch-vega: Implement PDE2 and PDE1 as PTE
Page directory entries (PDEs) can be interpreted as leaf node page
table entries (PTEs) if the "p" bit is set. This is used for flexible
page sizes in Vega. Currently there is only support for PDE level 0
entries which can be interpreted as 2MB pages. This changeset adds
support for PDE1 and PDE2 which can be used to represent 1GB and 512GB
pages. PDE1-as-PTE entries can be tested and were verified on
applications by allocating >2GB of data. PDE0 is untested due to being
too large for simulation, but the implementation is similar to PDE0
and PDE1.
Change-Id: I801cbb5ec79110d57d2db760cc689c2e5778f9bb
Reviewed-on: https://gem5-review.googlesource.com/c/public/gem5/+/64451
Maintainer: Matt Sinclair <mattdsincl...@gmail.com>
Tested-by: kokoro <noreply+kok...@google.com>
Reviewed-by: Matt Sinclair <mattdsincl...@gmail.com>
---
M src/arch/amdgpu/vega/pagetable_walker.cc
1 file changed, 43 insertions(+), 2 deletions(-)
Approvals:
kokoro: Regressions pass
Matt Sinclair: Looks good to me, approved; Looks good to me, approved
diff --git a/src/arch/amdgpu/vega/pagetable_walker.cc
b/src/arch/amdgpu/vega/pagetable_walker.cc
index 2a8238f..96ac0fe 100644
--- a/src/arch/amdgpu/vega/pagetable_walker.cc
+++ b/src/arch/amdgpu/vega/pagetable_walker.cc
@@ -245,7 +245,16 @@
switch(state) {
case PDE2:
- fatal_if(pde.p, "Fragment in PDE2 not implemented");
+ if (pde.p) {
+ DPRINTF(GPUPTWalker, "Treating PDE2 as PTE: %#016x frag: %d\n",
+ (uint64_t)pte, pte.fragment);
+ entry.pte = pte;
+ int fragment = pte.fragment;
+ entry.logBytes = PageShift + std::min(3*9, fragment);
+ entry.vaddr <<= PageShift;
+ entry.vaddr = entry.vaddr & ~mask(entry.logBytes);
+ doEndWalk = true;
+ }
// Read the pde1Addr
part1 = ((((uint64_t)pte) >> 6) << 3);
@@ -257,7 +266,16 @@
nextState = PDE1;
break;
case PDE1:
- fatal_if(pde.p, "Fragment in PDE1 not implemented");
+ if (pde.p) {
+ DPRINTF(GPUPTWalker, "Treating PDE1 as PTE: %#016x frag: %d\n",
+ (uint64_t)pte, pte.fragment);
+ entry.pte = pte;
+ int fragment = pte.fragment;
+ entry.logBytes = PageShift + std::min(2*9, fragment);
+ entry.vaddr <<= PageShift;
+ entry.vaddr = entry.vaddr & ~mask(entry.logBytes);
+ doEndWalk = true;
+ }
// Read the pde0Addr
part1 = ((((uint64_t)pte) >> 6) << 3);
--
To view, visit
https://gem5-review.googlesource.com/c/public/gem5/+/64451?usp=email
To unsubscribe, or for help writing mail filters, visit
https://gem5-review.googlesource.com/settings
Gerrit-Project: public/gem5
Gerrit-Branch: develop
Gerrit-Change-Id: I801cbb5ec79110d57d2db760cc689c2e5778f9bb
Gerrit-Change-Number: 64451
Gerrit-PatchSet: 2
Gerrit-Owner: Matthew Poremba <matthew.pore...@amd.com>
Gerrit-Reviewer: Alexandru Duțu (Alex) <alexandru.d...@amd.com>
Gerrit-Reviewer: Jason Lowe-Power <power...@gmail.com>
Gerrit-Reviewer: Matt Sinclair <mattdsincl...@gmail.com>
Gerrit-Reviewer: Matthew Poremba <matthew.pore...@amd.com>
Gerrit-Reviewer: kokoro <noreply+kok...@google.com>
Gerrit-MessageType: merged
_______________________________________________
gem5-dev mailing list -- gem5-dev@gem5.org
To unsubscribe send an email to gem5-dev-le...@gem5.org
