T HP support for PPC64
Get series with b4
b4 am [email protected]Versions newest first
- v1 Latest26 patches 6 repliesIn review
Changes from V1 from the cover letter
* Address review comments
* More patch split
* Add batch hpte invalidate for hugepages.
Patches in v1 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/26 | powerpc: Use signed formatting when printing errorarch/ | +1 | 0 replies | |
| 2/26 | powerpc: Save DAR and DSISR in pt_regs on MCEarch/ arch/ | +10 | 0 replies | |
| 3/26 | powerpc: Don't hard code the size of pte pagearch/ arch/ | +8 | 0 replies | |
| 4/26 | powerpc: Reduce the PTE_INDEX_SIZEarch/ | +2 | 0 replies | |
| 5/26 | powerpc: Move the pte free routines from common headerarch/ arch/ arch/ | +189 | 0 replies | |
| 6/26 | powerpc: Reduce PTE table memory wastagearch/ arch/ arch/ +4 more | +264 | 0 replies | |
| 7/26 | powerpc: Add size argument to pgtable_cache_addarch/ arch/ | +8 | 3 replies | |
| 8/26 | powerpc: Clarify __pgtable_cache_add by renaming shift to indexarch/ arch/ | +9 | 0 replies | |
| 9/26 | powerpc: Use encode avpn where we need only avpn valuesarch/ arch/ arch/ | +10 | 0 replies | |
| 10/26 | powerpc: Decode the pte-lp-encoding bits correctly.arch/ arch/ arch/ +7 more | +208 | 1 reply | |
| 11/26 | powerpc: Fix hpte_decode to use the correct decoding for page sizesarch/ | +22 | 1 reply | |
| 12/26 | powerpc: Return all the valid pte ecndoing in KVM_PPC_GET_SMMU_INFO ioctlarch/ | +9 | 1 reply | |
| 13/26 | powerpc: Update tlbie/tlbiel as per ISA docarch/ | +40 | 0 replies | |
| 14/26 | powerpc: print both base and actual page size on hash failurearch/ arch/ arch/ | +10 | 0 replies | |
| 15/26 | powerpc: Print page size info during bootarch/ | +5 | 0 replies | |
| 16/26 | mm/THP: HPAGE_SHIFT is not a #define on some archinclude/ mm/ | +6 | 0 replies | |
| 17/26 | mm/THP: Add pmd args to pgtable deposit and withdraw APIsarch/ arch/ arch/ +4 more | +27 | 0 replies | |
| 18/26 | mm/THP: withdraw the pgtable after pmdp related operationsmm/ | +2 | 0 replies | |
| 19/26 | powerpc/THP: Implement transparent huge pages for ppc64arch/ arch/ arch/ +5 more | +597 | 0 replies | |
| 20/26 | powerpc/THP: Differentiate THP PMD entries from HUGETLB PMD entriesarch/ arch/ arch/ | +17 | 0 replies | |
| 21/26 | powerpc/THP: Add code to handle HPTE faults for large pagesarch/ arch/ arch/ +11 more | +283 | 0 replies | |
| 22/26 | powerpc: Handle huge page in perf callchainarch/ | +21 | 0 replies | |
| 23/26 | powerpc/THP: hypervisor require few WIMG bit setarch/ | +2 | 0 replies | |
| 24/26 | powerpc/THP: get_user_pages_fast changesarch/ | +82 | 0 replies | |
| 25/26 | powerpc/THP: Enable THP on PPC64arch/ | +28 | 0 replies | |
| 26/26 | powerpc: Optimize hugepage invalidatearch/ arch/ arch/ +1 more | +210 | 0 replies |
Reviews trailers per patch
7 of 26 patches have a Reviewed-by or Acked-by.
| Patch | Paul Mackerras |
|---|---|
| 1/26 | A-b |
| 2/26 | A-b |
| 3/26 | A-b |
| 4/26 | A-b |
| 5/26 | A-b |
| 6/26 | |
| 7/26 | |
| 8/26 | |
| 9/26 | A-b |
| 10/26 | |
| 11/26 | A-b |
| 12/26 | |
| 13/26 | |
| 14/26 | |
| 15/26 | |
| 16/26 | |
| 17/26 | |
| 18/26 | |
| 19/26 | |
| 20/26 | |
| 21/26 | |
| 22/26 | |
| 23/26 | |
| 24/26 | |
| 25/26 | |
| 26/26 |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [PATCH -V2 00/26]T HP support for PPC64
Hi,
This patchset adds transparent huge page support for PPC64.
TODO:
* ppc64 KVM related changes
* powernv still doesn't boot
* hash preload support in update_mmu_cache_pmd
Some numbers:
The latency measurements code from Anton found at
http://ozlabs.org/~anton/junkcode/latency2001.c
THP disabled 64K page size
------------------------
[root@llmp24l02 ~]# ./latency2001 8G
8589934592 731.73 cycles 205.77 ns
[root@llmp24l02 ~]# ./latency2001 8G
8589934592 743.39 cycles 209.05 ns
[root@llmp24l02 ~]#
THP disabled large page via hugetlbfs
-------------------------------------
[root@llmp24l02 ~]# ./latency2001 -l 8G
8589934592 416.09 cycles 117.01 ns
[root@llmp24l02 ~]# ./latency2001 -l 8G
8589934592 415.74 cycles 116.91 ns
THP enabled 64K page size.
----------------
[root@llmp24l02 ~]# ./latency2001 8G
8589934592 405.07 cycles 113.91 ns
[root@llmp24l02 ~]# ./latency2001 8G
8589934592 411.82 cycles 115.81 ns
[root@llmp24l02 ~]#
We are close to hugetlbfs in latency and we can achieve this with zero config/page reservation. Most of the allocations above are fault allocated. I haven't really measured the collapse alloc impact.
Another test that does 50000000 random access over 1GB area goes from 2.65 seconds to 1.07 seconds with this patchset.
Changes from V1
* Address review comments
* More patch split
* Add batch hpte invalidate for hugepages.
Changes from RFC V2:
* Address review comments
* More code cleanup and patch split
Changes from RFC V1:
* HugeTLB fs now works
* Compile issues fixed
* rebased to v3.8
* Patch series reorded so that ppc64 cleanups and MM THP changes are moved early in the series. This should help in picking those patches early.
Thanks,
-aneesh