On 7/28/20 1:12 AM, Yu Xu wrote:
On 7/27/20 7:05 PM, Catalin Marinas wrote:
On Mon, Jul 27, 2020 at 03:31:16PM +0800, Yu Xu wrote:
On 7/25/20 4:22 AM, Linus Torvalds wrote:
On Fri, Jul 24, 2020 at 12:27 PM Linus Torvalds
<torvalds@xxxxxxxxxxxxxxxxxxxx> wrote:
It *may* make sense to say "ok, don't bother flushing the TLB if this
is a retry, because we already did that originally". MAYBE.
[...]
We could say that we never need it at all for FAULT_FLAG_RETRY. That
makes a lot of sense to me.
So a patch that does something like the appended (intentionally
whitespace-damaged) seems sensible.
I tested your patch on our aarch64 box, with 128 online CPUs.
[...]
There are two points to sum up.
1) the performance of page_fault3_process is restored, while the
performance
of page_fault3_thread is about ~80% of the vanilla, except the case
of 128
threads.
2) in the case of 128 threads, test worker threads seem to get stuck,
making
no progress in the iterations of mmap-write-munmap until a period of
time
later. the test result is 0 because only first 16 samples are
counted, and
they are all 0. This situation is easy to re-produce with large
number of
threads (not necessarily 128), and the stack of one stuck thread is
shown
below.
[<0>] __switch_to+0xdc/0x150
[<0>] wb_wait_for_completion+0x84/0xb0
[<0>] __writeback_inodes_sb_nr+0x9c/0xe8
[<0>] try_to_writeback_inodes_sb+0x6c/0x88
[<0>] ext4_nonda_switch+0x90/0x98 [ext4]
[<0>] ext4_page_mkwrite+0x248/0x4c0 [ext4]
[<0>] do_page_mkwrite+0x4c/0x100
[<0>] do_fault+0x2ac/0x3e0
[<0>] handle_pte_fault+0xb4/0x258
[<0>] __handle_mm_fault+0x1d8/0x3a8
[<0>] handle_mm_fault+0x104/0x1d0
[<0>] do_page_fault+0x16c/0x490
[<0>] do_translation_fault+0x60/0x68
[<0>] do_mem_abort+0x58/0x100
[<0>] el0_da+0x24/0x28
[<0>] 0xffffffffffffffff
It seems quite normal, right? and I've run out of ideas.
If threads get stuck here, it could be a stale TLB entry that's not
flushed with Linus' patch. Since that's a write fault, I think it hits
the FAULT_FLAG_TRIED case.
There must be some changes in my test box, because I find that even the
vanilla kernel (89b15332af7c^) get result of 0 in 128t testcase. And I
just directly used the history test data as the baseline. I will dig
into this then.
Hi all, I reset the test box, and re-run the whole test, the result this
time makes more sense.
Test 89b15332a^ Linus Catalin Yang
1p 100 90.10 79.20 86.19 %
1t 100 89.56 88.74 92.21 %
32p 100 98.22 97.36 98.91 %
32t 100 75.45 76.06 75.75 %
64p 100 99.97 100.01 99.97 %
64t 100 70.44 74.53 61.75 %
96p 100 99.95 99.91 100.00 %
96t 100 67.95 72.56 63.88 %
128p 100 99.92 99.93 100.12 %
128t 100 73.23 73.85 73.16 %
Sorry for previously confusing test data. Performance drop in thread
mode is now the remaining issue.
Thanks
Yu
And do we still need to concern the ~20% performance drop in thread mode?
Could you give my patch here a try as an alternative:
https://lore.kernel.org/linux-mm/20200725155841.GA14490@gaia/
I ran the same test on the same aarch64 box, with your patch, the result
is as follows.
test vanilla kernel patched kernel
parameter (89b15332af7c^) (Catalin's patch)
1p 829299 787676 (96.36 %)
1t 998007 789284 (78.36 %)
32p 18916718 17921100 (94.68 %)
32t 2020918 1644146 (67.64 %)
64p 18965168 18983580 (100.0 %)
64t 1415404 1093750 (48.03 %)
96p 18949438 18963921 (100.1 %)
96t 1622876 1262878 (63.72 %)
128p 18926813 1680146 (8.89 %)
128t 1643109 0 (0.00 % ) # ignore this temporarily
Thanks
Yu
It leaves the spurious flush in place but only local (though note that
in a guest under KVM, all local TLBIs are upgraded to inner-shareable,
so you'd not get the performance benefit).
