Message-ID: <156cf812-a2de-4480-80dc-31c38ca0eb57@redhat.com>
Date: Tue, 9 Apr 2024 16:45:12 +0200
From: David Hildenbrand <david@...hat.com>
To: Ryan Roberts <ryan.roberts@....com>,
 Itaru Kitayama <itaru.kitayama@...ux.dev>
Cc: Catalin Marinas <Catalin.Marinas@....com>, Will Deacon <will@...nel.org>,
 Mark Rutland <Mark.Rutland@....com>, Ard Biesheuvel <ardb@...nel.org>,
 Donald Dutile <ddutile@...hat.com>, Eric Chanudet <echanude@...hat.com>,
 Linux ARM <linux-arm-kernel@...ts.infradead.org>,
 "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2 0/4] Speed up boot with faster linear map creation

On 09.04.24 16:39, Ryan Roberts wrote:
> On 09/04/2024 15:29, David Hildenbrand wrote:
>> On 09.04.24 16:13, Ryan Roberts wrote:
>>> On 09/04/2024 12:51, David Hildenbrand wrote:
>>>> On 09.04.24 13:29, David Hildenbrand wrote:
>>>>> On 09.04.24 13:22, David Hildenbrand wrote:
>>>>>> On 09.04.24 12:13, Itaru Kitayama wrote:
>>>>>>>
>>>>>>>
>>>>>>>> On Apr 9, 2024, at 19:04, Ryan Roberts <ryan.roberts@....com> wrote:
>>>>>>>>
>>>>>>>> On 09/04/2024 01:10, Itaru Kitayama wrote:
>>>>>>>>> Hi Ryan,
>>>>>>>>>
>>>>>>>>>> On Apr 8, 2024, at 16:30, Ryan Roberts <ryan.roberts@....com> wrote:
>>>>>>>>>>
>>>>>>>>>> On 06/04/2024 11:31, Itaru Kitayama wrote:
>>>>>>>>>>> Hi Ryan,
>>>>>>>>>>>
>>>>>>>>>>> On Sat, Apr 06, 2024 at 09:32:34AM +0100, Ryan Roberts wrote:
>>>>>>>>>>>> Hi Itaru,
>>>>>>>>>>>>
>>>>>>>>>>>> On 05/04/2024 08:39, Itaru Kitayama wrote:
>>>>>>>>>>>>> On Thu, Apr 04, 2024 at 03:33:04PM +0100, Ryan Roberts wrote:
>>>>>>>>>>>>>> Hi All,
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> It turns out that creating the linear map can take a significant
>>>>>>>>>>>>>> proportion of
>>>>>>>>>>>>>> the total boot time, especially when rodata=full. And most of the
>>>>>>>>>>>>>> time is spent
>>>>>>>>>>>>>> waiting on superfluous tlb invalidation and memory barriers. This
>>>>>>>>>>>>>> series reworks
>>>>>>>>>>>>>> the kernel pgtable generation code to significantly reduce the number
>>>>>>>>>>>>>> of those
>>>>>>>>>>>>>> TLBIs, ISBs and DSBs. See each patch for details.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> The below shows the execution time of map_mem() across a couple of
>>>>>>>>>>>>>> different
>>>>>>>>>>>>>> systems with different RAM configurations. We measure after applying
>>>>>>>>>>>>>> each patch
>>>>>>>>>>>>>> and show the improvement relative to base (v6.9-rc2):
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>                     | Apple M2 VM | Ampere Altra| Ampere Altra| Ampere
>>>>>>>>>>>>>> Altra
>>>>>>>>>>>>>>                     | VM, 16G     | VM, 64G     | VM, 256G    | Metal,
>>>>>>>>>>>>>> 512G
>>>>>>>>>>>>>> ---------------|-------------|-------------|-------------|-------------
>>>>>>>>>>>>>>                     |   ms    (%) |   ms    (%) |   ms    (%) |
>>>>>>>>>>>>>> ms    (%)
>>>>>>>>>>>>>> ---------------|-------------|-------------|-------------|-------------
>>>>>>>>>>>>>> base           |  153   (0%) | 2227   (0%) | 8798   (0%) | 17442
>>>>>>>>>>>>>> (0%)
>>>>>>>>>>>>>> no-cont-remap  |   77 (-49%) |  431 (-81%) | 1727 (-80%) |  3796
>>>>>>>>>>>>>> (-78%)
>>>>>>>>>>>>>> batch-barriers |   13 (-92%) |  162 (-93%) |  655 (-93%) |  1656
>>>>>>>>>>>>>> (-91%)
>>>>>>>>>>>>>> no-alloc-remap |   11 (-93%) |  109 (-95%) |  449 (-95%) |  1257
>>>>>>>>>>>>>> (-93%)
>>>>>>>>>>>>>> lazy-unmap     |    6 (-96%) |   61 (-97%) |  257 (-97%) |   838
>>>>>>>>>>>>>> (-95%)
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> This series applies on top of v6.9-rc2. All mm selftests pass. I've
>>>>>>>>>>>>>> compile and
>>>>>>>>>>>>>> boot tested various PAGE_SIZE and VA size configs.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> ---
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Changes since v1 [1]
>>>>>>>>>>>>>> ====================
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>        - Added Tested-by tags (thanks to Eric and Itaru)
>>>>>>>>>>>>>>        - Renamed ___set_pte() -> __set_pte_nosync() (per Ard)
>>>>>>>>>>>>>>        - Reordered patches (biggest impact & least controversial first)
>>>>>>>>>>>>>>        - Reordered alloc/map/unmap functions in mmu.c to aid reader
>>>>>>>>>>>>>>        - pte_clear() -> __pte_clear() in clear_fixmap_nosync()
>>>>>>>>>>>>>>        - Reverted generic p4d_index() which caused x86 build error.
>>>>>>>>>>>>>> Replaced with
>>>>>>>>>>>>>>          unconditional p4d_index() define under arm64.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> [1]
>>>>>>>>>>>>>> https://lore.kernel.org/linux-arm-kernel/20240326101448.3453626-1-ryan.roberts@arm.com/<https://lore.kernel.org/linux-arm-kernel/20240326101448.3453626-1-ryan.roberts@arm.com/>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Thanks,
>>>>>>>>>>>>>> Ryan
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> Ryan Roberts (4):
>>>>>>>>>>>>>>        arm64: mm: Don't remap pgtables per-cont(pte|pmd) block
>>>>>>>>>>>>>>        arm64: mm: Batch dsb and isb when populating pgtables
>>>>>>>>>>>>>>        arm64: mm: Don't remap pgtables for allocate vs populate
>>>>>>>>>>>>>>        arm64: mm: Lazily clear pte table mappings from fixmap
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> arch/arm64/include/asm/fixmap.h  |   5 +-
>>>>>>>>>>>>>> arch/arm64/include/asm/mmu.h     |   8 +
>>>>>>>>>>>>>> arch/arm64/include/asm/pgtable.h |  13 +-
>>>>>>>>>>>>>> arch/arm64/kernel/cpufeature.c   |  10 +-
>>>>>>>>>>>>>> arch/arm64/mm/fixmap.c           |  11 +
>>>>>>>>>>>>>> arch/arm64/mm/mmu.c              | 377 +++++++++++++++++++++++--------
>>>>>>>>>>>>>> 6 files changed, 319 insertions(+), 105 deletions(-)
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> -- 
>>>>>>>>>>>>>> 2.25.1
>>>>>>>>>>>>>>
>>>>>>>>>>>>>
>>>>>>>>>>>>> I've build and boot tested the v2 on FVP, base is taken from your
>>>>>>>>>>>>> linux-rr repo. Running run_vmtests.sh on v2 left some gup longterm not
>>>>>>>>>>>>> oks, would you take a look at it? The mm ksefltests used is from your
>>>>>>>>>>>>> linux-rr repo too.
>>>>>>>>>>>>
>>>>>>>>>>>> Thanks for taking a look at this.
>>>>>>>>>>>>
>>>>>>>>>>>> I can't reproduce your issue unfortunately; steps as follows on Apple
>>>>>>>>>>>> M2 VM:
>>>>>>>>>>>>
>>>>>>>>>>>> Config: arm64 defconfig + the following:
>>>>>>>>>>>>
>>>>>>>>>>>> # Squashfs for snaps, xfs for large file folios.
>>>>>>>>>>>> ./scripts/config --enable CONFIG_SQUASHFS_LZ4
>>>>>>>>>>>> ./scripts/config --enable CONFIG_SQUASHFS_LZO
>>>>>>>>>>>> ./scripts/config --enable CONFIG_SQUASHFS_XZ
>>>>>>>>>>>> ./scripts/config --enable CONFIG_SQUASHFS_ZSTD
>>>>>>>>>>>> ./scripts/config --enable CONFIG_XFS_FS
>>>>>>>>>>>>
>>>>>>>>>>>> # For general mm debug.
>>>>>>>>>>>> ./scripts/config --enable CONFIG_DEBUG_VM
>>>>>>>>>>>> ./scripts/config --enable CONFIG_DEBUG_VM_MAPLE_TREE
>>>>>>>>>>>> ./scripts/config --enable CONFIG_DEBUG_VM_RB
>>>>>>>>>>>> ./scripts/config --enable CONFIG_DEBUG_VM_PGFLAGS
>>>>>>>>>>>> ./scripts/config --enable CONFIG_DEBUG_VM_PGTABLE
>>>>>>>>>>>> ./scripts/config --enable CONFIG_PAGE_TABLE_CHECK
>>>>>>>>>>>>
>>>>>>>>>>>> # For mm selftests.
>>>>>>>>>>>> ./scripts/config --enable CONFIG_USERFAULTFD
>>>>>>>>>>>> ./scripts/config --enable CONFIG_TEST_VMALLOC
>>>>>>>>>>>> ./scripts/config --enable CONFIG_GUP_TEST
>>>>>>>>>>>>
>>>>>>>>>>>> Running on VM with 12G memory, split across 2 (emulated) NUMA nodes
>>>>>>>>>>>> (needed by
>>>>>>>>>>>> some mm selftests), with kernel command line to reserve hugetlbs and
>>>>>>>>>>>> other
>>>>>>>>>>>> features required by some mm selftests:
>>>>>>>>>>>>
>>>>>>>>>>>> "
>>>>>>>>>>>> transparent_hugepage=madvise earlycon root=/dev/vda2 secretmem.enable
>>>>>>>>>>>> hugepagesz=1G hugepages=0:2,1:2 hugepagesz=32M hugepages=0:2,1:2
>>>>>>>>>>>> default_hugepagesz=2M hugepages=0:64,1:64 hugepagesz=64K
>>>>>>>>>>>> hugepages=0:2,1:2
>>>>>>>>>>>> "
>>>>>>>>>>>>
>>>>>>>>>>>> Ubuntu userspace running off XFS rootfs. Build and run mm selftests
>>>>>>>>>>>> from same
>>>>>>>>>>>> git tree.
>>>>>>>>>>>>
>>>>>>>>>>>>
>>>>>>>>>>>> Although I don't think any of this config should make a difference to
>>>>>>>>>>>> gup_longterm.
>>>>>>>>>>>>
>>>>>>>>>>>> Looks like your errors are all "ftruncate() failed". I've seen this
>>>>>>>>>>>> problem on
>>>>>>>>>>>> our CI system. There it is due to running the tests from NFS file
>>>>>>>>>>>> system. What
>>>>>>>>>>>> filesystem are you using? Perhaps you are sharing into the FVP using
>>>>>>>>>>>> 9p? That
>>>>>>>>>>>> might also be problematic.
>>>>>>>>>>>
>>>>>>>>>>> That was it. This time I booted up the kernel including your series on
>>>>>>>>>>> QEMU on my M1 and executed the gup_longterm program without the ftruncate
>>>>>>>>>>> failures. When testing your kernel on FVP, I was executing the script
>>>>>>>>>>> from the FVP's host filesystem using 9p.
>>>>>>>>>>
>>>>>>>>>> I'm not sure exactly what the root cause is. Perhaps there isn't enough
>>>>>>>>>> space on
>>>>>>>>>> the disk? It might be worth enhancing the error log to provide the
>>>>>>>>>> errno in
>>>>>>>>>> tools/testing/selftests/mm/gup_longterm.c.
>>>>>>>>>>
>>>>>>>>>
>>>>>>>>> Attached is the strace’d gup_longterm executiong log on your
>>>>>>>>> pgtable-boot-speedup-v2 kernel.
>>>>>>>>
>>>>>>>> Sorry are you saying that it only fails with the pgtable-boot-speedup-v2
>>>>>>>> patch
>>>>>>>> set applied? I thought we previously concluded that it was independent of
>>>>>>>> that?
>>>>>>>> I was under the impression that it was filesystem related and not something
>>>>>>>> that
>>>>>>>> I was planning to investigate.
>>>>>>>
>>>>>>> No, irrespective of the kernel, if using 9p on FVP the test program fails.
>>>>>>> It is indeed 9p filesystem related, as I switched to using NFS all the
>>>>>>> issues are gone.
>>>>>>
>>>>>> Did it never work on 9p? If so, we might have to SKIP that test.
>>>>>>
>>>>>> openat(AT_FDCWD, "gup_longterm.c_tmpfile_BLboOt", O_RDWR|O_CREAT|O_EXCL,
>>>>>> 0600) = 3
>>>>>> unlinkat(AT_FDCWD, "gup_longterm.c_tmpfile_BLboOt", 0) = 0
>>>>>> fstatfs(3, 0xffffe505a840)              = -1 EOPNOTSUPP (Operation not
>>>>>> supported)
>>>>>> ftruncate(3, 4096)                      = -1 ENOENT (No such file or
>>>>>> directory)
>>>>>
>>>>> Note: I'm wondering if the unlinkat here is the problem that makes
>>>>> ftruncate() with 9p result in weird errors (e.g., the hypervisor
>>>>> unlinked the file and cannot reopen it for the fstatfs/ftruncate. ...
>>>>> which gives us weird errors here).
>>>>>
>>>>> Then, we should lookup the fs type in run_with_local_tmpfile() before
>>>>> the unlink() and simply skip the test if it is 9p.
>>>>
>>>> The unlink with 9p most certainly was a known issue in the past:
>>>>
>>>> https://gitlab.com/qemu-project/qemu/-/issues/103
>>>>
>>>> Maybe it's still an issue with older hypervisors (QEMU?)? Or it was never
>>>> completely resolved?
>>>
>>> I believe Itaru is running on FVP (Fixed Virtual Platform - "fast model" -
>>> Arm's architecture emulator). So QEMU won't be involved here. The FVP emulates
>>> a 9p device, so perhaps the bug is in there.
>>
>> Very likely.
>>
>>>
>>> Note that I see lots of "fallocate() failed" failures in gup_longterm when
>>> running on our CI system. This is a completely different setup; Real HW with
>>> Linux running bare metal using an NFS rootfs. I'm not sure if this is related.
>>> Logs show it failing consistently for the "tmpfile" and "local tmpfile" test
>>> configs. I also see a couple of these fails in the cow tests.
>>
>> What is the fallocate() errno you are getting? strace log would help (to see if
>> statfs also fails already)! Likely a similar NFS issue.
> 
> Unfortunately this is a system I don't have access to. I've requested some of
> this triage to be done, but its fairly low priority unfortunately.

To work around these BUGs (?) elsewhere, we could simply skip the test 
if get_fs_type() is not able to detect the FS type. Likely that's an 
early indicator that the unlink() messed something up.

.. doesn't feel right, though.

-- 
Cheers,

David / dhildenb

Powered by blists - more mailing lists