mirror of
https://mirrors.bfsu.edu.cn/git/linux.git
synced 2024-11-28 06:34:12 +08:00
dd81dc0559
This series aims to improve the scalability of XFS transaction commits on large CPU count machines. My 32p machine hits contention limits in xlog_cil_commit() at about 700,000 transaction commits a section. It hits this at 16 thread workloads, and 32 thread workloads go no faster and just burn CPU on the CIL spinlocks. This patchset gets rid of spinlocks and global serialisation points in the xlog_cil_commit() path. It does this by moving to a combination of per-cpu counters, unordered per-cpu lists and post-ordered per-cpu lists. This results in transaction commit rates exceeding 1.4 million commits/s under unlink certain workloads, and while the log lock contention is largely gone there is still significant lock contention in the VFS (dentry cache, inode cache and security layers) at >600,000 transactions/s that still limit scalability. The changes to the CIL accounting and behaviour, combined with the structural changes to xlog_write() in prior patchsets make the per-cpu restructuring possible and sane. This allows us to move to precalculated reservation requirements that allow for reservation stealing to be accounted across multiple CPUs accurately. That is, instead of trying to account for continuation log opheaders on a "growth" basis, we pre-calculate how many iclogs we'll need to write out a maximally sized CIL checkpoint and steal that reserveD that space one commit at a time until the CIL has a full reservation. If we ever run a commit when we are already at the hard limit (because post-throttling) we simply take an extra reservation from each commit that is run when over the limit. Hence we don't need to do space usage math in the fast path and so never need to sum the per-cpu counters in this fast path. Similarly, per-cpu lists have the problem of ordering - we can't remove an item from a per-cpu list if we want to move it forward in the CIL. We solve this problem by using an atomic counter to give every commit a sequence number that is copied into the log items in that transaction. Hence relogging items just overwrites the sequence number in the log item, and does not move it in the per-cpu lists. Once we reaggregate the per-cpu lists back into a single list in the CIL push work, we can run it through list-sort() and reorder it back into a globally ordered list. This costs a bit of CPU time, but now that the CIL can run multiple works and pipelines properly, this is not a limiting factor for performance. It does increase fsync latency when the CIL is full, but workloads issuing large numbers of fsync()s or sync transactions end up with very small CILs and so the latency impact or sorting is not measurable for such workloads. OVerall, this pushes the transaction commit bottleneck out to the lockless reservation grant head updates. These atomic updates don't start to be a limiting fact until > 1.5 million transactions/s are being run, at which point the accounting functions start to show up in profiles as the highest CPU users. Still, this series doubles transaction throughput without increasing CPU usage before we get to that cacheline contention breakdown point... ` Signed-off-by: Dave Chinner <dchinner@redhat.com> -----BEGIN PGP SIGNATURE----- iQJIBAABCgAyFiEEmJOoJ8GffZYWSjj/regpR/R1+h0FAmLHai8UHGRhdmlkQGZy b21vcmJpdC5jb20ACgkQregpR/R1+h3JZQ//bb9HyBiBkeuK9MvqH40hOfazfGXD 8+pdP9r22qWp9LHhjz/EtH4Wy1sYe6a99mtPxqlsT3DqSl8GiolA1VFn+T3Sadu4 nqmB/ppzMLE0LLzKoVrb3/Zw+mEaz5Is3WLpr86CpK5gNW6gBHCj4B68lWiBtvjs OW5fTm0E44BnNORh/AdSUkJxxEB2OQhVk5omY/Op8vO5frviG5yqYakAeoQ3vFpS UKadwlGjei91c63g9se360Re+DXTBhzbgXz0oNV4YbgWba2O9lnut5zqlcJMvVAU YgGBxttT0OqCdSNp0vtwOG8UFeUqfWSY+AFwfDkNycltLASvU53efqC94kQHouoh 9++2VrPwPg0KOcQsvQo5WViQqWrr0+KlsaiTRO/TE0XCGFx4xQKEuhZ6QAnHiiVU en34SMqY51qa5D3LSbs6F278rEZNcLQguiH6Urxe5KRmkJDfoxtsWQ/DpV8itbnk raCUFlhW8GIBrRvizB7Na+hDWj1/HGQRIEs+xlfqPcFDV9bkECE/IpbD04+JDbil wsDoy2IO15oG/rX05/bkXAY7fFuhWbnVAbKrqvl+50w8Oo5w0+X3ZHlqhiLqCzVr e/TL5lc+9Ciq4uG8TCwal4HoktYLwqez4qxz396YpE4LN1ax2ICFgR9HyY4GLqmU 0H1qSxZmOkeueCU= =vLZn -----END PGP SIGNATURE----- Merge tag 'xfs-cil-scale-5.20' of git://git.kernel.org/pub/scm/linux/kernel/git/dgc/linux-xfs into xfs-5.20-mergeA xfs: improve CIL scalability This series aims to improve the scalability of XFS transaction commits on large CPU count machines. My 32p machine hits contention limits in xlog_cil_commit() at about 700,000 transaction commits a section. It hits this at 16 thread workloads, and 32 thread workloads go no faster and just burn CPU on the CIL spinlocks. This patchset gets rid of spinlocks and global serialisation points in the xlog_cil_commit() path. It does this by moving to a combination of per-cpu counters, unordered per-cpu lists and post-ordered per-cpu lists. This results in transaction commit rates exceeding 1.4 million commits/s under unlink certain workloads, and while the log lock contention is largely gone there is still significant lock contention in the VFS (dentry cache, inode cache and security layers) at >600,000 transactions/s that still limit scalability. The changes to the CIL accounting and behaviour, combined with the structural changes to xlog_write() in prior patchsets make the per-cpu restructuring possible and sane. This allows us to move to precalculated reservation requirements that allow for reservation stealing to be accounted across multiple CPUs accurately. That is, instead of trying to account for continuation log opheaders on a "growth" basis, we pre-calculate how many iclogs we'll need to write out a maximally sized CIL checkpoint and steal that reserveD that space one commit at a time until the CIL has a full reservation. If we ever run a commit when we are already at the hard limit (because post-throttling) we simply take an extra reservation from each commit that is run when over the limit. Hence we don't need to do space usage math in the fast path and so never need to sum the per-cpu counters in this fast path. Similarly, per-cpu lists have the problem of ordering - we can't remove an item from a per-cpu list if we want to move it forward in the CIL. We solve this problem by using an atomic counter to give every commit a sequence number that is copied into the log items in that transaction. Hence relogging items just overwrites the sequence number in the log item, and does not move it in the per-cpu lists. Once we reaggregate the per-cpu lists back into a single list in the CIL push work, we can run it through list-sort() and reorder it back into a globally ordered list. This costs a bit of CPU time, but now that the CIL can run multiple works and pipelines properly, this is not a limiting factor for performance. It does increase fsync latency when the CIL is full, but workloads issuing large numbers of fsync()s or sync transactions end up with very small CILs and so the latency impact or sorting is not measurable for such workloads. OVerall, this pushes the transaction commit bottleneck out to the lockless reservation grant head updates. These atomic updates don't start to be a limiting fact until > 1.5 million transactions/s are being run, at which point the accounting functions start to show up in profiles as the highest CPU users. Still, this series doubles transaction throughput without increasing CPU usage before we get to that cacheline contention breakdown point... ` Signed-off-by: Dave Chinner <dchinner@redhat.com> Signed-off-by: Darrick J. Wong <djwong@kernel.org> * tag 'xfs-cil-scale-5.20' of git://git.kernel.org/pub/scm/linux/kernel/git/dgc/linux-xfs: xfs: expanding delayed logging design with background material xfs: xlog_sync() manually adjusts grant head space xfs: avoid cil push lock if possible xfs: move CIL ordering to the logvec chain xfs: convert log vector chain to use list heads xfs: convert CIL to unordered per cpu lists xfs: Add order IDs to log items in CIL xfs: convert CIL busy extents to per-cpu xfs: track CIL ticket reservation in percpu structure xfs: implement percpu cil space used calculation xfs: introduce per-cpu CIL tracking structure xfs: rework per-iclog header CIL reservation xfs: lift init CIL reservation out of xc_cil_lock xfs: use the CIL space used counter for emptiness checks |
||
|---|---|---|
| .. | ||
| ABI | ||
| accounting | ||
| admin-guide | ||
| arc | ||
| arm | ||
| arm64 | ||
| block | ||
| bpf | ||
| cdrom | ||
| core-api | ||
| cpu-freq | ||
| crypto | ||
| dev-tools | ||
| devicetree | ||
| doc-guide | ||
| driver-api | ||
| fault-injection | ||
| fb | ||
| features | ||
| filesystems | ||
| firmware_class | ||
| firmware-guide | ||
| fpga | ||
| gpu | ||
| hid | ||
| hwmon | ||
| i2c | ||
| ia64 | ||
| iio | ||
| images | ||
| infiniband | ||
| input | ||
| isdn | ||
| kbuild | ||
| kernel-hacking | ||
| leds | ||
| litmus-tests | ||
| livepatch | ||
| locking | ||
| loongarch | ||
| m68k | ||
| maintainer | ||
| mhi | ||
| mips | ||
| misc-devices | ||
| netlabel | ||
| networking | ||
| nios2 | ||
| nvdimm | ||
| openrisc | ||
| parisc | ||
| PCI | ||
| pcmcia | ||
| peci | ||
| power | ||
| powerpc | ||
| process | ||
| RCU | ||
| riscv | ||
| s390 | ||
| scheduler | ||
| scsi | ||
| security | ||
| sh | ||
| sound | ||
| sparc | ||
| sphinx | ||
| sphinx-static | ||
| spi | ||
| staging | ||
| target | ||
| timers | ||
| tools | ||
| trace | ||
| translations | ||
| usb | ||
| userspace-api | ||
| virt | ||
| vm | ||
| w1 | ||
| watchdog | ||
| x86 | ||
| xtensa | ||
| .gitignore | ||
| arch.rst | ||
| asm-annotations.rst | ||
| atomic_bitops.txt | ||
| atomic_t.txt | ||
| Changes | ||
| CodingStyle | ||
| conf.py | ||
| docutils.conf | ||
| dontdiff | ||
| index.rst | ||
| Kconfig | ||
| Makefile | ||
| memory-barriers.txt | ||
| SubmittingPatches | ||