KVM supports msgsndp in guests by trapping and emulating the
instruction, so it was decided to always use XIVE for IPIs if it is
available. However on PowerVM systems, msgsndp can be used and gives
better performance. On large systems, high XIVE interrupt rates can
have sub-linear scaling, and using msgsndp can reduce the load on
the interrupt controller.
So switch to using core local doorbells even if XIVE is available.
This reduces performance for KVM guests with an SMT topology by
about 50% for ping-pong context switching between SMT vCPUs. An
option vector (or dt-cpu-ftrs) could be defined to disable msgsndp
to get KVM performance back.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Tested-by: Cédric Le Goater <clg@kaod.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200726035155.1424103-3-npiggin@gmail.com
Commit 9908c826d5 ("powerpc/perf: Add Power10 PMU feature to DT CPU
features") defines MMCRA_BHRB_DISABLE as `0x2000000000UL`. Binutils
version less than 2.28 doesn't support UL suffix.
arch/powerpc/kernel/cpu_setup_power.S: Assembler messages:
arch/powerpc/kernel/cpu_setup_power.S:250: Error: found 'L', expected: ')'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: junk at end of line, first unrecognized character is `L'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: found 'L', expected: ')'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: found 'L', expected: ')'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: junk at end of line, first unrecognized character is `L'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: found 'L', expected: ')'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: found 'L', expected: ')'
arch/powerpc/kernel/cpu_setup_power.S:250: Error: operand out of range (0x0000002000000000 is not between 0xffffffffffff8000 and 0x000000000000ffff)
Fix this by wrapping it with the `_UL` macro.
Fixes: 9908c826d5 ("Add Power10 PMU feature to DT CPU features")
Suggested-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Athira Rajeev <atrajeev@linux.vnet.ibm.com>
Reviewed-by: Madhavan Srinivasan <maddy@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/1595996214-5833-1-git-send-email-atrajeev@linux.vnet.ibm.com
skiroot_defconfig fails:
arch/powerpc/kernel/fadump.c:48:17: error: ‘cpus_in_fadump’ defined but not used
48 | static atomic_t cpus_in_fadump;
Fix it by moving the definition into the #ifdef where it's used.
Fixes: ba608c4fa1 ("powerpc/fadump: fix race between pstore write and fadump crash trigger")
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200727070341.595634-1-mpe@ellerman.id.au
The sparse tool complains as follows:
arch/powerpc/platforms/pseries/papr_scm.c:97:1: warning:
symbol 'papr_nd_regions' was not declared. Should it be static?
arch/powerpc/platforms/pseries/papr_scm.c:98:1: warning:
symbol 'papr_ndr_lock' was not declared. Should it be static?
Those variables are not used outside of papr_scm.c, so this
commit marks them static.
Fixes: 85343a8da2 ("powerpc/papr/scm: Add bad memory ranges to nvdimm bad ranges")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wei Yongjun <weiyongjun1@huawei.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200725091949.75234-1-weiyongjun1@huawei.com
Clang's objdump emits slightly different output from GNU's objdump,
causing a list of warnings to be emitted during relocatable builds.
E.g., clang's objdump emits this:
c000000000000004: 2c 00 00 48 b 0xc000000000000030
...
c000000000005c6c: 10 00 82 40 bf 2, 0xc000000000005c7c
while GNU objdump emits:
c000000000000004: 2c 00 00 48 b c000000000000030 <__start+0x30>
...
c000000000005c6c: 10 00 82 40 bne c000000000005c7c <masked_interrupt+0x3c>
Adjust llvm-objdump's output to remove the extraneous '0x' and convert
'bf' and 'bt' to 'bne' and 'beq' resp. to more closely match GNU
objdump's output.
Note that clang's objdump doesn't yet output the relocation symbols on
PPC.
Signed-off-by: Bill Wendling <morbo@google.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/191c67db31264b69cf6b566fd69851beb3dd0abb.1595630874.git.morbo@google.com
This implements the generic paravirt qspinlocks using H_PROD and
H_CONFER to kick and wait.
This uses an un-directed yield to any CPU rather than the directed
yield to a pre-empted lock holder that paravirtualised simple
spinlocks use, that requires no kick hcall. This is something that
could be investigated and improved in future.
Performance results can be found in the commit which added queued
spinlocks.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Waiman Long <longman@redhat.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200724131423.1362108-5-npiggin@gmail.com
These have shown significantly improved performance and fairness when
spinlock contention is moderate to high on very large systems.
With this series including subsequent patches, on a 16 socket 1536
thread POWER9, a stress test such as same-file open/close from all
CPUs gets big speedups, 11620op/s aggregate with simple spinlocks vs
384158op/s (33x faster), where the difference in throughput between
the fastest and slowest thread goes from 7x to 1.4x.
Thanks to the fast path being identical in terms of atomics and
barriers (after a subsequent optimisation patch), single threaded
performance is not changed (no measurable difference).
On smaller systems, performance and fairness seems to be generally
improved. Using dbench on tmpfs as a test (that starts to run into
kernel spinlock contention), a 2-socket OpenPOWER POWER9 system was
tested with bare metal and KVM guest configurations. Results can be
found here:
https://github.com/linuxppc/issues/issues/305#issuecomment-663487453
Observations are:
- Queued spinlocks are equal when contention is insignificant, as
expected and as measured with microbenchmarks.
- When there is contention, on bare metal queued spinlocks have better
throughput and max latency at all points.
- When virtualised, queued spinlocks are slightly worse approaching
peak throughput, but significantly better throughput and max latency
at all points beyond peak, until queued spinlock maximum latency
rises when clients are 2x vCPUs.
The regressions haven't been analysed very well yet, there are a lot
of things that can be tuned, particularly the paravirtualised locking,
but the numbers already look like a good net win even on relatively
small systems.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Waiman Long <longman@redhat.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200724131423.1362108-4-npiggin@gmail.com
MAX_NUMNODES is a theoretical maximum number of nodes thats is
supported by the kernel. Device tree properties exposes the number of
possible nodes on the current platform. The kernel would detected this
and would use it for most of its resource allocations. If the platform
now increases the nodes to over what was already exposed, then it may
lead to inconsistencies. Hence limit it to the already exposed nodes.
Suggested-by: Nathan Lynch <nathanl@linux.ibm.com>
Signed-off-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200724105809.24733-1-srikar@linux.vnet.ibm.com
Userspace applications may use /dev/adb to send Talk requests. Such
requests always have req->reply_expected == 1. The same is true of Talk
requests sent by the kernel, except for poll requests queued internally
by the via-macii driver. Those requests have req->reply_expected == 0.
Consequently, poll reply packets get treated like autopoll reply packets.
(It doesn't make sense to try to distinguish them.) Always enter 'reading'
state after a poll request, so that the reply gets collected and passed
to adb_input(), and none go missing.
All Talk replies passed to adb_input() come from polling or autopolling,
so call adb_input() with the autopoll parameter set to 1.
Fixes: d95fd5fce8 ("m68k: Mac II ADB fixes") # v5.0+
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/754cddfa045e5bfa53e5da199831de02e7d2f27f.1593318192.git.fthain@telegraphics.com.au
I'm told that the /CTLR_IRQ signal from the ADB transceiver gets
interpreted by MacOS to mean SRQ, bus timeout or end-of-packet depending
on the circumstances, and that Linux's via-macii driver does not
correctly interpret this signal.
Instead, the via-macii driver interprets certain received byte values
(0x00 and 0xFF) as signalling end of packet and bus timeout
(respectively). Problem is, those values can also appear under other
circumstances.
This patch changes the bus timeout, end of packet and SRQ detection logic
to bring it closer to the logic that MacOS reportedly uses.
Fixes: 1da177e4c3 ("Linux-2.6.12-rc2") # v5.0+
Reported-by: Mark Cave-Ayland <mark.cave-ayland@ilande.co.uk>
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/6541fda1d8db3ae87c3abe17d189a10dc96e2382.1593318192.git.fthain@telegraphics.com.au
drivers/macintosh/adb-iop.c:215:28: warning: Using plain integer as NULL pointer
drivers/macintosh/adb-iop.c:170:5: warning: symbol 'adb_iop_probe' was not declared. Should it be static?
drivers/macintosh/adb-iop.c:177:5: warning: symbol 'adb_iop_init' was not declared. Should it be static?
drivers/macintosh/adb-iop.c:184:5: warning: symbol 'adb_iop_send_request' was not declared. Should it be static?
drivers/macintosh/adb-iop.c:230:5: warning: symbol 'adb_iop_autopoll' was not declared. Should it be static?
drivers/macintosh/adb-iop.c:236:6: warning: symbol 'adb_iop_poll' was not declared. Should it be static?
drivers/macintosh/adb-iop.c:241:5: warning: symbol 'adb_iop_reset_bus' was not declared. Should it be static?
Signed-off-by: Finn Thain <fthain@telegraphics.com.au>
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/25edf4450abd20e002b166ba3a11189dc1efa906.1590880623.git.fthain@telegraphics.com.au
Previously iov->vfs_expanded was used for two purposes.
1) To work out how much we need to multiple the per-VF BAR size to figure
out the total space required for the IOV BAR.
2) To indicate that IOV is not usable with this device (vfs_expanded == 0).
We don't really need the field for either since the multiple in 1) is
always the number PEs supported by the PHB. Similarly, we don't really need
it in 2) either since the IOV data field will be NULL if we can't use IOV
with the device.
Signed-off-by: Oliver O'Halloran <oohall@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200722065715.1432738-16-oohall@gmail.com
