When switching processes, currently all user SLBEs are cleared, and a
few (exec_base, pc, and stack) are preloaded. In trivial testing with
small apps, this tends to miss the heap and low 256MB segments, and it
will also miss commonly accessed segments on large memory workloads.
Add a simple round-robin preload cache that just inserts the last SLB
miss into the head of the cache and preloads those at context switch
time. Every 256 context switches, the oldest entry is removed from the
cache to shrink the cache and require fewer slbmte if they are unused.
Much more could go into this, including into the SLB entry reclaim
side to track some LRU information etc, which would require a study of
large memory workloads. But this is a simple thing we can do now that
is an obvious win for common workloads.
With the full series, process switching speed on the context_switch
benchmark on POWER9/hash (with kernel speculation security masures
disabled) increases from 140K/s to 178K/s (27%).
POWER8 does not change much (within 1%), it's unclear why it does not
see a big gain like POWER9.
Booting to busybox init with 256MB segments has SLB misses go down
from 945 to 69, and with 1T segments 900 to 21. These could almost all
be eliminated by preloading a bit more carefully with ELF binary
loading.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
PPR is the odd register out when it comes to interrupt handling, it is
saved in current->thread.ppr while all others are saved on the stack.
The difficulty with this is that accessing thread.ppr can cause a SLB
fault, but the SLB fault handler implementation in C change had
assumed the normal exception entry handlers would not cause an SLB
fault.
Fix this by allocating room in the interrupt stack to save PPR.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This reverts commits:
5e46e29e6a ("powerpc/64s/hash: convert SLB miss handlers to C")
8fed04d0f6 ("powerpc/64s/hash: remove user SLB data from the paca")
655deecf67 ("powerpc/64s/hash: SLB allocation status bitmaps")
2e1626744e ("powerpc/64s/hash: provide arch_setup_exec hooks for hash slice setup")
89ca4e126a ("powerpc/64s/hash: Add a SLB preload cache")
This series had a few bugs, and the fixes are not all trivial. So
revert most of it for now.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
When switching processes, currently all user SLBEs are cleared, and a
few (exec_base, pc, and stack) are preloaded. In trivial testing with
small apps, this tends to miss the heap and low 256MB segments, and it
will also miss commonly accessed segments on large memory workloads.
Add a simple round-robin preload cache that just inserts the last SLB
miss into the head of the cache and preloads those at context switch
time. Every 256 context switches, the oldest entry is removed from the
cache to shrink the cache and require fewer slbmte if they are unused.
Much more could go into this, including into the SLB entry reclaim
side to track some LRU information etc, which would require a study of
large memory workloads. But this is a simple thing we can do now that
is an obvious win for common workloads.
With the full series, process switching speed on the context_switch
benchmark on POWER9/hash (with kernel speculation security masures
disabled) increases from 140K/s to 178K/s (27%).
POWER8 does not change much (within 1%), it's unclear why it does not
see a big gain like POWER9.
Booting to busybox init with 256MB segments has SLB misses go down
from 945 to 69, and with 1T segments 900 to 21. These could almost all
be eliminated by preloading a bit more carefully with ELF binary
loading.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
It's called 'fs' for historical reasons, it's named after the x86 'FS'
register. But we don't have to use that name for the member of
thread_struct, and in fact arch/x86 doesn't even call it 'fs' anymore.
So rename it to 'addr_limit', which better reflects what it's used
for, and is also the name used on other arches.
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Implement a new function to invoke stop, power9_offline_stop, which is
like power9_idle_stop but used by the cpu hotplug code.
Move KVM secondary state manipulation code to the offline case.
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Reviewed-by: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
This patch increases the max virtual (effective) address value to 4PB.
With 4K page size config we continue to limit ourself to 64TB.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
[mpe: Keep the H_PGTABLE_RANGE test, update it to work]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
For addresses above 512TB we allocate additional mmu contexts. To make
it all easy, addresses above 512TB are handled with IR/DR=1 and with
stack frame setup.
The mmu_context_t is also updated to track the new extended_ids. To
support upto 4PB we need a total 8 contexts.
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
[mpe: Minor formatting tweaks and comment wording, switch BUG to WARN
in get_ea_context().]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Store and restore the AMR, IAMR and UAMOR register state of the task
before scheduling out and after scheduling in, respectively.
Signed-off-by: Ram Pai <linuxram@us.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
A CP_ABORT instruction is required in processes that have mapped a VAS
"paste address" with the intention of using COPY/PASTE instructions.
But since CP_ABORT is expensive, we want to restrict it to only
processes that use/intend to use COPY/PASTE.
Define an interface, set_thread_uses_vas(), that VAS can use to
indicate that the current process opened a send window. During context
switch, issue CP_ABORT only for processes that have the flag set.
Thanks for input from Nick Piggin, Michael Ellerman.
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
[mpe: Fix to not use new_thread after _switch() returns]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
We need the SPRN_TIDR to be set for use with fast thread-wakeup (core-
to-core wakeup) and also with CAPI.
Each thread in a process needs to have a unique id within the process.
But for now, we assign globally unique thread ids to all threads in
the system.
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.vnet.ibm.com>
Signed-off-by: Philippe Bergheaud <felix@linux.vnet.ibm.com>
Signed-off-by: Christophe Lombard <clombard@linux.vnet.ibm.com>
[mpe: Simplify tidr clearing on fork() and ctx switch code]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Pull powerpc updates from Michael Ellerman:
"Highlights include:
- Support for STRICT_KERNEL_RWX on 64-bit server CPUs.
- Platform support for FSP2 (476fpe) board
- Enable ZONE_DEVICE on 64-bit server CPUs.
- Generic & powerpc spin loop primitives to optimise busy waiting
- Convert VDSO update function to use new update_vsyscall() interface
- Optimisations to hypercall/syscall/context-switch paths
- Improvements to the CPU idle code on Power8 and Power9.
As well as many other fixes and improvements.
Thanks to: Akshay Adiga, Andrew Donnellan, Andrew Jeffery, Anshuman
Khandual, Anton Blanchard, Balbir Singh, Benjamin Herrenschmidt,
Christophe Leroy, Christophe Lombard, Colin Ian King, Dan Carpenter,
Gautham R. Shenoy, Hari Bathini, Ian Munsie, Ivan Mikhaylov, Javier
Martinez Canillas, Madhavan Srinivasan, Masahiro Yamada, Matt Brown,
Michael Neuling, Michal Suchanek, Murilo Opsfelder Araujo, Naveen N.
Rao, Nicholas Piggin, Oliver O'Halloran, Paul Mackerras, Pavel Machek,
Russell Currey, Santosh Sivaraj, Stephen Rothwell, Thiago Jung
Bauermann, Yang Li"
* tag 'powerpc-4.13-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (158 commits)
powerpc/Kconfig: Enable STRICT_KERNEL_RWX for some configs
powerpc/mm/radix: Implement STRICT_RWX/mark_rodata_ro() for Radix
powerpc/mm/hash: Implement mark_rodata_ro() for hash
powerpc/vmlinux.lds: Align __init_begin to 16M
powerpc/lib/code-patching: Use alternate map for patch_instruction()
powerpc/xmon: Add patch_instruction() support for xmon
powerpc/kprobes/optprobes: Use patch_instruction()
powerpc/kprobes: Move kprobes over to patch_instruction()
powerpc/mm/radix: Fix execute permissions for interrupt_vectors
powerpc/pseries: Fix passing of pp0 in updatepp() and updateboltedpp()
powerpc/64s: Blacklist rtas entry/exit from kprobes
powerpc/64s: Blacklist functions invoked on a trap
powerpc/64s: Un-blacklist system_call() from kprobes
powerpc/64s: Move system_call() symbol to just after setting MSR_EE
powerpc/64s: Blacklist system_call() and system_call_common() from kprobes
powerpc/64s: Convert .L__replay_interrupt_return to a local label
powerpc64/elfv1: Only dereference function descriptor for non-text symbols
cxl: Export library to support IBM XSL
powerpc/dts: Use #include "..." to include local DT
powerpc/perf/hv-24x7: Aggregate result elements on POWER9 SMT8
...
Merge our fixes branch, a few of them are tripping people up while
working on top of next, and we also have a dependency between the CXL
fixes and new CXL code we want to merge into next.
The only user of thread_saved_pc() in non-arch-specific code was removed
in commit 8243d55977 ("sched/core: Remove pointless printout in
sched_show_task()"). Remove the implementations as well.
Some architectures use thread_saved_pc() in their arch-specific code.
Leave their thread_saved_pc() intact.
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Acked-by: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This simplifies the asm and fixes irq-off tracing over sleep
instructions.
Also move powersave_nap check for POWER8 into C code, and move
PSSCR register value calculation for POWER9 into C.
Reviewed-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Supporting 512TB requires us to do a order 3 allocation for level 1 page
table (pgd). This results in page allocation failures with certain workloads.
For now limit 4k linux page size config to 64TB.
Fixes: f6eedbba7a ("powerpc/mm/hash: Increase VA range to 128TB")
Reported-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Commit f4ea6dcb08 ("powerpc/mm: Enable mappings above 128TB") increased
the task size on book3s, and introduced a mechanism to dynamically
control whether a task uses these larger addresses. While the change to
the task size itself was ifdef-protected to only apply on book3s, the
change to STACK_TOP_USER64 was not. On book3e, this had the effect of
trying to use addresses up to 128TiB for the stack despite a 64TiB task
size limit -- which broke 64-bit userspace producing the following errors:
Starting init: /sbin/init exists but couldn't execute it (error -14)
Starting init: /bin/sh exists but couldn't execute it (error -14)
Kernel panic - not syncing: No working init found. Try passing init= option to kernel. See Linux Documentation/admin-guide/init.rst for guidance.
Fixes: f4ea6dcb08 ("powerpc/mm: Enable mappings above 128TB")
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Scott Wood <oss@buserror.net>
Not all user space application is ready to handle wide addresses. It's
known that at least some JIT compilers use higher bits in pointers to
encode their information. It collides with valid pointers with 512TB
addresses and leads to crashes.
To mitigate this, we are not going to allocate virtual address space
above 128TB by default.
But userspace can ask for allocation from full address space by
specifying hint address (with or without MAP_FIXED) above 128TB.
If hint address set above 128TB, but MAP_FIXED is not specified, we try
to look for unmapped area by specified address. If it's already
occupied, we look for unmapped area in *full* address space, rather than
from 128TB window.
This approach helps to easily make application's memory allocator aware
about large address space without manually tracking allocated virtual
address space.
This is going to be a per mmap decision. ie, we can have some mmaps with
larger addresses and other that do not.
A sample memory layout looks like:
10000000-10010000 r-xp 00000000 fc:00 9057045 /home/max_addr_512TB
10010000-10020000 r--p 00000000 fc:00 9057045 /home/max_addr_512TB
10020000-10030000 rw-p 00010000 fc:00 9057045 /home/max_addr_512TB
10029630000-10029660000 rw-p 00000000 00:00 0 [heap]
7fff834a0000-7fff834b0000 rw-p 00000000 00:00 0
7fff834b0000-7fff83670000 r-xp 00000000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so
7fff83670000-7fff83680000 r--p 001b0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so
7fff83680000-7fff83690000 rw-p 001c0000 fc:00 9177190 /lib/powerpc64le-linux-gnu/libc-2.23.so
7fff83690000-7fff836a0000 rw-p 00000000 00:00 0
7fff836a0000-7fff836c0000 r-xp 00000000 00:00 0 [vdso]
7fff836c0000-7fff83700000 r-xp 00000000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so
7fff83700000-7fff83710000 r--p 00030000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so
7fff83710000-7fff83720000 rw-p 00040000 fc:00 9177193 /lib/powerpc64le-linux-gnu/ld-2.23.so
7fffdccf0000-7fffdcd20000 rw-p 00000000 00:00 0 [stack]
1000000000000-1000000010000 rw-p 00000000 00:00 0
1ffff83710000-1ffff83720000 rw-p 00000000 00:00 0
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
We update the hash linux page table layout such that we can support
512TB. But we limit the TASK_SIZE to 128TB. We can switch to 128TB by
default without conditional because that is the max virtual address
supported by other architectures. We will later add a mechanism to
on-demand increase the application's effective address range to 512TB.
Having the page table layout changed to accommodate 512TB makes testing
large memory configuration easier with less code changes to kernel
Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
The power9_idle_stop method currently takes only the requested stop
level as a parameter and picks up the rest of the PSSCR bits from a
hand-coded macro. This is not a very flexible design, especially when
the firmware has the capability to communicate the psscr value and the
mask associated with a particular stop state via device tree.
This patch modifies the power9_idle_stop API to take as parameters the
PSSCR value and the PSSCR mask corresponding to the stop state that
needs to be set. These PSSCR value and mask are respectively obtained
by parsing the "ibm,cpu-idle-state-psscr" and
"ibm,cpu-idle-state-psscr-mask" fields from the device tree.
In addition to this, the patch adds support for handling stop states
for which ESL and EC bits in the PSSCR are zero. As per the
architecture, a wakeup from these stop states resumes execution from
the subsequent instruction as opposed to waking up at the System
Vector.
The older firmware sets only the Requested Level (RL) field in the
psscr and psscr-mask exposed in the device tree. For older firmware
where psscr-mask=0xf, this patch will set the default sane values that
the set for for remaining PSSCR fields (i.e PSLL, MTL, ESL, EC, and
TR). For the new firmware, the patch will validate that the invariants
required by the ISA for the psscr values are maintained by the
firmware.
This skiboot patch that exports fully populated PSSCR values and the
mask for all the stop states can be found here:
https://lists.ozlabs.org/pipermail/skiboot/2016-September/004869.html
[Optimize the number of instructions before entering STOP with
ESL=EC=0, validate the PSSCR values provided by the firimware
maintains the invariants required as per the ISA suggested by Balbir
Singh]
Acked-by: Balbir Singh <bsingharora@gmail.com>
Signed-off-by: Gautham R. Shenoy <ego@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Nick Desaulniers
Nicholas Piggin
Michael Ellerman
Christophe Leroy
Aneesh Kumar K.V
Ram Pai
Sukadev Bhattiprolu
Linus Torvalds
Tobias Klauser
Scott Wood
Gautham R. Shenoy