Pull scheduler fixes from Ingo Molnar:
"Fix a performance regression on large SMT systems, an Intel SMT4
balancing bug, and a topology setup bug on (Intel) hybrid processors"
* tag 'sched-urgent-2023-09-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/sched: Restore the SD_ASYM_PACKING flag in the DIE domain
sched/fair: Fix SMT4 group_smt_balance handling
sched/fair: Optimize should_we_balance() for large SMT systems
Pull WARN fix from Ingo Molnar:
"Fix a missing preempt-enable in the WARN() slowpath"
* tag 'core-urgent-2023-09-17' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
panic: Reenable preemption in WARN slowpath
Pull power management fixes from Rafael Wysocki:
"Fix the handling of block devices in the test_resume mode of
hibernation (Chen Yu)"
* tag 'pm-6.6-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
PM: hibernate: Fix the exclusive get block device in test_resume mode
PM: hibernate: Rename function parameter from snapshot_test to exclusive
Commit:
5a5d7e9bad ("cpuidle: lib/bug: Disable rcu_is_watching() during WARN/BUG")
amended warn_slowpath_fmt() to disable preemption until the WARN splat
has been emitted.
However the commit neglected to reenable preemption in the !fmt codepath,
i.e. when a WARN splat is emitted without additional format string.
One consequence is that users may see more splats than intended. E.g. a
WARN splat emitted in a work item results in at least two extra splats:
BUG: workqueue leaked lock or atomic
(emitted by process_one_work())
BUG: scheduling while atomic
(emitted by worker_thread() -> schedule())
Ironically the point of the commit was to *avoid* extra splats. ;)
Fix it.
Fixes: 5a5d7e9bad ("cpuidle: lib/bug: Disable rcu_is_watching() during WARN/BUG")
Signed-off-by: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Paul E. McKenney <paulmck@kernel.org>
Link: https://lore.kernel.org/r/3ec48fde01e4ee6505f77908ba351bad200ae3d1.1694763684.git.lukas@wunner.de
Pull tracing fixes from Steven Rostedt:
- Add missing LOCKDOWN checks for eventfs callers
When LOCKDOWN is active for tracing, it causes inconsistent state
when some functions succeed and others fail.
- Use dput() to free the top level eventfs descriptor
There was a race between accesses and freeing it.
- Fix a long standing bug that eventfs exposed due to changing timings
by dynamically creating files. That is, If a event file is opened for
an instance, there's nothing preventing the instance from being
removed which will make accessing the files cause use-after-free
bugs.
- Fix a ring buffer race that happens when iterating over the ring
buffer while writers are active. Check to make sure not to read the
event meta data if it's beyond the end of the ring buffer sub buffer.
- Fix the print trigger that disappeared because the test to create it
was looking for the event dir field being filled, but now it has the
"ef" field filled for the eventfs structure.
- Remove the unused "dir" field from the event structure.
- Fix the order of the trace_dynamic_info as it had it backwards for
the offset and len fields for which one was for which endianess.
- Fix NULL pointer dereference with eventfs_remove_rec()
If an allocation fails in one of the eventfs_add_*() functions, the
caller of it in event_subsystem_dir() or event_create_dir() assigns
the result to the structure. But it's assigning the ERR_PTR and not
NULL. This was passed to eventfs_remove_rec() which expects either a
good pointer or a NULL, not ERR_PTR. The fix is to not assign the
ERR_PTR to the structure, but to keep it NULL on error.
- Fix list_for_each_rcu() to use list_for_each_srcu() in
dcache_dir_open_wrapper(). One iteration of the code used RCU but
because it had to call sleepable code, it had to be changed to use
SRCU, but one of the iterations was missed.
- Fix synthetic event print function to use "as_u64" instead of passing
in a pointer to the union. To fix big/little endian issues, the u64
that represented several types was turned into a union to define the
types properly.
* tag 'trace-v6.6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
eventfs: Fix the NULL pointer dereference bug in eventfs_remove_rec()
tracefs/eventfs: Use list_for_each_srcu() in dcache_dir_open_wrapper()
tracing/synthetic: Print out u64 values properly
tracing/synthetic: Fix order of struct trace_dynamic_info
selftests/ftrace: Fix dependencies for some of the synthetic event tests
tracing: Remove unused trace_event_file dir field
tracing: Use the new eventfs descriptor for print trigger
ring-buffer: Do not attempt to read past "commit"
tracefs/eventfs: Free top level files on removal
ring-buffer: Avoid softlockup in ring_buffer_resize()
tracing: Have event inject files inc the trace array ref count
tracing: Have option files inc the trace array ref count
tracing: Have current_trace inc the trace array ref count
tracing: Have tracing_max_latency inc the trace array ref count
tracing: Increase trace array ref count on enable and filter files
tracefs/eventfs: Use dput to free the toplevel events directory
tracefs/eventfs: Add missing lockdown checks
tracefs: Add missing lockdown check to tracefs_create_dir()
For SMT4, any group with more than 2 tasks will be marked as
group_smt_balance. Retain the behaviour of group_has_spare by marking
the busiest group as the group which has the least number of idle_cpus.
Also, handle rounding effect of adding (ncores_local + ncores_busy) when
the local is fully idle and busy group imbalance is less than 2 tasks.
Local group should try to pull at least 1 task in this case so imbalance
should be set to 2 instead.
Fixes: fee1759e4f ("sched/fair: Determine active load balance for SMT sched groups")
Acked-by: Shrikanth Hegde <sshegde@linux.vnet.ibm.com>
Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Link: http://lkml.kernel.org/r/6cd1633036bb6b651af575c32c2a9608a106702c.camel@linux.intel.com
Commit 5904de0d73 ("PM: hibernate: Do not get block device exclusively
in test_resume mode") fixes a hibernation issue under test_resume mode.
That commit is supposed to open the block device in non-exclusive mode
when in test_resume. However the code does the opposite, which is against
its description.
In summary, the swap device is only opened exclusively by swsusp_check()
with its corresponding *close(), and must be in non test_resume mode.
This is to avoid the race condition that different processes scribble the
device at the same time. All the other cases should use non-exclusive mode.
Fix it by really disabling exclusive mode under test_resume.
Fixes: 5904de0d73 ("PM: hibernate: Do not get block device exclusively in test_resume mode")
Closes: https://lore.kernel.org/lkml/000000000000761f5f0603324129@google.com/
Reported-by: Pengfei Xu <pengfei.xu@intel.com>
Signed-off-by: Chen Yu <yu.c.chen@intel.com>
Tested-by: Chenzhou Feng <chenzhoux.feng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Several functions reply on snapshot_test to decide whether to
open the resume device exclusively. However there is no strict
connection between the snapshot_test and the open mode. Rename
the 'snapshot_test' input parameter to 'exclusive' to better reflect
the use case.
No functional change is expected.
Signed-off-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Pull more RISC-V updates from Palmer Dabbelt:
- The kernel now dynamically probes for misaligned access speed, as
opposed to relying on a table of known implementations.
- Support for non-coherent devices on systems using the Andes AX45MP
core, including the RZ/Five SoCs.
- Support for the V extension in ptrace(), again.
- Support for KASLR.
- Support for the BPF prog pack allocator in RISC-V.
- A handful of bug fixes and cleanups.
* tag 'riscv-for-linus-6.6-mw2-2' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux: (25 commits)
soc: renesas: Kconfig: For ARCH_R9A07G043 select the required configs if dependencies are met
riscv: Kconfig.errata: Add dependency for RISCV_SBI in ERRATA_ANDES config
riscv: Kconfig.errata: Drop dependency for MMU in ERRATA_ANDES_CMO config
riscv: Kconfig: Select DMA_DIRECT_REMAP only if MMU is enabled
bpf, riscv: use prog pack allocator in the BPF JIT
riscv: implement a memset like function for text
riscv: extend patch_text_nosync() for multiple pages
bpf: make bpf_prog_pack allocator portable
riscv: libstub: Implement KASLR by using generic functions
libstub: Fix compilation warning for rv32
arm64: libstub: Move KASLR handling functions to kaslr.c
riscv: Dump out kernel offset information on panic
riscv: Introduce virtual kernel mapping KASLR
RISC-V: Add ptrace support for vectors
soc: renesas: Kconfig: Select the required configs for RZ/Five SoC
cache: Add L2 cache management for Andes AX45MP RISC-V core
dt-bindings: cache: andestech,ax45mp-cache: Add DT binding documentation for L2 cache controller
riscv: mm: dma-noncoherent: nonstandard cache operations support
riscv: errata: Add Andes alternative ports
riscv: asm: vendorid_list: Add Andes Technology to the vendors list
...
Pull dma-mapping fixes from Christoph Hellwig:
- move a dma-debug call that prints a message out from a lock that's
causing problems with the lock order in serial drivers (Sergey
Senozhatsky)
- fix the CONFIG_DMA_NUMA_CMA Kconfig entry to have the right
dependency and not default to y (Christoph Hellwig)
- move an ifdef a bit to remove a __maybe_unused that seems to trip up
some sensitivities (Christoph Hellwig)
- revert a bogus check in the CMA allocator (Zhenhua Huang)
* tag 'dma-mapping-6.6-2023-09-09' of git://git.infradead.org/users/hch/dma-mapping:
Revert "dma-contiguous: check for memory region overlap"
dma-pool: remove a __maybe_unused label in atomic_pool_expand
dma-contiguous: fix the Kconfig entry for CONFIG_DMA_NUMA_CMA
dma-debug: don't call __dma_entry_alloc_check_leak() under free_entries_lock
The check to create the print event "trigger" was using the obsolete "dir"
value of the trace_event_file to determine if it should create the trigger
or not. But that value will now be NULL because it uses the event file
descriptor.
Change it to test the "ef" field of the trace_event_file structure so that
the trace_marker "trigger" file appears again.
Link: https://lkml.kernel.org/r/20230908022001.371815239@goodmis.org
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Ajay Kaher <akaher@vmware.com>
Fixes: 27152bceea ("eventfs: Move tracing/events to eventfs")
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
When iterating over the ring buffer while the ring buffer is active, the
writer can corrupt the reader. There's barriers to help detect this and
handle it, but that code missed the case where the last event was at the
very end of the page and has only 4 bytes left.
The checks to detect the corruption by the writer to reads needs to see the
length of the event. If the length in the first 4 bytes is zero then the
length is stored in the second 4 bytes. But if the writer is in the process
of updating that code, there's a small window where the length in the first
4 bytes could be zero even though the length is only 4 bytes. That will
cause rb_event_length() to read the next 4 bytes which could happen to be off the
allocated page.
To protect against this, fail immediately if the next event pointer is
less than 8 bytes from the end of the commit (last byte of data), as all
events must be a minimum of 8 bytes anyway.
Link: https://lore.kernel.org/all/20230905141245.26470-1-Tze-nan.Wu@mediatek.com/
Link: https://lore.kernel.org/linux-trace-kernel/20230907122820.0899019c@gandalf.local.home
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Reported-by: Tze-nan Wu <Tze-nan.Wu@mediatek.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Pull printk fix from Petr Mladek:
- Revert exporting symbols needed for dumping the raw printk buffer in
panic().
I pushed the export prematurely before the user was ready for merging
into the mainline.
* tag 'printk-for-6.6-fixup' of git://git.kernel.org/pub/scm/linux/kernel/git/printk/linux:
Revert "printk: export symbols for debug modules"
Puranjay Mohan <puranjay12@gmail.com> says:
Here is some data to prove the V2 fixes the problem:
Without this series:
root@rv-selftester:~/src/kselftest/bpf# time ./test_tag
test_tag: OK (40945 tests)
real 7m47.562s
user 0m24.145s
sys 6m37.064s
With this series applied:
root@rv-selftester:~/src/selftest/bpf# time ./test_tag
test_tag: OK (40945 tests)
real 7m29.472s
user 0m25.865s
sys 6m18.401s
BPF programs currently consume a page each on RISCV. For systems with many BPF
programs, this adds significant pressure to instruction TLB. High iTLB pressure
usually causes slow down for the whole system.
Song Liu introduced the BPF prog pack allocator[1] to mitigate the above issue.
It packs multiple BPF programs into a single huge page. It is currently only
enabled for the x86_64 BPF JIT.
I enabled this allocator on the ARM64 BPF JIT[2]. It is being reviewed now.
This patch series enables the BPF prog pack allocator for the RISCV BPF JIT.
======================================================
Performance Analysis of prog pack allocator on RISCV64
======================================================
Test setup:
===========
Host machine: Debian GNU/Linux 11 (bullseye)
Qemu Version: QEMU emulator version 8.0.3 (Debian 1:8.0.3+dfsg-1)
u-boot-qemu Version: 2023.07+dfsg-1
opensbi Version: 1.3-1
To test the performance of the BPF prog pack allocator on RV, a stresser
tool[4] linked below was built. This tool loads 8 BPF programs on the system and
triggers 5 of them in an infinite loop by doing system calls.
The runner script starts 20 instances of the above which loads 8*20=160 BPF
programs on the system, 5*20=100 of which are being constantly triggered.
The script is passed a command which would be run in the above environment.
The script was run with following perf command:
./run.sh "perf stat -a \
-e iTLB-load-misses \
-e dTLB-load-misses \
-e dTLB-store-misses \
-e instructions \
--timeout 60000"
The output of the above command is discussed below before and after enabling the
BPF prog pack allocator.
The tests were run on qemu-system-riscv64 with 8 cpus, 16G memory. The rootfs
was created using Bjorn's riscv-cross-builder[5] docker container linked below.
Results
=======
Before enabling prog pack allocator:
------------------------------------
Performance counter stats for 'system wide':
4939048 iTLB-load-misses
5468689 dTLB-load-misses
465234 dTLB-store-misses
1441082097998 instructions
60.045791200 seconds time elapsed
After enabling prog pack allocator:
-----------------------------------
Performance counter stats for 'system wide':
3430035 iTLB-load-misses
5008745 dTLB-load-misses
409944 dTLB-store-misses
1441535637988 instructions
60.046296600 seconds time elapsed
Improvements in metrics
=======================
It was expected that the iTLB-load-misses would decrease as now a single huge
page is used to keep all the BPF programs compared to a single page for each
program earlier.
--------------------------------------------
The improvement in iTLB-load-misses: -30.5 %
--------------------------------------------
I repeated this expriment more than 100 times in different setups and the
improvement was always greater than 30%.
This patch series is boot tested on the Starfive VisionFive 2 board[6].
The performance analysis was not done on the board because it doesn't
expose iTLB-load-misses, etc. The stresser program was run on the board to test
the loading and unloading of BPF programs
[1] https://lore.kernel.org/bpf/20220204185742.271030-1-song@kernel.org/
[2] https://lore.kernel.org/all/20230626085811.3192402-1-puranjay12@gmail.com/
[3] https://lore.kernel.org/all/20230626085811.3192402-2-puranjay12@gmail.com/
[4] https://github.com/puranjaymohan/BPF-Allocator-Bench
[5] https://github.com/bjoto/riscv-cross-builder
[6] https://www.starfivetech.com/en/site/boards
* b4-shazam-merge:
bpf, riscv: use prog pack allocator in the BPF JIT
riscv: implement a memset like function for text
riscv: extend patch_text_nosync() for multiple pages
bpf: make bpf_prog_pack allocator portable
Link: https://lore.kernel.org/r/20230831131229.497941-1-puranjay12@gmail.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
This reverts commit 3fa6456ebe.
The Commit broke the CMA region creation through DT on arm64,
as showed below logs with "memblock=debug":
[ 0.000000] memblock_phys_alloc_range: 41943040 bytes align=0x200000
from=0x0000000000000000 max_addr=0x00000000ffffffff
early_init_dt_alloc_reserved_memory_arch+0x34/0xa0
[ 0.000000] memblock_reserve: [0x00000000fd600000-0x00000000ffdfffff]
memblock_alloc_range_nid+0xc0/0x19c
[ 0.000000] Reserved memory: overlap with other memblock reserved region
>From call flow, region we defined in DT was always reserved before entering
into rmem_cma_setup. Also, rmem_cma_setup has one routine cma_init_reserved_mem
to ensure the region was reserved. Checking the region not reserved here seems
not correct.
early_init_fdt_scan_reserved_mem:
fdt_scan_reserved_mem
__reserved_mem_reserve_reg
early_init_dt_reserve_memory
memblock_reserve(using “reg” prop case)
fdt_init_reserved_mem
__reserved_mem_alloc_size
*early_init_dt_alloc_reserved_memory_arch*
memblock_reserve(dynamic alloc case)
__reserved_mem_init_node
rmem_cma_setup(region overlap check here should always fail)
Example DT can be used to reproduce issue:
dump_mem: mem_dump_region {
compatible = "shared-dma-pool";
alloc-ranges = <0x0 0x00000000 0x0 0xffffffff>;
reusable;
size = <0 0x2800000>;
};
Signed-off-by: Zhenhua Huang <quic_zhenhuah@quicinc.com>
Pull networking updates from Jakub Kicinski:
"Including fixes from netfilter and bpf.
Current release - regressions:
- eth: stmmac: fix failure to probe without MAC interface specified
Current release - new code bugs:
- docs: netlink: fix missing classic_netlink doc reference
Previous releases - regressions:
- deal with integer overflows in kmalloc_reserve()
- use sk_forward_alloc_get() in sk_get_meminfo()
- bpf_sk_storage: fix the missing uncharge in sk_omem_alloc
- fib: avoid warn splat in flow dissector after packet mangling
- skb_segment: call zero copy functions before using skbuff frags
- eth: sfc: check for zero length in EF10 RX prefix
Previous releases - always broken:
- af_unix: fix msg_controllen test in scm_pidfd_recv() for
MSG_CMSG_COMPAT
- xsk: fix xsk_build_skb() dereferencing possible ERR_PTR()
- netfilter:
- nft_exthdr: fix non-linear header modification
- xt_u32, xt_sctp: validate user space input
- nftables: exthdr: fix 4-byte stack OOB write
- nfnetlink_osf: avoid OOB read
- one more fix for the garbage collection work from last release
- igmp: limit igmpv3_newpack() packet size to IP_MAX_MTU
- bpf, sockmap: fix preempt_rt splat when using raw_spin_lock_t
- handshake: fix null-deref in handshake_nl_done_doit()
- ip: ignore dst hint for multipath routes to ensure packets are
hashed across the nexthops
- phy: micrel:
- correct bit assignments for cable test errata
- disable EEE according to the KSZ9477 errata
Misc:
- docs/bpf: document compile-once-run-everywhere (CO-RE) relocations
- Revert "net: macsec: preserve ingress frame ordering", it appears
to have been developed against an older kernel, problem doesn't
exist upstream"
* tag 'net-6.6-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (95 commits)
net: enetc: distinguish error from valid pointers in enetc_fixup_clear_rss_rfs()
Revert "net: team: do not use dynamic lockdep key"
net: hns3: remove GSO partial feature bit
net: hns3: fix the port information display when sfp is absent
net: hns3: fix invalid mutex between tc qdisc and dcb ets command issue
net: hns3: fix debugfs concurrency issue between kfree buffer and read
net: hns3: fix byte order conversion issue in hclge_dbg_fd_tcam_read()
net: hns3: Support query tx timeout threshold by debugfs
net: hns3: fix tx timeout issue
net: phy: Provide Module 4 KSZ9477 errata (DS80000754C)
netfilter: nf_tables: Unbreak audit log reset
netfilter: ipset: add the missing IP_SET_HASH_WITH_NET0 macro for ip_set_hash_netportnet.c
netfilter: nft_set_rbtree: skip sync GC for new elements in this transaction
netfilter: nf_tables: uapi: Describe NFTA_RULE_CHAIN_ID
netfilter: nfnetlink_osf: avoid OOB read
netfilter: nftables: exthdr: fix 4-byte stack OOB write
selftests/bpf: Check bpf_sk_storage has uncharged sk_omem_alloc
bpf: bpf_sk_storage: Fix the missing uncharge in sk_omem_alloc
bpf: bpf_sk_storage: Fix invalid wait context lockdep report
s390/bpf: Pass through tail call counter in trampolines
...
When user resize all trace ring buffer through file 'buffer_size_kb',
then in ring_buffer_resize(), kernel allocates buffer pages for each
cpu in a loop.
If the kernel preemption model is PREEMPT_NONE and there are many cpus
and there are many buffer pages to be allocated, it may not give up cpu
for a long time and finally cause a softlockup.
To avoid it, call cond_resched() after each cpu buffer allocation.
Link: https://lore.kernel.org/linux-trace-kernel/20230906081930.3939106-1-zhengyejian1@huawei.com
Cc: <mhiramat@kernel.org>
Signed-off-by: Zheng Yejian <zhengyejian1@huawei.com>
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
