If an interrupt occurs in queued_spin_lock_slowpath() after we increment
qnodesp->count and before node->lock is initialized, another CPU might
see stale lock values in get_tail_qnode(). If the stale lock value happens
to match the lock on that CPU, then we write to the "next" pointer of
the wrong qnode. This causes a deadlock as the former CPU, once it becomes
the head of the MCS queue, will spin indefinitely until it's "next" pointer
is set by its successor in the queue.
Running stress-ng on a 16 core (16EC/16VP) shared LPAR, results in
occasional lockups similar to the following:
$ stress-ng --all 128 --vm-bytes 80% --aggressive \
--maximize --oomable --verify --syslog \
--metrics --times --timeout 5m
watchdog: CPU 15 Hard LOCKUP
......
NIP [c0000000000b78f4] queued_spin_lock_slowpath+0x1184/0x1490
LR [c000000001037c5c] _raw_spin_lock+0x6c/0x90
Call Trace:
0xc000002cfffa3bf0 (unreliable)
_raw_spin_lock+0x6c/0x90
raw_spin_rq_lock_nested.part.135+0x4c/0xd0
sched_ttwu_pending+0x60/0x1f0
__flush_smp_call_function_queue+0x1dc/0x670
smp_ipi_demux_relaxed+0xa4/0x100
xive_muxed_ipi_action+0x20/0x40
__handle_irq_event_percpu+0x80/0x240
handle_irq_event_percpu+0x2c/0x80
handle_percpu_irq+0x84/0xd0
generic_handle_irq+0x54/0x80
__do_irq+0xac/0x210
__do_IRQ+0x74/0xd0
0x0
do_IRQ+0x8c/0x170
hardware_interrupt_common_virt+0x29c/0x2a0
--- interrupt: 500 at queued_spin_lock_slowpath+0x4b8/0x1490
......
NIP [c0000000000b6c28] queued_spin_lock_slowpath+0x4b8/0x1490
LR [c000000001037c5c] _raw_spin_lock+0x6c/0x90
--- interrupt: 500
0xc0000029c1a41d00 (unreliable)
_raw_spin_lock+0x6c/0x90
futex_wake+0x100/0x260
do_futex+0x21c/0x2a0
sys_futex+0x98/0x270
system_call_exception+0x14c/0x2f0
system_call_vectored_common+0x15c/0x2ec
The following code flow illustrates how the deadlock occurs.
For the sake of brevity, assume that both locks (A and B) are
contended and we call the queued_spin_lock_slowpath() function.
CPU0 CPU1
---- ----
spin_lock_irqsave(A) |
spin_unlock_irqrestore(A) |
spin_lock(B) |
| |
▼ |
id = qnodesp->count++; |
(Note that nodes[0].lock == A) |
| |
▼ |
Interrupt |
(happens before "nodes[0].lock = B") |
| |
▼ |
spin_lock_irqsave(A) |
| |
▼ |
id = qnodesp->count++ |
nodes[1].lock = A |
| |
▼ |
Tail of MCS queue |
| spin_lock_irqsave(A)
▼ |
Head of MCS queue ▼
| CPU0 is previous tail
▼ |
Spin indefinitely ▼
(until "nodes[1].next != NULL") prev = get_tail_qnode(A, CPU0)
|
▼
prev == &qnodes[CPU0].nodes[0]
(as qnodes[CPU0].nodes[0].lock == A)
|
▼
WRITE_ONCE(prev->next, node)
|
▼
Spin indefinitely
(until nodes[0].locked == 1)
Thanks to Saket Kumar Bhaskar for help with recreating the issue
Fixes: 84990b169557 ("powerpc/qspinlock: add mcs queueing for contended waiters")
Cc: stable@vger.kernel.org # v6.2+
Reported-by: Geetika Moolchandani <geetika@linux.ibm.com>
Reported-by: Vaishnavi Bhat <vaish123@in.ibm.com>
Reported-by: Jijo Varghese <vargjijo@in.ibm.com>
Signed-off-by: Nysal Jan K.A. <nysal@linux.ibm.com>
Reviewed-by: Nicholas Piggin <npiggin@gmail.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20240829022830.1164355-1-nysal@linux.ibm.com
Pull powerpc updates from Michael Ellerman: - Enable BPF Kernel Functions (kfuncs) in the powerpc BPF JIT. - Allow per-process DEXCR (Dynamic Execution Control Register) settings via prctl, notably NPHIE which controls hashst/hashchk for ROP protection. - Install powerpc selftests in sub-directories. Note this changes the way run_kselftest.sh needs to be invoked for powerpc selftests. - Change fadump (Firmware Assisted Dump) to better handle memory add/remove. - Add support for passing additional parameters to the fadump kernel. - Add support for updating the kdump image on CPU/memory add/remove events. - Other small features, cleanups and fixes. Thanks to Andrew Donnellan, Andy Shevchenko, Aneesh Kumar K.V, Arnd Bergmann, Benjamin Gray, Bjorn Helgaas, Christian Zigotzky, Christophe Jaillet, Christophe Leroy, Colin Ian King, Cédric Le Goater, Dr. David Alan Gilbert, Erhard Furtner, Frank Li, GUO Zihua, Ganesh Goudar, Geoff Levand, Ghanshyam Agrawal, Greg Kurz, Hari Bathini, Joel Stanley, Justin Stitt, Kunwu Chan, Li Yang, Lidong Zhong, Madhavan Srinivasan, Mahesh Salgaonkar, Masahiro Yamada, Matthias Schiffer, Naresh Kamboju, Nathan Chancellor, Nathan Lynch, Naveen N Rao, Nicholas Miehlbradt, Ran Wang, Randy Dunlap, Ritesh Harjani, Sachin Sant, Shirisha Ganta, Shrikanth Hegde, Sourabh Jain, Stephen Rothwell, sundar, Thorsten Blum, Vaibhav Jain, Xiaowei Bao, Yang Li, and Zhao Chenhui. * tag 'powerpc-6.10-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (85 commits) powerpc/fadump: Fix section mismatch warning powerpc/85xx: fix compile error without CONFIG_CRASH_DUMP powerpc/fadump: update documentation about bootargs_append powerpc/fadump: pass additional parameters when fadump is active powerpc/fadump: setup additional parameters for dump capture kernel powerpc/pseries/fadump: add support for multiple boot memory regions selftests/powerpc/dexcr: Fix spelling mistake "predicition" -> "prediction" KVM: PPC: Book3S HV nestedv2: Fix an error handling path in gs_msg_ops_kvmhv_nestedv2_config_fill_info() KVM: PPC: Fix documentation for ppc mmu caps KVM: PPC: code cleanup for kvmppc_book3s_irqprio_deliver KVM: PPC: Book3S HV nestedv2: Cancel pending DEC exception powerpc/xmon: Check cpu id in commands "c#", "dp#" and "dx#" powerpc/code-patching: Use dedicated memory routines for patching powerpc/code-patching: Test patch_instructions() during boot powerpc64/kasan: Pass virtual addresses to kasan_init_phys_region() powerpc: rename SPRN_HID2 define to SPRN_HID2_750FX powerpc: Fix typos powerpc/eeh: Fix spelling of the word "auxillary" and update comment macintosh/ams: Fix unused variable warning powerpc/Makefile: Remove bits related to the previous use of -mcmodel=large ...
There are places where CONFIG_MODULES guards the code that depends on memory allocation being done with module_alloc(). Replace CONFIG_MODULES with CONFIG_EXECMEM in such places. Signed-off-by: Mike Rapoport (IBM) <rppt@kernel.org> Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
The patching page set up as a writable alias may be in quadrant 0 (userspace) if the temporary mm path is used. This causes sanitiser failures if so. Sanitiser failures also occur on the non-mm path because the plain memset family is instrumented, and KASAN treats the patching window as poisoned. Introduce locally defined patch_* variants of memset that perform an uninstrumented lower level set, as well as detecting write errors like the original single patch variant does. copy_to_user() is not correct here, as the PTE makes it a proper kernel page (the EAA is privileged access only, RW). It just happens to be in quadrant 0 because that's the hardware's mechanism for using the current PID vs PID 0 in translations. Importantly, it's incorrect to allow user page accesses. Now that the patching memsets are used, we also propagate a failure up to the caller as the single patch variant does. Signed-off-by: Benjamin Gray <bgray@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20240325052815.854044-2-bgray@linux.ibm.com
patch_instructions() introduces new behaviour with a couple of variations. Test each case of * a repeated 32-bit instruction, * a repeated 64-bit instruction (ppc64), and * a copied sequence of instructions for both on a single page and when it crosses a page boundary. Signed-off-by: Benjamin Gray <bgray@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20240325052815.854044-1-bgray@linux.ibm.com
All supported compilers today (gcc v5.1+ and clang v11+) have support for -mcmodel=medium. As such, NO_MINIMAL_TOC is no longer being set. Remove NO_MINIMAL_TOC as well as the fallback to -mminimal-toc. Reviewed-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Naveen N Rao <naveen@kernel.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20240110141237.3179199-1-naveen@kernel.org
JUMP_LABEL_FEATURE_CHECK_DEBUG used static_key_intialized to determine
whether {cpu,mmu}_has_feature() is used before static keys were
initialized. However, {cpu,mmu}_has_feature() should not be used before
setup_feature_keys() is called but static_key_initialized is set well
before this by the call to jump_label_init() in early_init_devtree().
This creates a window in which JUMP_LABEL_FEATURE_CHECK_DEBUG will not
detect misuse and report errors. Add a flag specifically to indicate
when {cpu,mmu}_has_feature() is safe to use.
Signed-off-by: Nicholas Miehlbradt <nicholas@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://msgid.link/20240408052358.5030-1-nicholas@linux.ibm.com
arch/powerpc/lib/xor_vmx.o is built with '-msoft-float' (from the main powerpc Makefile) and '-maltivec' (from its CFLAGS), which causes an error when building with clang after a recent change in main: error: option '-msoft-float' cannot be specified with '-maltivec' make[6]: *** [scripts/Makefile.build:243: arch/powerpc/lib/xor_vmx.o] Error 1 Explicitly add '-mhard-float' before '-maltivec' in xor_vmx.o's CFLAGS to override the previous inclusion of '-msoft-float' (as the last option wins), which matches how other areas of the kernel use '-maltivec', such as AMDGPU. Cc: stable@vger.kernel.org Closes: https://github.com/ClangBuiltLinux/linux/issues/1986 Link: https://github.com/llvm/llvm-project/commit/4792f912b232141ecba4cbae538873be3c28556c Signed-off-by: Nathan Chancellor <nathan@kernel.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20240127-ppc-xor_vmx-drop-msoft-float-v1-1-f24140e81376@kernel.org
There's an almost identical code sequence to specify load/store access hints in __copy_tofrom_user_power7(), copypage_power7() and memcpy_power7(). Move the sequence into a common macro, which is passed the registers to use as they differ slightly. There also needs to be a copy in the selftests, it could be shared in future if the headers are cleaned up / refactored. Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20240229122521.762431-1-mpe@ellerman.id.au
There is a nice macro to check user mode. Use it instead of open coding anding with MSR_PR to increase readability and avoid having to comment what that anding is for. Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/fbf74887dcf1f1ba9e1680fc3247cbb581b00662.1708078228.git.christophe.leroy@csgroup.eu