linux-stable.git/arch/powerpc/kernel/time.c, branch linux-7.1.y

powerpc/time: Remove redundant preempt_disable|enable() calls from arch_irq_work_raise()

2026-05-14T05:45:26+00:00

A kernel panic is observed when handling machine check exceptions from
real mode.

  BUG: Unable to handle kernel data access on read at 0xc00000006be21300
  Oops: Kernel access of bad area, sig: 11 [#1]
  MSR:  8000000000001003   CR: 88222248  XER: 00000005
  CFAR: c00000000003ffc4 DAR: c00000006be21300 DSISR: 40000000 IRQMASK: 0
  NIP [c000000000029e40] arch_irq_work_raise+0x10/0x70
  LR [c00000000003ffc8] machine_check_queue_event+0xa8/0x150
  Call Trace:
  [c0000000179d3c70] [c00000000003ff64] machine_check_queue_event+0x44/0x150
  [c0000000179d3d30] [c0000000000084e0] machine_check_early_common+0x1f0/0x2c0

The crash occurs because arch_irq_work_raise() calls preempt_disable()
from machine check exception (MCE) handlers running in real mode. In
this context, accessing the preempt_count can fault, leading to the panic.

The preempt_disable()/preempt_enable() pair in arch_irq_work_raise()
was originally added by commit 0fe1ac48bef0 ("powerpc/perf_event: Fix
oops due to perf_event_do_pending call") to avoid races while raising
irq work from exception context.

Later, commit 471ba0e686cb ("irq_work: Do not raise an IPI when
queueing work on the local CPU") added preemption protection in
irq_work_queue() path, while commit 20b876918c06 ("irq_work: Use per
cpu atomics instead of regular atomics") added equivalent
protection in irq_work_queue_on() before reaching arch_irq_work_raise():

  irq_work_queue() / irq_work_queue_on()
    -> preempt_disable()
      -> __irq_work_queue_local()
        -> irq_work_raise()
          -> arch_irq_work_raise()

As a result, callers other than mce_irq_work_raise() already execute
with preemption disabled, making the additional
preempt_disable()/preempt_enable() pair in arch_irq_work_raise()
redundant.

The arch_irq_work_raise() function executes in NMI context when called
from MCE handler. Hence we will not be preempted or scheduled out since
we are in NMI context with MSR[EE]=0. Therefore, it is safe to remove
the preempt_disable()/preempt_enable() calls from here.

Remove it to avoid accessing preempt_count from real mode context.

Fixes: cc15ff327569 ("powerpc/mce: Avoid using irq_work_queue() in realmode")
Suggested-by: Mahesh Salgaonkar 
Acked-by: Shrikanth Hegde 
Reviewed-by: Ritesh Harjani (IBM) 
Signed-off-by: Sayali Patil 
[Maddy: Fixed the commit title]
Signed-off-by: Madhavan Srinivasan 
Link: https://patch.msgid.link/20260513081413.222490-1-sayalip@linux.ibm.com

powerpc/time: Expose boot_tb via accessor

2025-09-22T09:18:56+00:00

- Define accessor function get_boot_tb() to safely return boot_tb value,
  this is only needed when running in SPLPAR environments, so the
  accessor is built conditionally under CONFIG_PPC_SPLPAR.

- Tag boot_tb as __ro_after_init since it is written once at initialized
  and never updated afterwards.

Signed-off-by: Aboorva Devarajan 
Tested-by: Tejas Manhas 
Tested-by: Venkat Rao Bagalkote 
Signed-off-by: Madhavan Srinivasan 
Link: https://patch.msgid.link/20250915102947.26681-2-atrajeev@linux.ibm.com

Merge tag 'powerpc-6.15-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux

2025-03-28T02:39:08+00:00

Pull powerpc updates from Madhavan Srinivasan:

 - Remove support for IBM Cell Blades

 - SMP support for microwatt platform

 - Support for inline static calls on PPC32

 - Enable pmu selftests for power11 platform

 - Enable hardware trace macro (HTM) hcall support

 - Support for limited address mode capability

 - Changes to RMA size from 512 MB to 768 MB to handle fadump

 - Misc fixes and cleanups

Thanks to Abhishek Dubey, Amit Machhiwal, Andreas Schwab, Arnd Bergmann,
Athira Rajeev, Avnish Chouhan, Christophe Leroy, Disha Goel, Donet Tom,
Gaurav Batra, Gautam Menghani, Hari Bathini, Kajol Jain, Kees Cook,
Mahesh Salgaonkar, Michael Ellerman, Paul Mackerras, Ritesh Harjani
(IBM), Sathvika Vasireddy, Segher Boessenkool, Sourabh Jain, Vaibhav
Jain, and Venkat Rao Bagalkote.

* tag 'powerpc-6.15-1' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux: (61 commits)
  powerpc/kexec: fix physical address calculation in clear_utlb_entry()
  crypto: powerpc: Mark ghashp8-ppc.o as an OBJECT_FILES_NON_STANDARD
  powerpc: Fix 'intra_function_call not a direct call' warning
  powerpc/perf: Fix ref-counting on the PMU 'vpa_pmu'
  KVM: PPC: Enable CAP_SPAPR_TCE_VFIO on pSeries KVM guests
  powerpc/prom_init: Fixup missing #size-cells on PowerBook6,7
  powerpc/microwatt: Add SMP support
  powerpc: Define config option for processors with broadcast TLBIE
  powerpc/microwatt: Define an idle power-save function
  powerpc/microwatt: Device-tree updates
  powerpc/microwatt: Select COMMON_CLK in order to get the clock framework
  net: toshiba: Remove reference to PPC_IBM_CELL_BLADE
  net: spider_net: Remove powerpc Cell driver
  cpufreq: ppc_cbe: Remove powerpc Cell driver
  genirq: Remove IRQ_EDGE_EOI_HANDLER
  docs: Remove reference to removed CBE_CPUFREQ_SPU_GOVERNOR
  powerpc: Remove UDBG_RTAS_CONSOLE
  powerpc/io: Use standard barrier macros in io.c
  powerpc/io: Rename _insw_ns() etc.
  powerpc/io: Use generic raw accessors
  ...

powerpc/time: Define div128_by_32() static and __init

2025-02-24T06:56:21+00:00

div128_by_32() used to be called from outside time.c in the old days
but since v2.6.15 it hasn't been used outside time.c

$ git grep div128_by_32 v2.6.14
v2.6.14:arch/ppc64/kernel/iSeries_setup.c:	div128_by_32(1024 * 1024, 0, tb_ticks_per_sec, &divres);
v2.6.14:arch/ppc64/kernel/pmac_time.c:	div128_by_32( 1024*1024, 0, tb_ticks_per_sec, &divres );
v2.6.14:arch/ppc64/kernel/time.c:				div128_by_32( XSEC_PER_SEC, 0, tb_ticks_per_sec, &divres );
v2.6.14:arch/ppc64/kernel/time.c:	div128_by_32(1024*1024, 0, tb_ticks_per_sec, &divres);
v2.6.14:arch/ppc64/kernel/time.c:	div128_by_32(1000000000, 0, tb_ticks_per_sec, &res);
v2.6.14:arch/ppc64/kernel/time.c:	div128_by_32( 1024*1024, 0, new_tb_ticks_per_sec, &divres );
v2.6.14:arch/ppc64/kernel/time.c:void div128_by_32( unsigned long dividend_high, unsigned long dividend_low,
v2.6.14:include/asm-ppc64/time.h:void div128_by_32( unsigned long dividend_high, unsigned long dividend_low,

$ git grep div128_by_32 v2.6.15
v2.6.15:arch/powerpc/kernel/time.c:				div128_by_32( XSEC_PER_SEC, 0, tb_ticks_per_sec, &divres );
v2.6.15:arch/powerpc/kernel/time.c:	div128_by_32(1024*1024, 0, tb_ticks_per_sec, &res);
v2.6.15:arch/powerpc/kernel/time.c:	div128_by_32(1000000000, 0, tb_ticks_per_sec, &res);
v2.6.15:arch/powerpc/kernel/time.c:	div128_by_32(1024*1024, 0, new_tb_ticks_per_sec, &divres);
v2.6.15:arch/powerpc/kernel/time.c:void div128_by_32(u64 dividend_high, u64 dividend_low,
v2.6.15:include/asm-powerpc/time.h:extern void div128_by_32(u64 dividend_high, u64 dividend_low,

Move it above its only caller which is time_init() and define it
static and __init.

Signed-off-by: Christophe Leroy 
Signed-off-by: Madhavan Srinivasan 
Link: https://patch.msgid.link/50810349bf1eee378fbeab72a36e4b6553a60c3d.1738749246.git.christophe.leroy@csgroup.eu

powerpc/vdso: Switch to generic storage implementation

2025-02-21T08:54:02+00:00

The generic storage implementation provides the same features as the
custom one. However it can be shared between architectures, making
maintenance easier.

Co-developed-by: Nam Cao 
Signed-off-by: Nam Cao 
Signed-off-by: Thomas Weißschuh 
Signed-off-by: Thomas Gleixner 
Reviewed-by: Christophe Leroy 
Link: https://lore.kernel.org/all/20250204-vdso-store-rng-v3-14-13a4669dfc8c@linutronix.de

powerpc: Split systemcfg struct definitions out from vdso

2024-11-02T11:37:36+00:00

The systemcfg data has nothing to do anymore with the vdso.
Split it into a dedicated header file.

Signed-off-by: Thomas Weißschuh 
Signed-off-by: Thomas Gleixner 
Link: https://lore.kernel.org/all/20241010-vdso-generic-base-v1-27-b64f0842d512@linutronix.de

powerpc: Split systemcfg data out of vdso data page

2024-11-02T11:37:36+00:00

The systemcfg data only has minimal overlap with the vdso data.
Splitting the two avoids mapping the implementation-defined vdso data
into /proc/ppc64/systemcfg.
It is also a preparation for the standardization of vdso data storage.

The only field actually used by both systemcfg and vdso is
tb_ticks_per_sec and it is only changed once during time_init().
Initialize it in both structures there.

Signed-off-by: Thomas Weißschuh 
Signed-off-by: Thomas Gleixner 
Link: https://lore.kernel.org/all/20241010-vdso-generic-base-v1-26-b64f0842d512@linutronix.de

powerpc/vdso: Remove timekeeper includes

2024-10-15T15:50:29+00:00

Since the generic VDSO clock mode storage is used, this header file is
unused and can be removed.

This avoids including a non-VDSO header while building the VDSO,
which can lead to compilation errors.

Also drop the comment which is out of date and in the wrong place.

Signed-off-by: Thomas Weißschuh 
Signed-off-by: Thomas Gleixner 
Link: https://lore.kernel.org/all/20241010-vdso-generic-arch_update_vsyscall-v1-4-7fe5a3ea4382@linutronix.de

powerpc/4xx: Remove CONFIG_BOOKE_OR_40x

2024-06-28T12:28:48+00:00

Now that 40x is gone, replace CONFIG_BOOKE_OR_40x by CONFIG_BOOKE.

Signed-off-by: Michael Ellerman 
Link: https://msgid.link/20240628121201.130802-5-mpe@ellerman.id.au

sched/vtime: Get rid of generic vtime_task_switch() implementation

2024-04-17T11:37:20+00:00

The generic vtime_task_switch() implementation gets built only
if __ARCH_HAS_VTIME_TASK_SWITCH is not defined, but requires an
architecture to implement arch_vtime_task_switch() callback at
the same time, which is confusing.

Further, arch_vtime_task_switch() is implemented for 32-bit PowerPC
architecture only and vtime_task_switch() generic variant is rather
superfluous.

Simplify the whole vtime_task_switch() wiring by moving the existing
generic implementation to PowerPC.

Signed-off-by: Alexander Gordeev 
Signed-off-by: Ingo Molnar 
Reviewed-by: Frederic Weisbecker 
Reviewed-by: Nicholas Piggin 
Acked-by: Michael Ellerman 
Link: https://lore.kernel.org/r/2cb6e3caada93623f6d4f78ad938ac6cd0e2fda8.1712760275.git.agordeev@linux.ibm.com