On Wed, Mar 11, 2020 at 12:30:41PM +0800, kbuild test robot wrote: > tree: https://github.com/hnaz/linux-mm master > head: 5d9ee416b5701096536c7a63c04dbe25012baa9e > commit: 79181dfcc4b8c2091126a8b3175957a598ae61e1 [90/340] x86/mm: use helper fault_signal_pending() > reproduce: > # apt-get install sparse > # sparse version: v0.6.1-174-g094d5a94-dirty > git checkout 79181dfcc4b8c2091126a8b3175957a598ae61e1 > make ARCH=x86_64 allmodconfig > make C=1 CF='-fdiagnostic-prefix -D__CHECK_ENDIAN__' > > If you fix the issue, kindly add following tag > Reported-by: kbuild test robot <lkp@xxxxxxxxx> > > > sparse warnings: (new ones prefixed by >>) > > >> arch/x86/mm/fault.c:1468:34: sparse: sparse: incorrect type in argument 1 (different base types) @@ expected unsigned int fault_flags @@ got restricted vm_fault_t [assignedunsigned int fault_flags @@ > >> arch/x86/mm/fault.c:1468:34: sparse: expected unsigned int fault_flags > >> arch/x86/mm/fault.c:1468:34: sparse: got restricted vm_fault_t [assigned] [usertype] fault > >> include/linux/sched/signal.h:381:16: sparse: sparse: restricted vm_fault_t degrades to integer Hi, Andrew, I'm not sure whether this comes from the extra "-fdiagnostic-prefix" because I didn't encounter that, but if after the previous signal.h fixup squashed into patch "mm: Introduce fault_signal_pending()", then we have linux/mm_types.h included already, and it'll be pretty safe to just use vm_fault_t here. To be explicit, maybe another squash-in would work to the same patch: -------------8<------------- diff --git a/include/linux/sched/signal.h b/include/linux/sched/signal.h index c8e0e4f801e4..65ea429daaa2 100644 --- a/include/linux/sched/signal.h +++ b/include/linux/sched/signal.h @@ -377,7 +377,7 @@ static inline int signal_pending_state(long state, struct task_struct *p) * instead, especially with the case where we've got interrupted with * a VM_FAULT_RETRY. */ -static inline bool fault_signal_pending(unsigned int fault_flags, +static inline bool fault_signal_pending(vm_fault_t fault_flags, struct pt_regs *regs) { return unlikely((fault_flags & VM_FAULT_RETRY) && ------------->8------------- So the whole fixup is (against "mm: Introduce fault_signal_pending()"): -------------8<------------- diff --git a/arch/mips/mm/fault.c b/arch/mips/mm/fault.c index fd64b135fd7b..4a0eafe3d932 100644 --- a/arch/mips/mm/fault.c +++ b/arch/mips/mm/fault.c @@ -154,7 +154,7 @@ static void __kprobes __do_page_fault(struct pt_regs *regs, unsigned long write, */ fault = handle_mm_fault(vma, address, flags); - if (fault_signal_pending(regs)) + if (fault_signal_pending(fault, regs)) return; perf_sw_event(PERF_COUNT_SW_PAGE_FAULTS, 1, regs, address); diff --git a/include/linux/sched/signal.h b/include/linux/sched/signal.h index 09d40ce6a162..65ea429daaa2 100644 --- a/include/linux/sched/signal.h +++ b/include/linux/sched/signal.h @@ -10,6 +10,8 @@ #include <linux/cred.h> #include <linux/refcount.h> #include <linux/posix-timers.h> +#include <linux/mm_types.h> +#include <asm/ptrace.h> /* * Types defining task->signal and task->sighand and APIs using them: @@ -375,7 +377,7 @@ static inline int signal_pending_state(long state, struct task_struct *p) * instead, especially with the case where we've got interrupted with * a VM_FAULT_RETRY. */ -static inline bool fault_signal_pending(unsigned int fault_flags, +static inline bool fault_signal_pending(vm_fault_t fault_flags, struct pt_regs *regs) { return unlikely((fault_flags & VM_FAULT_RETRY) && ------------->8------------- Thanks, -- Peter Xu
