Re: [PATCH v2 1/5] powerpc/mm: Introduce temporary mm

[Daniel Axtens <dja@axtens.net>]

Hi Chris,
  
>  void __set_breakpoint(int nr, struct arch_hw_breakpoint *brk);
> +void __get_breakpoint(int nr, struct arch_hw_breakpoint *brk);
>  bool ppc_breakpoint_available(void);
>  #ifdef CONFIG_PPC_ADV_DEBUG_REGS
>  extern void do_send_trap(struct pt_regs *regs, unsigned long address,
> diff --git a/arch/powerpc/include/asm/mmu_context.h b/arch/powerpc/include/asm/mmu_context.h
> index 1a474f6b1992..9269c7c7b04e 100644
> --- a/arch/powerpc/include/asm/mmu_context.h
> +++ b/arch/powerpc/include/asm/mmu_context.h
> @@ -10,6 +10,7 @@
>  #include <asm/mmu.h>	
>  #include <asm/cputable.h>
>  #include <asm/cputhreads.h>
> +#include <asm/debug.h>
>  
>  /*
>   * Most if the context management is out of line
> @@ -300,5 +301,68 @@ static inline int arch_dup_mmap(struct mm_struct *oldmm,
>  	return 0;
>  }
>  
> +struct temp_mm {
> +	struct mm_struct *temp;
> +	struct mm_struct *prev;
> +	bool is_kernel_thread;
> +	struct arch_hw_breakpoint brk[HBP_NUM_MAX];
> +};

This is on the nitpicky end, but I wonder if this should be named
temp_mm, or should be labelled something else to capture its broader
purpose as a context for code patching? I'm thinking that a store of
breakpoints is perhaps unusual in a memory-managment structure?

I don't have a better suggestion off the top of my head and I'm happy
for you to leave it, I just wanted to flag it as a possible way we could
be clearer.

> +
> +static inline void init_temp_mm(struct temp_mm *temp_mm, struct mm_struct *mm)
> +{
> +	temp_mm->temp = mm;
> +	temp_mm->prev = NULL;
> +	temp_mm->is_kernel_thread = false;
> +	memset(&temp_mm->brk, 0, sizeof(temp_mm->brk));
> +}
> +
> +static inline void use_temporary_mm(struct temp_mm *temp_mm)
> +{
> +	lockdep_assert_irqs_disabled();
> +
> +	temp_mm->is_kernel_thread = current->mm == NULL;
> +	if (temp_mm->is_kernel_thread)
> +		temp_mm->prev = current->active_mm;

You don't seem to restore active_mm below. I don't know what active_mm
does, so I don't know if this is a problem.

> +	else
> +		temp_mm->prev = current->mm;
> +
> +	/*
> +	 * Hash requires a non-NULL current->mm to allocate a userspace address
> +	 * when handling a page fault. Does not appear to hurt in Radix either.
> +	 */
> +	current->mm = temp_mm->temp;
> +	switch_mm_irqs_off(NULL, temp_mm->temp, current);
> +
> +	if (ppc_breakpoint_available()) {

I wondered if this could be changed during a text-patching operation.
AIUI, it potentially can on a P9 via "dawr_enable_dangerous" in debugfs.

I don't know if that's a problem. My concern is that you could turn off
breakpoints, call 'use_temporary_mm', then turn them back on again
before 'unuse_temporary_mm' and get a breakpoint while that can access
the temporary mm. Is there something else that makes that safe?
disabling IRQs maybe?

> +		struct arch_hw_breakpoint null_brk = {0};
> +		int i = 0;
> +
> +		for (; i < nr_wp_slots(); ++i) {

super nitpicky, and I'm not sure if this is actually documented, but I'd
usually see this written as:

for (i = 0; i < nr_wp_slots(); i++) {

Not sure if there's any reason that it _shouldn't_ be written the way
you've written it (and I do like initialising the variable when it's
defined!), I'm just not used to it. (Likewise with the unuse function.)

> +			__get_breakpoint(i, &temp_mm->brk[i]);
> +			if (temp_mm->brk[i].type != 0)
> +				__set_breakpoint(i, &null_brk);
> +		}
> +	}
> +}
> +

Kind regards,
Daniel

> +static inline void unuse_temporary_mm(struct temp_mm *temp_mm)
> +{
> +	lockdep_assert_irqs_disabled();
> +
> +	if (temp_mm->is_kernel_thread)
> +		current->mm = NULL;
> +	else
> +		current->mm = temp_mm->prev;
> +	switch_mm_irqs_off(NULL, temp_mm->prev, current);
> +
> +	if (ppc_breakpoint_available()) {
> +		int i = 0;
> +
> +		for (; i < nr_wp_slots(); ++i)
> +			if (temp_mm->brk[i].type != 0)
> +				__set_breakpoint(i, &temp_mm->brk[i]);
> +	}
> +}
> +
>  #endif /* __KERNEL__ */
>  #endif /* __ASM_POWERPC_MMU_CONTEXT_H */
> diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c
> index 4650b9bb217f..b6c123bf5edd 100644
> --- a/arch/powerpc/kernel/process.c
> +++ b/arch/powerpc/kernel/process.c
> @@ -824,6 +824,11 @@ static inline int set_breakpoint_8xx(struct arch_hw_breakpoint *brk)
>  	return 0;
>  }
>  
> +void __get_breakpoint(int nr, struct arch_hw_breakpoint *brk)
> +{
> +	memcpy(brk, this_cpu_ptr(&current_brk[nr]), sizeof(*brk));
> +}
> +
>  void __set_breakpoint(int nr, struct arch_hw_breakpoint *brk)
>  {
>  	memcpy(this_cpu_ptr(&current_brk[nr]), brk, sizeof(*brk));
> -- 
> 2.27.0