Re: [PATCH] ftrace: Fix possible warning on checking all pages used in ftrace_process_locs()

[Zheng Yejian]

On 2023/7/10 22:46, Steven Rostedt wrote:
> On Tue, 11 Jul 2023 05:29:58 +0800
> Zheng Yejian <zhengyejian1@xxxxxxxxxx> wrote:
>
> > As comments in ftrace_process_locs(), there may be NULL pointers in
> > mcount_loc section:
> >   > Some architecture linkers will pad between
> >   > the different mcount_loc sections of different
> >   > object files to satisfy alignments.
> >   > Skip any NULL pointers.
> >
> > After 20e5227e9f55 ("ftrace: allow NULL pointers in mcount_loc"),
> > NULL pointers will be accounted when allocating ftrace pages but
> > skipped before adding into ftrace pages, this may result in some
> > pages not being used. Then after 706c81f87f84 ("ftrace: Remove extra
> > helper functions"), warning may occur at:
> >    WARN_ON(pg->next);
> >
> > So we may need to skip NULL pointers before allocating ftrace pages.
> >
> > Fixes: 706c81f87f84 ("ftrace: Remove extra helper functions")
> > Signed-off-by: Zheng Yejian <zhengyejian1@xxxxxxxxxx>
> > ---
> >   kernel/trace/ftrace.c | 10 ++++++++++
> >   1 file changed, 10 insertions(+)
> >
> > diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> > index 3740aca79fe7..5b474165df31 100644
> > --- a/kernel/trace/ftrace.c
> > +++ b/kernel/trace/ftrace.c
> > @@ -6485,6 +6485,16 @@ static int ftrace_process_locs(struct module *mod,
> >   	if (!count)
> >   		return 0;
> >   
> > +	p = start;
> > +	while (p < end) {
> > +		/*
> > +		 * Refer to conments below, there may be NULL pointers,
> > +		 * skip them before allocating pages
> > +		 */
> > +		addr = ftrace_call_adjust(*p++);
> > +		if (!addr)
> > +			count--;
> > +	}
>
> My main concern about this is the added overhead during boot to process
> this. There's 10s of thousands of functions, so this loop will be 10s of
> thousands. I also don't like that this is an unconditional loop (meaning it
> executes even when it is unnecessary to do so).

Agreed! The added overhead probably superfluousin in most cases.

> >   	/*
> >   	 * Sorting mcount in vmlinux at build time depend on
> >   	 * CONFIG_BUILDTIME_MCOUNT_SORT, while mcount loc in
>
> How about something like this?
>
> -- Steve
>
> diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c
> index b24c573934af..acd033371721 100644
> --- a/kernel/trace/ftrace.c
> +++ b/kernel/trace/ftrace.c
> @@ -6474,6 +6474,7 @@ static int ftrace_process_locs(struct module *mod,
>   	struct ftrace_page *start_pg;
>   	struct ftrace_page *pg;
>   	struct dyn_ftrace *rec;
> +	unsigned long skipped = 0;
>   	unsigned long count;
>   	unsigned long *p;
>   	unsigned long addr;
> @@ -6536,8 +6537,10 @@ static int ftrace_process_locs(struct module *mod,
>   		 * object files to satisfy alignments.
>   		 * Skip any NULL pointers.
>   		 */
> -		if (!addr)
> +		if (!addr) {
> +			skipped++;
>   			continue;
> +		}
>   
>   		end_offset = (pg->index+1) * sizeof(pg->records[0]);
>   		if (end_offset > PAGE_SIZE << pg->order) {
> @@ -6551,12 +6554,24 @@ static int ftrace_process_locs(struct module *mod,
>   		rec->ip = addr;
>   	}
>   
> -	/* We should have used all pages */
> -	WARN_ON(pg->next);
> -
>   	/* Assign the last page to ftrace_pages */
>   	ftrace_pages = pg;
>   
> +	/* We should have used all pages unless we skipped some */
> +	if (pg->next) {
> +		WARN_ON(!skipped);
> +		while (ftrace_pages->next) {
> +			pg = ftrace_pages->next;
> +			ftrace_pages->next = pg->next;
> +			if (pg->records) {
> +				free_pages((unsigned long)pg->records, pg->order);
> +				ftrace_number_of_pages -= 1 << pg->order;
> +			}
> +			kfree(pg);
> +			ftrace_number_of_groups--;
> +		}

Do we only need to free the pages that not being used?

> +	}
> +
>   	/*
>   	 * We only need to disable interrupts on start up
>   	 * because we are modifying code that an interrupt