Re: [PATCH v3 1/1] pgo: Fix allocate_node() v2

[Nathan Chancellor]

On 6/4/2021 9:58 AM, Jarmo Tiitto wrote:
> When clang instrumentation eventually calls allocate_node()
> the struct llvm_prf_data *p argument tells us from what section
> we should reserve the vnode: It either points into vmlinux's
> core __llvm_prf_data section or some loaded module's
> __llvm_prf_data section.
>
> But since we don't have access to corresponding
> __llvm_prf_vnds section(s) for any module, the function
> should return just NULL and ignore any profiling attempts
> from modules for now.
>
> Signed-off-by: Jarmo Tiitto <jarmo.tiitto@xxxxxxxxx>

Reviewed-by: Nathan Chancellor <nathan@xxxxxxxxxx>

> ---
> Based on Kees and others feedback here is v3 patch
> that clarifies why the current checks in allocate_node()
> are flawed. I did fair amount of KGDB time on it.
>
> The commit is based on kees/for-next/clang/features tree,
> hopefully this is ok. Should I have based it on linux-next
> instead?
>
> I grep -R'd where the memory_contains() can be found and it is only
> found in #include <asm-generic/sections.h>
>
> I cross my fingers and await if this is my first accepted patch. :-)
> ---
>   kernel/pgo/instrument.c | 23 ++++++++++++++---------
>   1 file changed, 14 insertions(+), 9 deletions(-)
>
> diff --git a/kernel/pgo/instrument.c b/kernel/pgo/instrument.c
> index 0e07ee1b17d9..c69b4f7ebaad 100644
> --- a/kernel/pgo/instrument.c
> +++ b/kernel/pgo/instrument.c
> @@ -18,6 +18,7 @@
>   
>   #define pr_fmt(fmt)	"pgo: " fmt
>   
> +#include <asm-generic/sections.h>
>   #include <linux/bitops.h>
>   #include <linux/kernel.h>
>   #include <linux/export.h>
> @@ -55,17 +56,21 @@ void prf_unlock(unsigned long flags)
>   static struct llvm_prf_value_node *allocate_node(struct llvm_prf_data *p,
>   						 u32 index, u64 value)
>   {
> -	if (&__llvm_prf_vnds_start[current_node + 1] >= __llvm_prf_vnds_end)
> -		return NULL; /* Out of nodes */
> -
> -	current_node++;
> -
> -	/* Make sure the node is entirely within the section */
> -	if (&__llvm_prf_vnds_start[current_node] >= __llvm_prf_vnds_end ||
> -	    &__llvm_prf_vnds_start[current_node + 1] > __llvm_prf_vnds_end)
> +	const int max_vnds = prf_vnds_count();
> +
> +	/*
> +	 * Check that p is within vmlinux __llvm_prf_data section.
> +	 * If not, don't allocate since we can't handle modules yet.
> +	 */
> +	if (!memory_contains(__llvm_prf_data_start,
> +		__llvm_prf_data_end, p, sizeof(*p)))
>   		return NULL;
>   
> -	return &__llvm_prf_vnds_start[current_node];
> +	if (WARN_ON_ONCE(current_node >= max_vnds))
> +		return NULL; /* Out of nodes */
> +
> +	/* reserve vnode for vmlinux */
> +	return &__llvm_prf_vnds_start[current_node++];
>   }
>   
>   /*
>
> base-commit: 5d0cda65918279ada060417c5fecb7e86ccb3def