Re: [RESEND] [PATCH] eventfs: Have inodes have unique inode numbers

[Mathieu Desnoyers]

On 2024-01-26 15:12, Steven Rostedt wrote:
[...]
> diff --git a/fs/tracefs/inode.c b/fs/tracefs/inode.c
> index e1b172c0e091..2187be6d7b23 100644
> --- a/fs/tracefs/inode.c
> +++ b/fs/tracefs/inode.c
> @@ -223,13 +223,41 @@ static const struct inode_operations tracefs_file_inode_operations = {
>   	.setattr	= tracefs_setattr,
>   };
>   
> +/* Copied from get_next_ino() but adds allocation for multiple inodes */
> +#define LAST_INO_BATCH 1024
> +#define LAST_INO_MASK (~(LAST_INO_BATCH - 1))
> +static DEFINE_PER_CPU(unsigned int, last_ino);
> +
> +unsigned int tracefs_get_next_ino(int files)
> +{
> +	unsigned int *p = &get_cpu_var(last_ino);
> +	unsigned int res = *p;
> +
> +#ifdef CONFIG_SMP
> +	/* Check if adding files+1 overflows */

How does it handle a @files input where:

* (files+1 > LAST_INO_BATCH) ?

* (files+1 == LAST_INO_BATCH) ?

> +	if (unlikely(!res || (res & LAST_INO_MASK) != ((res + files + 1) & LAST_INO_MASK))) {
> +		static atomic_t shared_last_ino;
> +		int next = atomic_add_return(LAST_INO_BATCH, &shared_last_ino);
> +
> +		res = next - LAST_INO_BATCH;
> +	}
> +#endif
> +
> +	res++;
> +	/* get_next_ino should not provide a 0 inode number */
> +	if (unlikely(!res))
> +		res++;

I suspect that bumping this res++ in the 0 case can cause inode range
reservation issues at (files+1 == LAST_INO_BATCH-1).

Thanks,

Mathieu

> +	*p = res + files;
> +	put_cpu_var(last_ino);
> +	return res;
> +}

--
Mathieu Desnoyers
EfficiOS Inc.
https://www.efficios.com