Re: [PATCH] exfat: speed up file creation in large directories

[Chi Zhiling]

Hi, Yang Wen


On 8/28/26 10:58 PM, Yang Wen wrote:
> Negative name lookups scan a directory from the beginning. Repeating
> this scan before creating each file makes bulk file creation approach
> O(N^2).
>
> Add a 64 KiB per-directory Bloom filter and retain the next empty entry
> found during directory scans. A definite Bloom filter miss skips the
> directory scan, while a possible match falls back to the normal lookup
> path, preserving correctness in the presence of hash collisions.

I think this is a nice optimization, and for large-directory workloads, 
the 64 KB overhead is acceptable to me.

>   static void exfat_free_inode(struct inode *inode)
>   {
> +	exfat_name_filter_free(inode);
The exfat_free_inode() could be called in softirq context, see 
destroy_inode() in fs/inode.c, which calls free_inode() via call_rcu(),

so you can't call exfat_name_filter_free() from exfat_free_inode(). It 
should be moved to exfat_evict_inode(), otherwise LOCKDEP will report a 
warning.


Thanks,