[PATCH] erofs: mark inodes without acls in erofs_read_inode()

[Gao Xiang]

On 2026/1/28 15:03, Hongbo Li wrote:
> Hi, Xiang
> 
> On 2026/1/28 11:54, Gao Xiang wrote:
>> Similar to commit 91ef18b567da ("ext4: mark inodes without acls in
>> __ext4_iget()"), the ACL state won't be read when the file owner
>> performs a lookup, and the RCU fast path for lookups won't work
>> because the ACL state remains unknown.
>>
>> If there are no extended attributes, or if the xattr filter
>> indicates that no ACL xattr is present, call cache_no_acl() directly.
>>
>> Signed-off-by: Gao Xiang <hsiangkao at linux.alibaba.com>
>> ---
>>   fs/erofs/inode.c |  5 +++++
>>   fs/erofs/xattr.c | 20 ++++++++++++++++++++
>>   fs/erofs/xattr.h |  1 +
>>   3 files changed, 26 insertions(+)
>>
>> diff --git a/fs/erofs/inode.c b/fs/erofs/inode.c
>> index bce98c845a18..2e02d4b466ce 100644
>> --- a/fs/erofs/inode.c
>> +++ b/fs/erofs/inode.c
>> @@ -137,6 +137,11 @@ static int erofs_read_inode(struct inode *inode)
>>           err = -EFSCORRUPTED;
>>           goto err_out;
>>       }
>> +
>> +    if (IS_ENABLED(CONFIG_EROFS_FS_POSIX_ACL) &&
>> +        erofs_inode_has_noacl(inode, ptr, ofs))
>> +        cache_no_acl(inode);
>> +
>>       switch (inode->i_mode & S_IFMT) {
>>       case S_IFDIR:
>>           vi->dot_omitted = (ifmt >> EROFS_I_DOT_OMITTED_BIT) & 1;
>> diff --git a/fs/erofs/xattr.c b/fs/erofs/xattr.c
>> index 512b998bdfff..14d22adc1476 100644
>> --- a/fs/erofs/xattr.c
>> +++ b/fs/erofs/xattr.c
>> @@ -574,4 +574,24 @@ struct posix_acl *erofs_get_acl(struct inode *inode, int type, bool rcu)
>>       kfree(value);
>>       return acl;
>>   }
>> +
>> +bool erofs_inode_has_noacl(struct inode *inode, void *kaddr, unsigned int ofs)
>> +{
> 
> How about put the definition and declare of erofs_inode_has_noacl before the erofs_get_acl helper? Since it is no need to resolved the context conflicts for page sharing.

I've resolved it manually.

> 
> Otherwise it looks good to me:
> Reviewed-by: Hongbo Li <lihongbo22 at huawei.com>

Thanks,
Gao Xiang

> 
> Thanks,
> Hongbo