Re: [PATCH v3 14/25] fuse: refactor io-uring header copying to ring

[Bernd Schubert <bernd@bsbernd.com>]

On 1/16/26 23:33, Joanne Koong wrote:
> On Sun, Jan 11, 2026 at 8:04 AM Bernd Schubert <bschubert@ddn.com> wrote:
>>
>> On 12/23/25 01:35, Joanne Koong wrote:
>>> Move header copying to ring logic into a new copy_header_to_ring()
>>> function. This consolidates error handling.
>>>
>>> Signed-off-by: Joanne Koong <joannelkoong@gmail.com>
>>> ---
>>>  fs/fuse/dev_uring.c | 39 +++++++++++++++++++++------------------
>>>  1 file changed, 21 insertions(+), 18 deletions(-)
>>>
>>> diff --git a/fs/fuse/dev_uring.c b/fs/fuse/dev_uring.c
>>> index 1efee4391af5..7962a9876031 100644
>>> --- a/fs/fuse/dev_uring.c
>>> +++ b/fs/fuse/dev_uring.c
>>> @@ -575,6 +575,18 @@ static int fuse_uring_out_header_has_err(struct fuse_out_header *oh,
>>>       return err;
>>>  }
>>>
>>> +static __always_inline int copy_header_to_ring(void __user *ring,
>>> +                                            const void *header,
>>> +                                            size_t header_size)
>>
>> Minor nit: The only part I don't like too much is the __always_inline. I
>> had at least two times a debug issue where I didn't get much out of the
>> trace and then used for fuse.ko
> 
> Unfortunately the __always_inline here is necessary else builds with
> CONFIG_HARDENED_USERCOPY will complain because there's no metadata
> visibility into the header object which means __builtin_object_size()
> can't correctly determine the header size.

Oh I see, thanks for pointing to CONFIG_HARDENED_USERCOPY.


Thanks,
Bernd