Re: [PATCH RFT 0/3] mm, drm: ensure .fault() does not have to be followed by .pfn_mkwrite() for write faults

[Paolo Bonzini]

On 8/3/26 10:55, David Hildenbrand (Arm) wrote:
> On 7/31/26 18:05, Paolo Bonzini wrote:
> > Reported-by: Sergio Lopez <slp@xxxxxxxxxx>
>
> Reported-by: without Fixes: is odd.

Fixes: 6da8e9634bb7 ("mm: new follow_pfnmap API") would also be odd :) 
but I can certainly add it.
> > +	 * @write_fault: if true, fail with -EFAULT unless the mapping is
>
> Just wondering whether EPERM would be better.

It would be EACCES if anything, not EPERM; but almost all callers 
already pass EFAULT to userspace, and write() to a PROT_READ area 
returns EFAULT, so I don't think EACCES is the right choice.

> > +	 * writable
> >  	 */
> >  	struct vm_area_struct *vma;
> >  	unsigned long address;
> > +	bool write_fault;
>
> "write_fault" is a rather odd name for this, given that this function will not
> trigger a write fault.
>
> You want something that matches FOLL_WRITE.
>
> "write_access" / "check_writable" maybe?
There are no for_write, write_access or check_write in mm/,  but there 
are a handful of each of these

	int write = (gup_flags & FOLL_WRITE);
	bool write = vmf->flags & FAULT_FLAG_WRITE;

so I'll go for just "write".

Thanks,

Paolo