Received-SPF: pass (google.com: domain of linux-kernel+bounces-74037-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Date: Wed, 21 Feb 2024 03:46:04 +0000
From: Matthew Wilcox <willy@infradead.org>
To: "Vishal Moola (Oracle)" <vishal.moola@gmail.com>
Cc: linux-mm@kvack.org, linux-kernel@vger.kernel.org,
	akpm@linux-foundation.org, muchun.song@linux.dev
Subject: Re: [PATCH 2/3] hugetlb: Use vmf_anon_prepare() instead of
 anon_vma_prepare()
Message-ID: <ZdVx_Jr9tm2lV0Ku@casper.infradead.org>
References: <20240220231424.126600-1-vishal.moola@gmail.com>
 <20240220231424.126600-3-vishal.moola@gmail.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20240220231424.126600-3-vishal.moola@gmail.com>

On Tue, Feb 20, 2024 at 03:14:23PM -0800, Vishal Moola (Oracle) wrote:
> +++ b/mm/hugetlb.c
> @@ -5834,9 +5834,15 @@ static vm_fault_t hugetlb_wp(struct mm_struct *mm, struct vm_area_struct *vma,
>  	struct folio *old_folio;
>  	struct folio *new_folio;
>  	int outside_reserve = 0;
> -	vm_fault_t ret = 0;
> +	vm_fault_t ret = 0, anon_ret = 0;

Do we need a second variable here?  Seems to me like we could
unconditionally assign to ret:

> -	if (unlikely(anon_vma_prepare(vma))) {
> -		ret = VM_FAULT_OOM;
> +	anon_ret = vmf_anon_prepare(&vmf);
> +	if (unlikely(anon_ret)) {
> +		ret = anon_ret;


>  	unsigned long haddr = address & huge_page_mask(h);
>  	struct mmu_notifier_range range;
> +	struct vm_fault vmf = {
> +				.vma = vma,
> +				.address = haddr,
> +				.real_address = address,
> +				.flags = flags,
> +	};

We don't usually indent quite so far.  One extra tab would be enough.

Also, I thought we talked about creating the vmf in hugetlb_fault(),
then passing it to hugetlb_wp() hugetlb_no_page() and handle_userfault()?
Was there a reason to abandon that idea?