linux-kernel - Re: [PATCH] mm: pgtable: Ensure pml spinlock gets unlock

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <b70f663f-5b72-41dd-bae4-14808a87f262@bytedance.com>
Date: Mon, 10 Feb 2025 12:05:05 +0800
From: Qi Zheng <zhengqi.arch@...edance.com>
To: I Hsin Cheng <richard120310@...il.com>
Cc: akpm@...ux-foundation.org, linux-mm@...ck.org,
 linux-kernel@...r.kernel.org
Subject: Re: [PATCH] mm: pgtable: Ensure pml spinlock gets unlock



On 2025/2/9 02:49, I Hsin Cheng wrote:
> When !start_pte is true, the "pml" spinlock is still being holded and
> the branch "out_pte" is taken. If "ptl" is equal to "pml", the lock
> "pml" will still be locked when the function returns.

No. When start_pte is NULL, the ptl must also be NULL, so the ptl and
pml will not be equal.

> 
> It'll be better to set a new branch "out_pte" and jump to it when
> !start_pte is true at the first place, therefore no additional check for
> "start_pte" or "ptl != pml" is needed, simply unlock "pml" and return.
> 
> Signed-off-by: I Hsin Cheng <richard120310@...il.com>
> ---
>   mm/pt_reclaim.c | 6 +++++-
>   1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/mm/pt_reclaim.c b/mm/pt_reclaim.c
> index 7e9455a18aae..163e38f1728d 100644
> --- a/mm/pt_reclaim.c
> +++ b/mm/pt_reclaim.c
> @@ -43,7 +43,7 @@ void try_to_free_pte(struct mm_struct *mm, pmd_t *pmd, unsigned long addr,
>   	pml = pmd_lock(mm, pmd);
>   	start_pte = pte_offset_map_rw_nolock(mm, pmd, addr, &pmdval, &ptl);
>   	if (!start_pte)
> -		goto out_ptl;
> +		goto out_pte;
>   	if (ptl != pml)
>   		spin_lock_nested(ptl, SINGLE_DEPTH_NESTING);
>   
> @@ -68,4 +68,8 @@ void try_to_free_pte(struct mm_struct *mm, pmd_t *pmd, unsigned long addr,
>   		pte_unmap_unlock(start_pte, ptl);
>   	if (ptl != pml)
>   		spin_unlock(pml);
> +	return;
> +
> +out_pte:
> +	spin_unlock(pml);
>   }