Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Wed, 11 May 2022 12:50:20 -0700
From:   Sultan Alsawaf <sultan@kerneltoast.com>
To:     Minchan Kim <minchan@kernel.org>
Cc:     stable@vger.kernel.org, Nitin Gupta <ngupta@vflare.org>,
        Sergey Senozhatsky <senozhatsky@chromium.org>,
        Andrew Morton <akpm@linux-foundation.org>, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
Subject: Re: [PATCH] zsmalloc: Fix races between asynchronous zspage free and
 page migration
Message-ID: <YnwTfBLn+6vYSe59@sultan-box.localdomain>
References: <20220509024703.243847-1-sultan@kerneltoast.com>
 <Ynv53fkx8cG0ixaE@google.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Ynv53fkx8cG0ixaE@google.com>
Precedence: bulk

On Wed, May 11, 2022 at 11:01:01AM -0700, Minchan Kim wrote:
> On Sun, May 08, 2022 at 07:47:02PM -0700, Sultan Alsawaf wrote:
> > Cc: stable@vger.kernel.org
> > Fixes: 48b4800a1c6a ("zsmalloc: page migration support")
> 
> Shouldn't the fix be Fixes: 77ff465799c6 ("zsmalloc: zs_page_migrate: skip
> unnecessary loops but not return -EBUSY if zspage is not inuse)?
> Because we didn't migrate ZS_EMPTY pages before.

Hi,

Yeah, 77ff465799c6 indeed seems like the commit that introduced the bug.

> I couldn't get the point here. Why couldn't we simple lock zspage migration?
> 
> diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c
> index 9152fbde33b5..05ff2315b7b1 100644
> --- a/mm/zsmalloc.c
> +++ b/mm/zsmalloc.c
> @@ -1987,7 +1987,10 @@ static void async_free_zspage(struct work_struct *work)
>  
>         list_for_each_entry_safe(zspage, tmp, &free_pages, list) {
>                 list_del(&zspage->list);
> +
> +               migrate_read_lock(zspage);
>                 lock_zspage(zspage);
> +               migrate_read_unlock(zspage);
>  
>                 get_zspage_mapping(zspage, &class_idx, &fullness);
>                 VM_BUG_ON(fullness != ZS_EMPTY);

There are two problems with this:
1. migrate_read_lock() is a rwlock and lock_page() can sleep.
2. This will cause a deadlock because it violates the lock ordering in
   zs_page_migrate(), since zs_page_migrate() takes migrate_write_lock() under
   the page lock.

Sultan