linux-kernel - Re: [PATCH v2] fs: dax: Adding new return type vm_fault

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: <20180421201711.GE14610@bombadil.infradead.org>
Date:   Sat, 21 Apr 2018 13:17:11 -0700
From:   Matthew Wilcox <willy@...radead.org>
To:     Souptick Joarder <jrdr.linux@...il.com>
Cc:     mawilcox@...rosoft.com, ross.zwisler@...ux.intel.com,
        viro@...iv.linux.org.uk, linux-fsdevel@...r.kernel.org,
        linux-kernel@...r.kernel.org, linux-mm@...ck.org
Subject: Re: [PATCH v2] fs: dax: Adding new return type vm_fault_t

On Sat, Apr 21, 2018 at 10:44:42PM +0530, Souptick Joarder wrote:
> @@ -1112,7 +1112,7 @@ int __dax_zero_page_range(struct block_device *bdev,
>  }
>  EXPORT_SYMBOL_GPL(dax_iomap_rw);
>  
> -static int dax_fault_return(int error)
> +static vm_fault_t dax_fault_return(int error)
>  {
>  	if (error == 0)
>  		return VM_FAULT_NOPAGE;

At some point, we'll want to get rid of dax_fault_return, but that can be
a follow-on patch after vmf_error is in.

>  		if (write)
> -			error = vm_insert_mixed_mkwrite(vma, vaddr, pfn);
> +			ret = vmf_insert_mixed_mkwrite(vma, vaddr, pfn);
>  		else
> -			error = vm_insert_mixed(vma, vaddr, pfn);
> +			ret = vmf_insert_mixed(vma, vaddr, pfn);
>  
> -		/* -EBUSY is fine, somebody else faulted on the same PTE */
> -		if (error == -EBUSY)
> -			error = 0;
> -		break;
> +		goto finish_iomap;

> @@ -1284,12 +1281,12 @@ static int dax_iomap_pte_fault(struct vm_fault *vmf, pfn_t *pfnp,
>  	}
>  
>   error_finish_iomap:
> -	vmf_ret = dax_fault_return(error) | major;
> +	ret = dax_fault_return(error) | major;
>   finish_iomap:

I think we lose VM_FAULT_MAJOR with this change.

I would suggest fixing this with ...

  error_finish_iomap:
-	vmf_ret = dax_fault_return(error) | major;
+	ret = dax_fault_return(error);
  finish_iomap:

[...]

  out:
-	trace_dax_pte_fault_done(inode, vmf, vmf_ret);
-	return vmf_ret;
+	trace_dax_pte_fault_done(inode, vmf, ret);
+	return ret | major;