Date: Mon, 14 Mar 2016 12:25:39 +0900 (JST)
Message-Id: <20160314.122539.338096075.d.hatayama@jp.fujitsu.com>
To: dyoung@redhat.com
Cc: akpm@linux-foundation.org, linux-kernel@vger.kernel.org, bhe@redhat.com,
        vgoyal@redhat.com, kexec@lists.infradead.org, nasa4836@gmail.com,
        mhuang@redhat.com
Subject: Re: [PATCH V2] proc-vmcore: wrong data type casting fix
From: HATAYAMA Daisuke <d.hatayama@jp.fujitsu.com>
In-Reply-To: <20160311084248.GA9417@dhcp-128-65.nay.redhat.com>
References: <20160311084248.GA9417@dhcp-128-65.nay.redhat.com>
Mime-Version: 1.0
Content-Type: Text/Plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 2889
Lines: 82

From: Dave Young <dyoung@redhat.com>
Subject: [PATCH V2] proc-vmcore: wrong data type casting fix
Date: Fri, 11 Mar 2016 16:42:48 +0800

> On i686 PAE enabled machine the contiguous physical area could be large
> and it can cause trimming down variables in below calculation in
> read_vmcore() and mmap_vmcore():
> 
> 	tsz = min_t(size_t, m->offset + m->size - *fpos, buflen);
> 
> Then the real size passed down is not correct any more.
> Suppose m->offset + m->size - *fpos being truncated to 0, buflen >0 then

That is, size_t and loff_t are defined as follows on i686:

    (gdb) ptype size_t
    type = unsigned int
    (gdb) ptype loff_t
    type = long long int

So casting by size_t means truncating a given value by 4GB.

Then, if (m->offset + m->size - *fpos) is equal to or larger than 4GB,
and is aligned with 4GB, the resulted value is 0, and

> we will get tsz = 0. It is of course not an expected result.
> 
> During our tests there are two problems caused by it:
> 1) read_vmcore will refuse to continue so makedumpfile fails.
> 2) mmap_vmcore will trigger BUG_ON() in remap_pfn_range().
>

we reach these errors.

If (m->offset + m->size - *fpos) is not aligned with 4GB,
read_vmcore() or mmap_vmcore() is performed with the truncated
non-zero value as size (of course, this is also not expected value but
the execution doesn't result in error). Then, fpos proceeds so that
(m->offset + m->size - *fpos) is aligned with 4GB in the next loop,
and we after all reach the errors.

I think your patch description needs a bit more detail.

It seems good to me that the patch itself.

> Use unsigned long long in min_t instead so that the variables are not
> truncated.
> 
> Signed-off-by: Baoquan He <bhe@redhat.com>
> Signed-off-by: Dave Young <dyoung@redhat.com>
> ---
> v1->v2: spelling fix in patch log
>  fs/proc/vmcore.c |    7 +++++--
>  1 file changed, 5 insertions(+), 2 deletions(-)
> 
> --- linux-x86.orig/fs/proc/vmcore.c
> +++ linux-x86/fs/proc/vmcore.c
> @@ -231,7 +231,9 @@ static ssize_t __read_vmcore(char *buffe
>  
>  	list_for_each_entry(m, &vmcore_list, list) {
>  		if (*fpos < m->offset + m->size) {
> -			tsz = min_t(size_t, m->offset + m->size - *fpos, buflen);
> +			tsz = (size_t)min_t(unsigned long long,
> +					    m->offset + m->size - *fpos,
> +					    buflen);
>  			start = m->paddr + *fpos - m->offset;
>  			tmp = read_from_oldmem(buffer, tsz, &start, userbuf);
>  			if (tmp < 0)
> @@ -461,7 +463,8 @@ static int mmap_vmcore(struct file *file
>  		if (start < m->offset + m->size) {
>  			u64 paddr = 0;
>  
> -			tsz = min_t(size_t, m->offset + m->size - start, size);
> +			tsz = (size_t)min_t(unsigned long long,
> +					    m->offset + m->size - start, size);
>  			paddr = m->paddr + start - m->offset;
>  			if (vmcore_remap_oldmem_pfn(vma, vma->vm_start + len,
>  						    paddr >> PAGE_SHIFT, tsz,
> 
--
Thanks.
HATAYAMA, Daisuke