linux-kernel - Re: [PATCH 1/3] NTFS: aops: Remove VLA usage

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] [day] [month] [year] [list]

Date:   Mon, 9 Jul 2018 18:04:11 -0700
From:   Kees Cook <keescook@...omium.org>
To:     Arnd Bergmann <arnd@...db.de>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Anton Altaparmakov <anton@...era.com>,
        linux-ntfs-dev@...ts.sourceforge.net,
        Al Viro <viro@...iv.linux.org.uk>, Jan Kara <jack@...e.cz>,
        Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 1/3] NTFS: aops: Remove VLA usage

On Thu, Jul 5, 2018 at 1:44 PM, Arnd Bergmann <arnd@...db.de> wrote:
> On Tue, Jun 26, 2018 at 7:29 PM, Kees Cook <keescook@...omium.org> wrote:
>> In the quest to remove all stack VLA usage from the kernel[1], this
>> uses the maximum size needed on the stack and adds a sanity check for
>> robustness: index.block_size cannot be larger than PAGE_SIZE nor less
>> than NTFS_BLOCK_SIZE.
>>
>> [1] https://lkml.kernel.org/r/CA+55aFzCG-zNmZwX4A2FQpadafLfEzK6CC=qPXydAacU1RqZWA@mail.gmail.com
>>
>> Cc: Anton Altaparmakov <anton@...era.com>
>> Cc: linux-ntfs-dev@...ts.sourceforge.net
>> Signed-off-by: Kees Cook <keescook@...omium.org>
>> ---
>>  fs/ntfs/aops.c | 5 ++++-
>>  1 file changed, 4 insertions(+), 1 deletion(-)
>>
>> diff --git a/fs/ntfs/aops.c b/fs/ntfs/aops.c
>> index 3a2e509c77c5..58dadff3e0e0 100644
>> --- a/fs/ntfs/aops.c
>> +++ b/fs/ntfs/aops.c
>> @@ -926,7 +926,7 @@ static int ntfs_write_mst_block(struct page *page,
>>         ntfs_volume *vol = ni->vol;
>>         u8 *kaddr;
>>         unsigned int rec_size = ni->itype.index.block_size;
>> -       ntfs_inode *locked_nis[PAGE_SIZE / rec_size];
>> +       ntfs_inode *locked_nis[PAGE_SIZE / NTFS_BLOCK_SIZE];
>>         struct buffer_head *bh, *head, *tbh, *rec_start_bh;
>>         struct buffer_head *bhs[MAX_BUF_PER_PAGE];
>>         runlist_element *rl;
>
> This has uncovered what looks like a preexisting bug on architectures
> with large page size, this is what I get with 64K pages on arm64:
>
> fs/ntfs/aops.c: In function 'ntfs_write_mst_block':
> fs/ntfs/aops.c:1328:1: error: the frame size of 2432 bytes is larger
> than 2048 bytes [-Werror=frame-larger-than=]
>
> Since both ntfs and 64k pages are fairly obscure features, we might
> get away with just disabling the combination of the two in Kconfig.
>
> Using dynamic allocation might be tricky here, since I assume this
> could be called during writeback in order to free memory, and I can't
> immediately see any better fix.

I'm open to whatever. In crypto, my series uses specifically 4096 for
PAGE_SIZE instead of using PAGE_SIZE, since it wasn't really related.
Here, though, I can't tell if it really IS a page size issue.

-Kees

-- 
Kees Cook
Pixel Security