Re: KASAN: slab-out-of-bounds Read in bpf_prog_create

[Eric Biggers]

Arnd and Al,

On Fri, Sep 27, 2019 at 08:15:10PM -0700, Eric Biggers wrote:
> Arnd and Al,
> 
> On Tue, Sep 17, 2019 at 11:49:06AM -0700, syzbot wrote:
> > Hello,
> > 
> > syzbot found the following crash on:
> > 
> > HEAD commit:    2015a28f Add linux-next specific files for 20190915
> > git tree:       linux-next
> > console output: https://syzkaller.appspot.com/x/log.txt?x=11880d69600000
> > kernel config:  https://syzkaller.appspot.com/x/.config?x=110691c2286b679a
> > dashboard link: https://syzkaller.appspot.com/bug?extid=eb853b51b10f1befa0b7
> > compiler:       gcc (GCC) 9.0.0 20181231 (experimental)
> > syz repro:      https://syzkaller.appspot.com/x/repro.syz?x=127c3481600000
> > C reproducer:   https://syzkaller.appspot.com/x/repro.c?x=1150a70d600000
> > 
> > The bug was bisected to:
> > 
> > commit 2f4fa2db75e26995709043c8d3de4632ebed5c4b
> > Author: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
> > Date:   Thu Apr 18 03:48:01 2019 +0000
> > 
> >     compat_ioctl: unify copy-in of ppp filters
> > 
> > bisection log:  https://syzkaller.appspot.com/x/bisect.txt?x=145eee1d600000
> > final crash:    https://syzkaller.appspot.com/x/report.txt?x=165eee1d600000
> > console output: https://syzkaller.appspot.com/x/log.txt?x=125eee1d600000
> > 
> > IMPORTANT: if you fix the bug, please add the following tag to the commit:
> > Reported-by: syzbot+eb853b51b10f1befa0b7@xxxxxxxxxxxxxxxxxxxxxxxxx
> > Fixes: 2f4fa2db75e2 ("compat_ioctl: unify copy-in of ppp filters")
> > 
> > ==================================================================
> > BUG: KASAN: slab-out-of-bounds in memcpy include/linux/string.h:404 [inline]
> > BUG: KASAN: slab-out-of-bounds in bpf_prog_create+0xe9/0x250
> > net/core/filter.c:1351
> > Read of size 32768 at addr ffff88809cf74000 by task syz-executor183/8575
> > 
> > CPU: 0 PID: 8575 Comm: syz-executor183 Not tainted 5.3.0-rc8-next-20190915
> > #0
> > Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS
> > Google 01/01/2011
> > Call Trace:
> >  __dump_stack lib/dump_stack.c:77 [inline]
> >  dump_stack+0x172/0x1f0 lib/dump_stack.c:113
> >  print_address_description.constprop.0.cold+0xd4/0x30b mm/kasan/report.c:374
> >  __kasan_report.cold+0x1b/0x41 mm/kasan/report.c:506
> >  kasan_report+0x12/0x20 mm/kasan/common.c:634
> >  check_memory_region_inline mm/kasan/generic.c:185 [inline]
> >  check_memory_region+0x134/0x1a0 mm/kasan/generic.c:192
> >  memcpy+0x24/0x50 mm/kasan/common.c:122
> >  memcpy include/linux/string.h:404 [inline]
> >  bpf_prog_create+0xe9/0x250 net/core/filter.c:1351
> >  get_filter.isra.0+0x108/0x1a0 drivers/net/ppp/ppp_generic.c:572
> >  ppp_get_filter drivers/net/ppp/ppp_generic.c:584 [inline]
> >  ppp_ioctl+0x129d/0x2590 drivers/net/ppp/ppp_generic.c:801
> 
> This is a correct bisection.  This commit needs:
> 
> diff --git a/drivers/net/ppp/ppp_generic.c b/drivers/net/ppp/ppp_generic.c
> index 267fe2c58087..f55d7937d6c5 100644
> --- a/drivers/net/ppp/ppp_generic.c
> +++ b/drivers/net/ppp/ppp_generic.c
> @@ -564,8 +564,9 @@ static struct bpf_prog *get_filter(struct sock_fprog *uprog)
>  		return NULL;
>  
>  	/* uprog->len is unsigned short, so no overflow here */
> -	fprog.len = uprog->len * sizeof(struct sock_filter);
> -	fprog.filter = memdup_user(uprog->filter, fprog.len);
> +	fprog.len = uprog->len;
> +	fprog.filter = memdup_user(uprog->filter,
> +				   uprog->len * sizeof(struct sock_filter));
>  	if (IS_ERR(fprog.filter))
>  		return ERR_CAST(fprog.filter);
>  

Why did you ignore this and merge the buggy commit to mainline anyway?
I even told you how to fix it...

- Eric