RE: [PATCH] Bluetooth: L2CAP: Elide a string overflow warning

["Elliott, Robert (Servers)" <elliott@hpe.com>]

> -----Original Message-----
> From: Siddh Raman Pant <code@siddh.me>
> Sent: Thursday, August 25, 2022 6:01 AM
> To: palmer@rivosinc.com
> Cc: davem@davemloft.net; edumazet@google.com; johan.hedberg@gmail.com;
> kuba@kernel.org; linux-bluetooth@vger.kernel.org; linux-
> kernel@vger.kernel.org; linux@rivosinc.com; luiz.dentz@gmail.com;
> marcel@holtmann.org; netdev@vger.kernel.org; pabeni@redhat.com
> Subject: Re: [PATCH] Bluetooth: L2CAP: Elide a string overflow warning
> 
> On Fri, 12 Aug 2022 11:22:49 +0530  Palmer Dabbelt  wrote:
> > From: Palmer Dabbelt <palmer@rivosinc.com>
> >
> > Without this I get a string op warning related to copying from a
> > possibly NULL pointer.  I think the warning is spurious, but it's
> > tripping up allmodconfig.
> 
> I think it is not spurious, and is due to the following commit:
> d0be8347c623 ("Bluetooth: L2CAP: Fix use-after-free caused by l2cap_chan_put")

That commit was OK - it added an "if (!c) continue" to handle if
the value c is changed to NULL.
 
> The following commit fixes a similar problem (added the NULL check on line
> 1996):
> 332f1795ca20 ("Bluetooth: L2CAP: Fix l2cap_global_chan_by_psm regression")

That commit wiped out the "if (!c) continue" path escape clause
from the previous patch, introducing a path back to code that
doesn't check for NULL:

-                               if (!c)
-                                       continue;
-
-                               read_unlock(&chan_list_lock);
-                               return c;
+                               if (c) {
+                                       read_unlock(&chan_list_lock);
+                                       return c;
+                               }
                        }
                        src_any = !bacmp(&c->src, BDADDR_ANY);
                        dst_any = !bacmp(&c->dst, BDADDR_ANY);
 

> >     inlined from 'l2cap_global_chan_by_psm' at /scratch/merges/ko-linux-
> next/linux/net/bluetooth/l2cap_core.c:2003:15:
> > /scratch/merges/ko-linux-next/linux/include/linux/fortify-string.h:44:33:
> error: '__builtin_memcmp' specified bound 6 exceeds source size 0 [-
> Werror=stringop-overread]
> >    44 | #define __underlying_memcmp     __builtin_memcmp
> >       |                                 ^
> > /scratch/merges/ko-linux-next/linux/include/linux/fortify-string.h:420:16:
> note: in expansion of macro '__underlying_memcmp'
> >   420 |         return __underlying_memcmp(p, q, size);
> >       |                ^~~~~~~~~~~~~~~~~~~
> > In function 'memcmp',
> >     inlined from 'bacmp' at /scratch/merges/ko-linux-
> next/linux/include/net/bluetooth/bluetooth.h:347:9,
> >     inlined from 'l2cap_global_chan_by_psm' at /scratch/merges/ko-linux-
...
> >
> > Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
> 
> Tested-by: Siddh Raman Pant <code@siddh.me>

This patch is necessary to get a successful cross-compile of 6.0-rc3 
allmodconfig for ARCH=mips on x86.

Tested-by: Robert Elliott <elliott@hpe.com>

I suggest you label this patch as:
Fixes: 332f1795ca20 ("Bluetooth: L2CAP: Fix l2cap_global_chan_by_psm regression")

> > ---
> >  net/bluetooth/l2cap_core.c | 12 +++++++-----
> >  1 file changed, 7 insertions(+), 5 deletions(-)
> >
> > diff --git a/net/bluetooth/l2cap_core.c b/net/bluetooth/l2cap_core.c
> > index cbe0cae73434..be7f47e52119 100644
> > --- a/net/bluetooth/l2cap_core.c
> > +++ b/net/bluetooth/l2cap_core.c
> > @@ -2000,11 +2000,13 @@ static struct l2cap_chan
> *l2cap_global_chan_by_psm(int state, __le16 psm,
> >  			}
> >
> >  			/* Closest match */
> > -			src_any = !bacmp(&c->src, BDADDR_ANY);
> > -			dst_any = !bacmp(&c->dst, BDADDR_ANY);
> > -			if ((src_match && dst_any) || (src_any && dst_match) ||
> > -			    (src_any && dst_any))
> > -				c1 = c;
> > +			if (c) {
> > +				src_any = !bacmp(&c->src, BDADDR_ANY);
> > +				dst_any = !bacmp(&c->dst, BDADDR_ANY);
> > +				if ((src_match && dst_any) || (src_any &&
> dst_match) ||
> > +				    (src_any && dst_any))
> > +					c1 = c;
> > +			}
> >  		}
> >  	}
> >
>