Skip to contentSkip to conversation

nfsd: avoid undefined signed overflow

PATCH In review Jim Rees 3 participants 6 messages May 17, 2013 – May 18, 2013

Subsystems (2)
lore
Share & tools
lore

Jim Rees

author
#0 lore raw

PATCH v2 1/1 nfsd: avoid undefined signed overflow

In C, signed integer overflow results in undefined behavior, but unsigned overflow wraps around. So do the subtraction first, then cast to signed.

---
 fs/nfsd/nfs4state.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/nfsd/nfs4state.c +1 -1
index 316ec84..9850329 100644
@@ -3427,7 +3427,7 @@ grace_disallows_io(struct net *net, struct inode *inode)
 /* Returns true iff a is later than b: */
 static bool stateid_generation_after(stateid_t *a, stateid_t *b)
 {
-	return (s32)a->si_generation - (s32)b->si_generation > 0;
+	return (s32)(a->si_generation - b->si_generation) > 0;
 }
 
 static __be32 check_stateid_generation(stateid_t *in, stateid_t *ref, bool has_session)
Signature · 1 line
1.8.2.3
#1 lore raw
Quoted 2 lines · Jim Rees

In C, signed integer overflow results in undefined behavior, but unsigned overflow wraps around. So do the subtraction first, then cast to signed.

Thanks! Applying for 3.11.

(I wonder if it actually bites anyone in practice? Are there common compilers or architectures where this makes a difference? Even if so I suppose that many generations is probably unlikely enough not to make this worth backporting to stable branches.)

--b.

Quoted 27 lines
---
 fs/nfsd/nfs4state.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/nfsd/nfs4state.c +1 -1
index 316ec84..9850329 100644
@@ -3427,7 +3427,7 @@ grace_disallows_io(struct net *net, struct inode *inode)
 /* Returns true iff a is later than b: */
 static bool stateid_generation_after(stateid_t *a, stateid_t *b)
 {
-	return (s32)a->si_generation - (s32)b->si_generation > 0;
+	return (s32)(a->si_generation - b->si_generation) > 0;
 }
 
 static __be32 check_stateid_generation(stateid_t *in, stateid_t *ref, bool has_session)
Signature · 6 lines
1.8.2.3

--
To unsubscribe from this list: send the line "unsubscribe linux-nfs" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
author
#2 lore raw

Bruce Fields wrote:

Quoted 2 lines · Jim Rees

In C, signed integer overflow results in undefined behavior, but unsigned overflow wraps around. So do the subtraction first, then cast to signed.

Thanks! Applying for 3.11.

(I wonder if it actually bites anyone in practice? Are there common compilers or architectures where this makes a difference? Even if so I suppose that many generations is probably unlikely enough not to make this worth backporting to stable branches.)

I tested on the architectures I have at my fingertips and it made no difference. My guess is you might need a one's complement architecture. Of course we'll never see wraparound in our lifetimes, so no sense backporting this.

I actually used a one's complement machine many years ago, a CDC 6600. I don't think linux has been ported to it.

#3 lore raw
In reply to Jim Rees
Quoted 14 lines · Jim Rees

In C, signed integer overflow results in undefined behavior, but unsigned overflow wraps around. So do the subtraction first, then cast to signed.

---
 fs/nfsd/nfs4state.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/nfsd/nfs4state.c +0 -0
index 316ec84..9850329 100644

fs/nfsd/nfs4state.c grace_disallows_io(struct net *net, struct inode *inode) · line 3431

Quoted 6 lines
fs/nfsd/nfs4state.c +1 -1
@@ -3427,7 +3427,7 @@ grace_disallows_io(struct net *net, struct inode *inode)
 /* Returns true iff a is later than b: */
 static bool stateid_generation_after(stateid_t *a, stateid_t *b)
 {
-	return (s32)a->si_generation - (s32)b->si_generation > 0;
+	return (s32)(a->si_generation - b->si_generation) > 0;
 }

Hmm, what is actually wrong with plain-old
---- snip ----
static bool stateid_generation_after(stateid_t *a, stateid_t *b)

{
	return a->si_generation > b->si_generation;
}

---- snip ----
?

Kind regards,
Bernd

Signature · 2 lines
Bernd Petrovitsch                  Email : [email protected]
                     LUGA : http://www.luga.at
author
#4 lore raw

Bernd Petrovitsch wrote:

  Hmm, what is actually wrong with plain-old
  ----  snip  ----
  static bool stateid_generation_after(stateid_t *a, stateid_t *b)
  {
  	return a->si_generation > b->si_generation;
  }
  ----  snip  ----
  ?

It doesn't account for wraparound. Try this:

    unsigned int a = 0xffff0000, b = 0x7fff0000;
    printf("%d %d\n", a > b, (int)(a - b) > 0);

I just realized that I said this needed a comment, then didn't provide one.

#5 lore raw
Quoted 11 lines · Jim Rees

Bernd Petrovitsch wrote:

  Hmm, what is actually wrong with plain-old
  ----  snip  ----
  static bool stateid_generation_after(stateid_t *a, stateid_t *b)
  {
  	return a->si_generation > b->si_generation;
  }
  ----  snip  ----
  ?
Quoted 6 lines

It doesn't account for wraparound. Try this:

    unsigned int a = 0xffff0000, b = 0x7fff0000;
    printf("%d %d\n", a > b, (int)(a - b) > 0);

I just realized that I said this needed a comment, then didn't provide one.

Ooops, yes, thx.
Sry for the noise;-)

Bernd

Signature · 2 lines
Bernd Petrovitsch                  Email : [email protected]
                     LUGA : http://www.luga.at