| 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
| ||
|
Message-ID: <49BAE95D.7060001@gmail.com>
Date: Sat, 14 Mar 2009 00:16:45 +0100
From: Roel Kluin <roel.kluin@...il.com>
To: "David S. Miller" <davem@...emloft.net>, netdev@...r.kernel.org,
lkml <linux-kernel@...r.kernel.org>
Subject: build bug on kfree(struct sk_buff*)
Hi David,
Maybe I should have sent this to you (and netdev) in the first place.
I wrote:
> I was thinking about forcing a build bug if anyone attempts to kfree a
> struct sk_buff*, like this:
>
> #define PTR_OR_BB_ON_TYPE(x, type) (&x[sizeof(char[0 - \
> __builtin_types_compatible_p(typeof(x), type)])])
>
> and:
>
> #define kfree(x) _kfree(PTR_OR_BB_ON_TYPE((x), struct sk_buff*))
>
> with _kfree as the function that is currently kfree.
>
> It appears to work in my test.c.
My question is, is there something wrong with this?
> I also compiled this with (defconfig) and there were breakages, which are
> fixed in the patch below.
I am now trying allyesconfig. The errors are due to:
* In net/core/dev.c +2674 a struct sk_buff* was freed,
* when assigning kfree to a function pointer.
* Several kfrees, but I don't know why gcc complains:
In block/genhd.c:
struct timer_rand_state *random,
In drivers/gpu/drm/i915/intel_modes.c:
struct edid *edid,
In kernel/exit.c:
struct futex_pi_state *pi_state_cache;
In drivers/char/ipmi/ipmi_si_intf.c:
struct si_sm_data *si_sm;
e.g:
/home/roel/dnld/src/kernel/git/linux-2.6/drivers/char/ipmi/ipmi_si_intf.c: In function 'try_smi_init':
/home/roel/dnld/src/kernel/git/linux-2.6/drivers/char/ipmi/ipmi_si_intf.c:2960: error: invalid use of undefined type 'struct si_sm_data'
/home/roel/dnld/src/kernel/git/linux-2.6/drivers/char/ipmi/ipmi_si_intf.c:2960: error: dereferencing pointer to incomplete type
/home/roel/dnld/src/kernel/git/linux-2.6/drivers/char/ipmi/ipmi_si_intf.c: In function 'cleanup_one_si':
/home/roel/dnld/src/kernel/git/linux-2.6/drivers/char/ipmi/ipmi_si_intf.c:3123: error: invalid use of undefined type 'struct si_sm_data'
/home/roel/dnld/src/kernel/git/linux-2.6/drivers/char/ipmi/ipmi_si_intf.c:3123: error: dereferencing pointer to incomplete type
However, allyes wasn't finished yet.
> Also increased sparse warnings, like this:
>
> In file included from /home/roel/dnld/src/kernel/git/linux-2.6/include/linux/vmstat.h:5,
> from /home/roel/dnld/src/kernel/git/linux-2.6/include/linux/mm.h:595,
> from /home/roel/dnld/src/kernel/git/linux-2.6/kernel/exit.c:7:
> /home/roel/dnld/src/kernel/git/linux-2.6/include/linux/percpu.h: In function 'percpu_free':
> /home/roel/dnld/src/kernel/git/linux-2.6/include/linux/percpu.h:98: warning: dereferencing 'void *' pointer
>
> This is mostly due to wrappers for kfree that take a void* argument
But also in cases where the variable is a void*, e.g. net/core/dev.c +4720
below.
I can quite easily change these kfree's to _kfree, however, I am not sure
if that's the best strategy.
Roel
Not meant for inclusion (yet)
---
block/genhd.c | 2 +-
drivers/char/ipmi/ipmi_si_intf.c | 4 ++--
drivers/firmware/dmi-id.c | 2 +-
drivers/gpu/drm/i915/intel_modes.c | 2 +-
drivers/s390/char/vmlogrdr.c | 2 +-
drivers/s390/net/netiucv.c | 2 +-
fs/nfsd/nfs4xdr.c | 4 ++--
include/linux/kernel.h | 5 +++++
include/linux/slab.h | 5 ++++-
kernel/exit.c | 2 +-
lib/kref.c | 2 +-
mm/slab.c | 4 ++--
mm/slob.c | 4 ++--
mm/slub.c | 4 ++--
net/core/dev.c | 4 ++--
security/selinux/ss/services.c | 2 +-
16 files changed, 29 insertions(+), 21 deletions(-)
diff --git a/block/genhd.c b/block/genhd.c
index a9ec910..e732530 100644
--- a/block/genhd.c
+++ b/block/genhd.c
@@ -972,7 +972,7 @@ static void disk_release(struct device *dev)
{
struct gendisk *disk = dev_to_disk(dev);
- kfree(disk->random);
+ _kfree(disk->random);
disk_replace_part_tbl(disk, NULL);
free_part_stats(&disk->part0);
kfree(disk);
diff --git a/drivers/char/ipmi/ipmi_si_intf.c b/drivers/char/ipmi/ipmi_si_intf.c
index 3000135..f86798d 100644
--- a/drivers/char/ipmi/ipmi_si_intf.c
+++ b/drivers/char/ipmi/ipmi_si_intf.c
@@ -2957,7 +2957,7 @@ static int try_smi_init(struct smi_info *new_smi)
if (new_smi->si_sm) {
if (new_smi->handlers)
new_smi->handlers->cleanup(new_smi->si_sm);
- kfree(new_smi->si_sm);
+ _kfree(new_smi->si_sm);
}
if (new_smi->addr_source_cleanup)
new_smi->addr_source_cleanup(new_smi);
@@ -3120,7 +3120,7 @@ static void cleanup_one_si(struct smi_info *to_clean)
to_clean->handlers->cleanup(to_clean->si_sm);
- kfree(to_clean->si_sm);
+ _kfree(to_clean->si_sm);
if (to_clean->addr_source_cleanup)
to_clean->addr_source_cleanup(to_clean);
diff --git a/drivers/firmware/dmi-id.c b/drivers/firmware/dmi-id.c
index 5a76d05..e4517d7 100644
--- a/drivers/firmware/dmi-id.c
+++ b/drivers/firmware/dmi-id.c
@@ -160,7 +160,7 @@ static int dmi_dev_uevent(struct device *dev, struct kobj_uevent_env *env)
static struct class dmi_class = {
.name = "dmi",
- .dev_release = (void(*)(struct device *)) kfree,
+ .dev_release = (void(*)(struct device *)) _kfree,
.dev_uevent = dmi_dev_uevent,
};
diff --git a/drivers/gpu/drm/i915/intel_modes.c b/drivers/gpu/drm/i915/intel_modes.c
index e42019e..0ca35e5 100644
--- a/drivers/gpu/drm/i915/intel_modes.c
+++ b/drivers/gpu/drm/i915/intel_modes.c
@@ -76,7 +76,7 @@ int intel_ddc_get_modes(struct intel_output *intel_output)
drm_mode_connector_update_edid_property(&intel_output->base,
edid);
ret = drm_add_edid_modes(&intel_output->base, edid);
- kfree(edid);
+ _kfree(edid);
}
return ret;
diff --git a/drivers/s390/char/vmlogrdr.c b/drivers/s390/char/vmlogrdr.c
index d8a2289..35471bc 100644
--- a/drivers/s390/char/vmlogrdr.c
+++ b/drivers/s390/char/vmlogrdr.c
@@ -736,7 +736,7 @@ static int vmlogrdr_register_device(struct vmlogrdr_priv_t *priv)
* directly here. (Probably a little bit obfuscating
* but legitime ...).
*/
- dev->release = (void (*)(struct device *))kfree;
+ dev->release = (void (*)(struct device *))_kfree;
} else
return -ENOMEM;
ret = device_register(dev);
diff --git a/drivers/s390/net/netiucv.c b/drivers/s390/net/netiucv.c
index 930e2fc..2ab4e65 100644
--- a/drivers/s390/net/netiucv.c
+++ b/drivers/s390/net/netiucv.c
@@ -1745,7 +1745,7 @@ static int netiucv_register_device(struct net_device *ndev)
* directly here. (Probably a little bit obfuscating
* but legitime ...).
*/
- dev->release = (void (*)(struct device *))kfree;
+ dev->release = (void (*)(struct device *))_kfree;
dev->driver = &netiucv_driver;
} else
return -ENOMEM;
diff --git a/fs/nfsd/nfs4xdr.c b/fs/nfsd/nfs4xdr.c
index f65953b..3c09abd 100644
--- a/fs/nfsd/nfs4xdr.c
+++ b/fs/nfsd/nfs4xdr.c
@@ -215,7 +215,7 @@ static char *savemem(struct nfsd4_compoundargs *argp, __be32 *p, int nbytes)
BUG_ON(p != argp->tmpp);
argp->tmpp = NULL;
}
- if (defer_free(argp, kfree, p)) {
+ if (defer_free(argp, _kfree, p)) {
kfree(p);
return NULL;
} else
@@ -292,7 +292,7 @@ nfsd4_decode_fattr(struct nfsd4_compoundargs *argp, u32 *bmval, struct iattr *ia
host_err = -ENOMEM;
goto out_nfserr;
}
- defer_free(argp, kfree, *acl);
+ defer_free(argp, _kfree, *acl);
(*acl)->naces = nace;
for (ace = (*acl)->aces; ace < (*acl)->aces + nace; ace++) {
diff --git a/include/linux/kernel.h b/include/linux/kernel.h
index 7fa3718..8809c9c 100644
--- a/include/linux/kernel.h
+++ b/include/linux/kernel.h
@@ -526,6 +526,11 @@ struct sysinfo {
aren't permitted). */
#define BUILD_BUG_ON_ZERO(e) (sizeof(char[1 - 2 * !!(e)]) - 1)
+/* If the type of pointer x matches type, force a compilation error,
+ otherwise produce x as a result */
+#define PTR_OR_BB_ON_TYPE(x, type) (&x[sizeof(char[0 - \
+ __builtin_types_compatible_p(typeof(x), type)])])
+
/* Trap pasters of __FUNCTION__ at compile-time */
#define __FUNCTION__ (__func__)
diff --git a/include/linux/slab.h b/include/linux/slab.h
index 24c5602..7a17c3e 100644
--- a/include/linux/slab.h
+++ b/include/linux/slab.h
@@ -121,12 +121,15 @@ int kmem_ptr_validate(struct kmem_cache *cachep, const void *ptr);
#define KMALLOC_MAX_SIZE (1UL << KMALLOC_SHIFT_HIGH)
#define KMALLOC_MAX_ORDER (KMALLOC_SHIFT_HIGH - PAGE_SHIFT)
+/* This ensures that kfree is not called with a struct sk_buff* */
+#define kfree(x) _kfree(PTR_OR_BB_ON_TYPE((x), struct sk_buff*))
+
/*
* Common kmalloc functions provided by all allocators
*/
void * __must_check __krealloc(const void *, size_t, gfp_t);
void * __must_check krealloc(const void *, size_t, gfp_t);
-void kfree(const void *);
+void _kfree(const void *);
void kzfree(const void *);
size_t ksize(const void *);
diff --git a/kernel/exit.c b/kernel/exit.c
index efd30cc..96eb98a 100644
--- a/kernel/exit.c
+++ b/kernel/exit.c
@@ -1103,7 +1103,7 @@ NORET_TYPE void do_exit(long code)
if (unlikely(!list_empty(&tsk->pi_state_list)))
exit_pi_state_list(tsk);
if (unlikely(current->pi_state_cache))
- kfree(current->pi_state_cache);
+ _kfree(current->pi_state_cache);
#endif
/*
* Make sure we are holding no locks:
diff --git a/lib/kref.c b/lib/kref.c
index 9ecd6e8..a6364df 100644
--- a/lib/kref.c
+++ b/lib/kref.c
@@ -62,7 +62,7 @@ void kref_get(struct kref *kref)
int kref_put(struct kref *kref, void (*release)(struct kref *kref))
{
WARN_ON(release == NULL);
- WARN_ON(release == (void (*)(struct kref *))kfree);
+ WARN_ON(release == (void (*)(struct kref *))_kfree);
if (atomic_dec_and_test(&kref->refcount)) {
release(kref);
diff --git a/mm/slab.c b/mm/slab.c
index 4d00855..3acefac 100644
--- a/mm/slab.c
+++ b/mm/slab.c
@@ -3711,7 +3711,7 @@ EXPORT_SYMBOL(kmem_cache_free);
* Don't free memory not originally allocated by kmalloc()
* or you will run into trouble.
*/
-void kfree(const void *objp)
+void _kfree(const void *objp)
{
struct kmem_cache *c;
unsigned long flags;
@@ -3726,7 +3726,7 @@ void kfree(const void *objp)
__cache_free(c, (void *)objp);
local_irq_restore(flags);
}
-EXPORT_SYMBOL(kfree);
+EXPORT_SYMBOL(_kfree);
unsigned int kmem_cache_size(struct kmem_cache *cachep)
{
diff --git a/mm/slob.c b/mm/slob.c
index 52bc8a2..1d8a3c4 100644
--- a/mm/slob.c
+++ b/mm/slob.c
@@ -487,7 +487,7 @@ void *__kmalloc_node(size_t size, gfp_t gfp, int node)
}
EXPORT_SYMBOL(__kmalloc_node);
-void kfree(const void *block)
+void _kfree(const void *block)
{
struct slob_page *sp;
@@ -502,7 +502,7 @@ void kfree(const void *block)
} else
put_page(&sp->page);
}
-EXPORT_SYMBOL(kfree);
+EXPORT_SYMBOL(_kfree);
/* can't use ksize for kmem_cache_alloc memory, only kmalloc */
size_t ksize(const void *block)
diff --git a/mm/slub.c b/mm/slub.c
index 0280eee..f9fa9e1 100644
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -2738,7 +2738,7 @@ size_t ksize(const void *object)
}
EXPORT_SYMBOL(ksize);
-void kfree(const void *x)
+void _kfree(const void *x)
{
struct page *page;
void *object = (void *)x;
@@ -2754,7 +2754,7 @@ void kfree(const void *x)
}
slab_free(page->slab, page, object, _RET_IP_);
}
-EXPORT_SYMBOL(kfree);
+EXPORT_SYMBOL(_kfree);
/*
* kmem_cache_shrink removes empty slabs from the partial lists and sorts
diff --git a/net/core/dev.c b/net/core/dev.c
index f112970..3a7a552 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -2671,7 +2671,7 @@ void netif_napi_del(struct napi_struct *napi)
struct sk_buff *skb, *next;
list_del_init(&napi->dev_list);
- kfree(napi->skb);
+ kfree_skb(napi->skb);
for (skb = napi->gro_list; skb; skb = next) {
next = skb->next;
@@ -4720,7 +4720,7 @@ struct net_device *alloc_netdev_mq(int sizeof_priv, const char *name,
if (!tx) {
printk(KERN_ERR "alloc_netdev: Unable to allocate "
"tx qdiscs.\n");
- kfree(p);
+ _kfree(p);
return NULL;
}
diff --git a/security/selinux/ss/services.c b/security/selinux/ss/services.c
index c65e4fe..cefe043 100644
--- a/security/selinux/ss/services.c
+++ b/security/selinux/ss/services.c
@@ -2862,7 +2862,7 @@ static void security_netlbl_cache_add(struct netlbl_lsm_secattr *secattr,
}
*sid_cache = sid;
- secattr->cache->free = kfree;
+ secattr->cache->free = _kfree;
secattr->cache->data = sid_cache;
secattr->flags |= NETLBL_SECATTR_CACHE;
}
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists