: Fix alignment checks in ebt_among.ko module.
Subsystems (4)
PATCH 1/1 : Fix alignment checks in ebt_among.ko module.
I've sent this patch some days ago to Bart, but with no answer...
When trying to do
# ebtables -A FORWARD --among-src 0:12:34:56:78:9a=192.168.0.10 -j ACCEPT
on x86_64 box the ebt_among->check() callback warns me that
ebtables: among: wrong size: 1060 against expected 1056, rounded to 1056
Checking the ebtables sources, I found that the alignment is done differently in the tool and the kernel. Tool makes it like this:
EBT_ALIGN(sizeof(struct ebt_among_info)) + X
while the kernel module like this:
EBT_ALIGN(sizeof(struct ebt_among_info) + X)
So the suggested fix is to move the alignment in the kernel. After the fix the rule is added and appears in the ebtables -L output.
Originally developed by Evgeny Kravtsunov.
---
net/bridge/netfilter/ebt_among.c +2 -2
@@ -182,7 +182,7 @@ static int ebt_among_check(const char *tablename, unsigned int hookmask, unsigned int datalen) {const struct ebt_among_info *info = data; -int expected_length =sizeof (struct ebt_among_info ); +int expected_length =EBT_ALIGN (sizeof (struct ebt_among_info ));const struct ebt_mac_wormhash *wh_dst, *wh_src;int err;
@@ -191,7 +191,7 @@ static int ebt_among_check(const char *tablename, unsigned int hookmask, expected_length +=ebt_mac_wormhash_size (wh_dst); expected_length +=ebt_mac_wormhash_size (wh_src); -if (datalen !=EBT_ALIGN (expected_length)) { +if (datalen != expected_length) { printk(KERN_WARNING "ebtables: among: wrong size: %d " "against expected %d, rounded to %Zd\n",
Quoted 13 lines · Pavel Emelyanov
I've sent this patch some days ago to Bart, but with no answer...
When trying to do
# ebtables -A FORWARD --among-src 0:12:34:56:78:9a=192.168.0.10 -j ACCEPTon
x86_64box theebt_among->check()callback warns me thatebtables: among: wrong size: 1060 against expected 1056, rounded to 1056
Checking the ebtables sources, I found that the alignment is done differently in the tool and the kernel. Tool makes it like this:
Quoted 8 lines
EBT_ALIGN(sizeof(struct ebt_among_info))+ Xwhile the kernel module like this:
EBT_ALIGN(sizeof(struct ebt_among_info) + X)So the suggested fix is to move the alignment in the kernel. After the fix the rule is added and appears in the ebtables -L output.
It seems the kernel is correct and userspace is doing it wrong, so I think userspace should be fixed instead. The problem with your patch is that is causes misalignment for following structures that contain u64 members.
Quoted 9 lines · Patrick McHardy
Quoted 8 lines · Pavel Emelyanov
I've sent this patch some days ago to Bart, but with no answer...
When trying to do
# ebtables -A FORWARD --among-src 0:12:34:56:78:9a=192.168.0.10 -j ACCEPTon
x86_64box theebt_among->check()callback warns me that
Quoted 7 lines
Quoted 2 lines
ebtables: among: wrong size: 1060 against expected 1056, rounded to 1056
[...]It seems the kernel is correct and userspace is doing it wrong, so I think userspace should be fixed instead. The problem with your patch is that is causes misalignment for following structures that contain u64 members.
ebt_among is a true mess! It hinders my efforts to port ebtables to xtables, but I could of course use a hack and will likely do just that.
Op vr, 29-02-2008 te 16:40 +0100, schreef Jan Engelhardt:
Quoted 8 lines
Quoted 4 lines
It seems the kernel is correct and userspace is doing it wrong, so I think userspace should be fixed instead. The problem with your patch is that is causes misalignment for following structures that contain u64 members.
ebt_amongis a true mess! It hinders my efforts to port ebtables to xtables, but I could of course use a hack and will likely do just that.
Ebtables' dynamic memory allocation scheme is probably not portable to iptables. I agree the fix is in userspace. Anyway, I think the author of that module did a great job and provided a very useful and often-used functionality.
Bart
Quoted 6 lines · Bart De Schuymer
Op vr, 29-02-2008 te 16:40 +0100, schreef Jan Engelhardt:
Quoted 5 lines
Quoted 4 lines
It seems the kernel is correct and userspace is doing it wrong, so I think userspace should be fixed instead. The problem with your patch is that is causes misalignment for following structures that contain u64 members.
Quoted 8 lines
Quoted 3 lines
ebt_amongis a true mess! It hinders my efforts to port ebtables to xtables, but I could of course use a hack and will likely do just that.Ebtables' dynamic memory allocation scheme is probably not portable to iptables. I agree the fix is in userspace. Anyway, I think the author of that module did a great job and provided a very useful and often-used functionality.
Actually yes, it's kind of a weird mixup of TLV and static blocks :p
Op vr, 29-02-2008 te 13:48 +0100, schreef Patrick McHardy:
Quoted 2 lines
Quoted 1 line · Pavel Emelyanov
I've sent this patch some days ago to Bart, but with no answer...
Sorry about that.
Quoted 9 lines
Quoted 9 lines
When trying to do
# ebtables -A FORWARD --among-src 0:12:34:56:78:9a=192.168.0.10 -j ACCEPTon
x86_64box theebt_among->check()callback warns me thatebtables: among: wrong size: 1060 against expected 1056, rounded to 1056
Quoted 8 lines
Quoted 2 lines
So the suggested fix is to move the alignment in the kernel. After the fix the rule is added and appears in the ebtables -L output.
It seems the kernel is correct and userspace is doing it wrong, so I think userspace should be fixed instead. The problem with your patch is that is causes misalignment for following structures that contain u64 members.
Pavel, please try the attached userspace patch to see if it fixes the problem (and let me know the result).
cheers,
Bart
extensions/ebt_among.c renamed from extensions/ebt_among.c.old +12 -8
@@ -68,6 +68,7 @@ static void print_help()" 00:00:00:fa:eb:fe=153.19.120.250,00:00:00:fa:eb:fe=192.168.0.1\n" ); } +static int old_size;static void init (struct ebt_entry_match *match) {
@@ -75,6 +76,7 @@ static void init(struct ebt_entry_match (struct ebt_among_info *) match->data;memset (amonginfo,0 ,sizeof (struct ebt_among_info )); + old_size =sizeof (struct ebt_among_info ); }static struct ebt_mac_wormhash *new_wormhash (int n)
@@ -308,7 +310,7 @@ static int parse(int c, char **argv, int (struct ebt_among_info *) (*match)->data;struct ebt_mac_wormhash *wh;struct ebt_entry_match *h; -int new_size, old_size; +int new_size;long flen;int fd;
@@ -354,21 +356,23 @@ static int parse(int c, char **argv, intif (ebt_errormsg[0 ] !='\0' )break ; - old_size =sizeof (struct ebt_entry_match ) + (**match).match_size; - h =malloc ((new_size = old_size +ebt_mac_wormhash_size (wh))); + new_size = old_size+ebt_mac_wormhash_size (wh); + h =malloc (sizeof (struct ebt_entry_match )+EBT_ALIGN (new_size));if (!h)ebt_print_memory (); -memcpy (h, *match, old_size); -memcpy ((char *) h + old_size, wh,ebt_mac_wormhash_size (wh)); - h->match_size = new_size -sizeof (struct ebt_entry_match ); +memcpy (h, *match, old_size+sizeof (struct ebt_entry_match )); +memcpy ((char *)h+old_size+sizeof (struct ebt_entry_match ), wh, +ebt_mac_wormhash_size (wh)); + h->match_size =EBT_ALIGN (new_size); info = (struct ebt_among_info *) h->data;if (c ==AMONG_DST ) { info->wh_dst_ofs = - old_size -sizeof (struct ebt_entry_match ); + old_size; }else { info->wh_src_ofs = - old_size -sizeof (struct ebt_entry_match ); + old_size; } + old_size = new_size;free (*match); *match = h;free (wh);
Quoted 17 lines · Bart De Schuymer
Op vr, 29-02-2008 te 13:48 +0100, schreef Patrick McHardy:
Quoted 2 lines
Quoted 1 line · Pavel Emelyanov
I've sent this patch some days ago to Bart, but with no answer...
Sorry about that.
Quoted 11 lines
Quoted 10 lines
When trying to do
# ebtables -A FORWARD --among-src 0:12:34:56:78:9a=192.168.0.10 -j ACCEPTon
x86_64box theebt_among->check()callback warns me thatebtables: among: wrong size: 1060 against expected 1056, rounded to 1056
So the suggested fix is to move the alignment in the kernel. After the fix the rule is added and appears in the ebtables -L output.
Quoted 7 lines
Quoted 4 lines
It seems the kernel is correct and userspace is doing it wrong, so I think userspace should be fixed instead. The problem with your patch is that is causes misalignment for following structures that contain u64 members.
Pavel, please try the attached userspace patch to see if it fixes the problem (and let me know the result).
Tested - works OK. Thanks.
Quoted 4 lines
cheers,
Bart