slip: fix slab-out-of-bounds write in slhc_uncompress()
Subsystems (2)
PATCH 1/1 slip: fix slab-out-of-bounds write in slhc_uncompress()
sl_bump() reserves only 80 bytes of expansion headroom before calling slhc_uncompress(), but the reconstructed IP + TCP header is up to ip->ihl*4 + thp->doff*4 bytes. IHL and TCP doff are 4-bit fields and both can legitimately reach 15, so the header can grow to 2*15*4 = 120 bytes. A VJ-uncompressed primer with ihl=15, doff=15 followed by a compressed frame of size buffsize - 80 therefore writes up to 33 bytes past the kmalloc(buffsize + 4) rbuff allocation, with attacker-controlled content:
BUG: KASAN: slab-out-of-bounds in slhc_uncompress Write of size 1069 at addr ffff88800ba93078 by task kworker/u8:1/32 Workqueue: events_unbound flush_to_ldisc Call Trace: __asan_memmove+0x3f/0x70 slhc_uncompress (drivers/net/slip/slhc.c:614) slip_receive_buf (drivers/net/slip/slip.c:342) tty_ldisc_receive_buf flush_to_ldisc
Raise the reservation to match the real worst case. The ppp_generic receive path already enforces skb_tailroom >= 124 and is unaffected.
--- drivers/net/slip/slip.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-)
drivers/net/slip/slip.c +7 -3
@@ -333,9 +333,13 @@ static void sl_bump(struct slip *sl)printk (KERN_WARNING "%s: compressed packet ignored\n" , dev->name);return ; } -/* make sure we've reserved enough space for uncompress -to use */ -if (count +80 > sl->buffsize) { +/* slhc_uncompress() prepends up to +* ip->ihl * 4 + thp->doff * 4 bytes of reconstructed +* IPv4 + TCP header. IHL and doff are 4-bit fields +* (max 15) counting 4-byte units, so the header is +* at most 2 * 15 * 4 = 120 bytes. +*/ +if (count +2 *15 *4 > sl->buffsize) { dev->stats.rx_over_errors++;return ; }
Signature · 1 line
2.43.0
Quoted 19 lines · Weiming Shi
sl_bump()reserves only 80 bytes of expansion headroom before callingslhc_uncompress(), but the reconstructed IP + TCP header is up to ip->ihl*4 + thp->doff*4 bytes. IHL and TCP doff are 4-bit fields and both can legitimately reach 15, so the header can grow to 2*15*4 = 120 bytes. A VJ-uncompressed primer with ihl=15, doff=15 followed by a compressed frame of size buffsize - 80 therefore writes up to 33 bytes past the kmalloc(buffsize + 4) rbuff allocation, with attacker-controlled content:BUG: KASAN: slab-out-of-bounds in slhc_uncompress Write of size 1069 at addr ffff88800ba93078 by task kworker/u8:1/32 Workqueue: events_unbound flush_to_ldisc Call Trace: __asan_memmove+0x3f/0x70 slhc_uncompress (drivers/net/slip/slhc.c:614) slip_receive_buf (drivers/net/slip/slip.c:342) tty_ldisc_receive_buf flush_to_ldisc
Quoted 5 lines
Raise the reservation to match the real worst case. The
ppp_genericreceive path already enforcesskb_tailroom>= 124 and is unaffected.Fixes b5451d783ade ("slip: Move the SLIP drivers")Reported-by Simon Horman <[email protected]>
FTR, I was mainly passing on a review generated by Sashiko
Quoted 1 line
Signed-off-by Weiming Shi <[email protected]>
As usual I'll comment on the review of this patch by Sashiko.
TL;DR: I don't think it should block progress of this patch.
The review by Sashiko flags out of bounds errors. However, these are addressed by one of your other patches:
- [PATCH net] slip: bound decode() reads against the compressed packet length
https://lore.kernel.org/netdev/[email protected]/
As noted in my review of that patch, while it seems too late for these patches, please consider bundling related patches in a patchset in future.
Quoted 19 lines · Weiming Shi
sl_bump()reserves only 80 bytes of expansion headroom before callingslhc_uncompress(), but the reconstructed IP + TCP header is up to ip->ihl*4 + thp->doff*4 bytes. IHL and TCP doff are 4-bit fields and both can legitimately reach 15, so the header can grow to 2*15*4 = 120 bytes. A VJ-uncompressed primer with ihl=15, doff=15 followed by a compressed frame of size buffsize - 80 therefore writes up to 33 bytes past the kmalloc(buffsize + 4) rbuff allocation, with attacker-controlled content:BUG: KASAN: slab-out-of-bounds in slhc_uncompress Write of size 1069 at addr ffff88800ba93078 by task kworker/u8:1/32 Workqueue: events_unbound flush_to_ldisc Call Trace: __asan_memmove+0x3f/0x70 slhc_uncompress (drivers/net/slip/slhc.c:614) slip_receive_buf (drivers/net/slip/slip.c:342) tty_ldisc_receive_buf flush_to_ldisc
Quoted 5 lines
Raise the reservation to match the real worst case. The
ppp_genericreceive path already enforcesskb_tailroom>= 124 and is unaffected.Fixes b5451d783ade ("slip: Move the SLIP drivers")Reported-by Simon Horman <[email protected]>
FTR, I was mainly passing on information flagged by Sashiko.
Quoted 1 line
Signed-off-by Weiming Shi <[email protected]>
Let me summarise my understanding of Sashiko's review of this patch.
TL;DR: I don't think that review should block progress of this patch.
1. The issue wrt concurrent MTU changes appears to be a separate, pre-existing problem. Maybe you've looked into it already, if not, you may wish to.
2. The bounds checking problems are addressed by other patches in flight.
- [PATCH net v2] slip: reject VJ receive packets on instances with no rstate array
https://lore.kernel.org/netdev/[email protected]/
- [PATCH net] slip: fix slab-out-of-bounds write in slhc_uncompress()
https://lore.kernel.org/netdev/[email protected]/
In future you might want to consider creating patch sets for related patches. But I think it's too late in the case of these patches.
...
Simon Horman
reviewerQuoted 23 lines · Simon Horman
Quoted 22 lines · Weiming Shi
sl_bump()reserves only 80 bytes of expansion headroom before callingslhc_uncompress(), but the reconstructed IP + TCP header is up to ip->ihl*4 + thp->doff*4 bytes. IHL and TCP doff are 4-bit fields and both can legitimately reach 15, so the header can grow to 2*15*4 = 120 bytes. A VJ-uncompressed primer with ihl=15, doff=15 followed by a compressed frame of size buffsize - 80 therefore writes up to 33 bytes past the kmalloc(buffsize + 4) rbuff allocation, with attacker-controlled content:BUG: KASAN: slab-out-of-bounds in slhc_uncompress Write of size 1069 at addr ffff88800ba93078 by task kworker/u8:1/32 Workqueue: events_unbound flush_to_ldisc Call Trace: __asan_memmove+0x3f/0x70 slhc_uncompress (drivers/net/slip/slhc.c:614) slip_receive_buf (drivers/net/slip/slip.c:342) tty_ldisc_receive_buf flush_to_ldiscRaise the reservation to match the real worst case. The
ppp_genericreceive path already enforcesskb_tailroom>= 124 and is unaffected.
Quoted 8 lines
Quoted 2 lines
Fixes b5451d783ade ("slip: Move the SLIP drivers")Reported-by Simon Horman <[email protected]>FTR, I was mainly passing on information flagged by Sashiko.
Quoted 1 line
Signed-off-by Weiming Shi <[email protected]>Reviewed-by Simon Horman <[email protected]>
I'm very sorry but the text below below was for a different, albeit related, patch:
- [PATCH net] slip: bound decode() reads against the compressed packet length
https://lore.kernel.org/netdev/[email protected]/
The corresponding text relating to this patch was posted as:
https://lore.kernel.org/netdev/[email protected]/
Sorry for the mix up!
Quoted 20 lines
Let me summarise my understanding of Sashiko's review of this patch.
TL;DR: I don't think that review should block progress of this patch.
1. The issue wrt concurrent MTU changes appears to be a separate, pre-existing problem. Maybe you've looked into it already, if not, you may wish to.
2. The bounds checking problems are addressed by other patches in flight.
- [PATCH net v2] slip: reject VJ receive packets on instances with no rstate array
https://lore.kernel.org/netdev/[email protected]/- [PATCH net] slip: fix slab-out-of-bounds write in
slhc_uncompress()https://lore.kernel.org/netdev/[email protected]/In future you might want to consider creating patch sets for related patches. But I think it's too late in the case of these patches.
...
Simon Horman
reviewerQuoted 36 lines · Simon Horman
Quoted 19 lines · Weiming Shi
sl_bump()reserves only 80 bytes of expansion headroom before callingslhc_uncompress(), but the reconstructed IP + TCP header is up to ip->ihl*4 + thp->doff*4 bytes. IHL and TCP doff are 4-bit fields and both can legitimately reach 15, so the header can grow to 2*15*4 = 120 bytes. A VJ-uncompressed primer with ihl=15, doff=15 followed by a compressed frame of size buffsize - 80 therefore writes up to 33 bytes past the kmalloc(buffsize + 4) rbuff allocation, with attacker-controlled content:BUG: KASAN: slab-out-of-bounds in slhc_uncompress Write of size 1069 at addr ffff88800ba93078 by task kworker/u8:1/32 Workqueue: events_unbound flush_to_ldisc Call Trace: __asan_memmove+0x3f/0x70 slhc_uncompress (drivers/net/slip/slhc.c:614) slip_receive_buf (drivers/net/slip/slip.c:342) tty_ldisc_receive_buf flush_to_ldiscQuoted 5 lines
Raise the reservation to match the real worst case. The
ppp_genericreceive path already enforcesskb_tailroom>= 124 and is unaffected.Fixes b5451d783ade ("slip: Move the SLIP drivers")Reported-by Simon Horman <[email protected]>FTR, I was mainly passing on a review generated by Sashiko
Quoted 1 line
Signed-off-by Weiming Shi <[email protected]>Reviewed-by Simon Horman <[email protected]>As usual I'll comment on the review of this patch by Sashiko.
TL;DR: I don't think it should block progress of this patch.
Quoted 8 lines
The review by Sashiko flags out of bounds errors. However, these are addressed by one of your other patches:
- [PATCH net] slip: bound
decode()reads against the compressed packet length
https://lore.kernel.org/netdev/[email protected]/As noted in my review of that patch, while it seems too late for these patches, please consider bundling related patches in a patchset in future.
I'm very sorry but the text above results from me muddling up my response to different slip patches.
I'll post the correct text for this patch elsewhere in this thread.
Simon Horman
reviewerQuoted 21 lines · Simon Horman
Quoted 20 lines · Simon Horman
Quoted 19 lines · Weiming Shi
sl_bump()reserves only 80 bytes of expansion headroom before callingslhc_uncompress(), but the reconstructed IP + TCP header is up to ip->ihl*4 + thp->doff*4 bytes. IHL and TCP doff are 4-bit fields and both can legitimately reach 15, so the header can grow to 2*15*4 = 120 bytes. A VJ-uncompressed primer with ihl=15, doff=15 followed by a compressed frame of size buffsize - 80 therefore writes up to 33 bytes past the kmalloc(buffsize + 4) rbuff allocation, with attacker-controlled content:BUG: KASAN: slab-out-of-bounds in slhc_uncompress Write of size 1069 at addr ffff88800ba93078 by task kworker/u8:1/32 Workqueue: events_unbound flush_to_ldisc Call Trace: __asan_memmove+0x3f/0x70 slhc_uncompress (drivers/net/slip/slhc.c:614) slip_receive_buf (drivers/net/slip/slip.c:342) tty_ldisc_receive_buf flush_to_ldisc
Quoted 4 lines
Quoted 4 lines
Quoted 4 lines
Raise the reservation to match the real worst case. The
ppp_genericreceive path already enforcesskb_tailroom>= 124 and is unaffected.Fixes b5451d783ade ("slip: Move the SLIP drivers")
AI review flags that this patch moved the code, rather than adding the bug. It suggests the bug has been present since the beginning of git history, so:
Quoted 11 lines
Quoted 7 lines
Quoted 1 line
Reported-by Simon Horman <[email protected]>FTR, I was mainly passing on information flagged by Sashiko.
Quoted 1 line
Signed-off-by Weiming Shi <[email protected]>Reviewed-by Simon Horman <[email protected]>I'm very sorry but the text below below was for a different, albeit related, patch:
Quoted 8 lines
- [PATCH net] slip: bound
decode()reads against the compressed packet length
https://lore.kernel.org/netdev/[email protected]/The corresponding text relating to this patch was posted as:
https://lore.kernel.org/netdev/[email protected]/
Sorry for the mix up!
Actually, that's not right either.
I will try one more time:
TL;DR: I don't think that review should block progress of this patch.
1. The issue wrt concurrent MTU changes appears to be a separate, pre-existing problem. Maybe you've looked into it already, if not, you may wish to.
2. The bounds checking problems are addressed by other patches in flight.
- [PATCH net v2] slip: reject VJ receive packets on instances with no rstate array
https://lore.kernel.org/netdev/[email protected]/
- [PATCH net] slip: bound decode() reads against the compressed packet length
https://lore.kernel.org/netdev/[email protected]/
In future you might want to consider creating patch sets for related patches. But I think it's too late in the case of these patches.