Bartek wrote: > Hopefully, this time it my bug report should be ok :): > > Apr 11 23:53:38 localhost pppd[31289]: rcvd [proto=0x7689] e1 cd 33 f6 > fd f7 52 e6 58 c9 73 98 bc ff ad d5 b5 a3 e5 d9 1e 77 76 0a 1c 87 59 > bf 44 cc ac 3b ... > Apr 11 23:53:38 localhost pppd[31289]: Unsupported protocol 0x7689 received > Apr 11 23:53:38 localhost pppd[31289]: sent [LCP ProtRej id=0x9 76 89 > e1 cd 33 f6 fd f7 52 e6 58 c9 73 98 bc ff ad d5 b5 a3 e5 d9 1e 77 76 > 0a 1c 87 59 bf 44 cc ...] > Apr 11 23:53:38 localhost pppd[31289]: rcvd [proto=0xda7d] 15 19 45 3c > e0 ac 44 92 3b c4 8e 75 6b b8 4a 9f 4a 3a 22 63 d3 a1 56 98 47 62 bc > cd a6 8e d5 77 ... > Apr 11 23:53:38 localhost pppd[31289]: Unsupported protocol 0xda7d received > Apr 11 23:53:38 localhost pppd[31289]: sent [LCP ProtRej id=0xa da 7d > 15 19 45 3c e0 ac 44 92 3b c4 8e 75 6b b8 4a 9f 4a 3a 22 63 d3 a1 56 > 98 47 62 bc cd a6 8e ...] > Apr 11 23:53:40 localhost kernel: skb_under_panic: text:f8c62c0e > len:291 put:1 head:ddc94800 data:ddc947ff tail:ddc94922 end:ddc94e00 > dev:<NULL>
It seems we fail to reserve enough headroom for the case buf[0] == PPP_ALLSTATIONS and buf[1] != PPP_UI. Can you try this patch please?
diff --git a/drivers/net/ppp_async.c b/drivers/net/ppp_async.c index 933e2f3..c68e37f 100644 --- a/drivers/net/ppp_async.c +++ b/drivers/net/ppp_async.c @@ -890,6 +890,8 @@ ppp_async_input(struct asyncppp *ap, const unsigned char *buf, ap->rpkt = skb; } if (skb->len == 0) { + int headroom = 0; + /* Try to get the payload 4-byte aligned. * This should match the * PPP_ALLSTATIONS/PPP_UI/compressed tests in @@ -897,7 +899,10 @@ ppp_async_input(struct asyncppp *ap, const unsigned char *buf, * enough chars here to test buf[1] and buf[2]. */ if (buf[0] != PPP_ALLSTATIONS) - skb_reserve(skb, 2 + (buf[0] & 1)); + headroom += 2; + if (buf[0] & 1) + headroom += 1; + skb_reserve(skb, headroom); } if (n > skb_tailroom(skb)) { /* packet overflowed MRU */