grsecurity forums

Hello,

I ran in to a couple problems building with grsec and 4.3.3 for arm5. It's probably easiest to explain with the errors and patches I hacked up to fix it.

Code: Select all

In file included from /mnt/scratch/fixes/linux-4.3.3/include/linux/atomic.h:4:0,
                 from /mnt/scratch/fixes/linux-4.3.3/include/linux/spinlock.h:417,
                 from /mnt/scratch/fixes/linux-4.3.3/include/linux/seqlock.h:35,
                 from /mnt/scratch/fixes/linux-4.3.3/include/linux/time.h:5,
                 from /mnt/scratch/fixes/linux-4.3.3/include/uapi/linux/timex.h:56,
                 from /mnt/scratch/fixes/linux-4.3.3/include/linux/timex.h:56,
                 from /mnt/scratch/fixes/linux-4.3.3/include/linux/sched.h:19,
                 from /mnt/scratch/fixes/linux-4.3.3/arch/arm/kernel/asm-offsets.c:14:
/mnt/scratch/fixes/linux-4.3.3/arch/arm/include/asm/atomic.h: In function ‘atomic_inc_and_test_unchecked’:
/mnt/scratch/fixes/linux-4.3.3/arch/arm/include/asm/atomic.h:326:9: error: implicit declaration of function ‘atomic_add_return_unchecked_relaxed’ [-Werror=implicit-function-declaration]
  return atomic_add_return_unchecked_relaxed(1, v) == 0;
         ^

There is probably a more elegant fix, assuming this isn't just outright wrong:

Code: Select all

--- linux-4.3.3/arch/arm/include/asm/atomic.h   2015-12-27 01:57:10.042670349 -0600
+++ /mnt/scratch/linux-4.3.3/arch/arm/include/asm/atomic.h   2015-12-27 01:58:20.369607745 -0600
@@ -323,13 +323,21 @@
 #define atomic_inc_and_test(v)   (atomic_add_return(1, v) == 0)
 static inline int atomic_inc_and_test_unchecked(atomic_unchecked_t *v)
 {
+#if __LINUX_ARM_ARCH__ >= 6
    return atomic_add_return_unchecked_relaxed(1, v) == 0;
+#else
+   return atomic_add_return_unchecked(1, v) == 0;
+#endif
 }
 #define atomic_dec_and_test(v)   (atomic_sub_return(1, v) == 0)
 #define atomic_inc_return(v)    (atomic_add_return(1, v))
 static inline int atomic_inc_return_unchecked(atomic_unchecked_t *v)
 {
+#if __LINUX_ARM_ARCH__ >= 6
    return atomic_add_return_unchecked_relaxed(1, v);
+#else
+   return atomic_add_return_unchecked(1, v);
+#endif
 }
 #define atomic_dec_return(v)    (atomic_sub_return(1, v))
 #define atomic_sub_and_test(i, v) (atomic_sub_return(i, v) == 0)

Also this

Code: Select all

In file included from /mnt/scratch/fixes/linux-4.3.3/arch/arm/include/asm/uaccess.h:18:0,
                 from /mnt/scratch/fixes/linux-4.3.3/include/linux/uaccess.h:5,
                 from /mnt/scratch/fixes/linux-4.3.3/arch/arm/kernel/ptrace.c:22:
/mnt/scratch/fixes/linux-4.3.3/arch/arm/include/asm/uaccess.h: In function ‘set_fs’:
/mnt/scratch/fixes/linux-4.3.3/arch/arm/include/asm/uaccess.h:103:36: error: ‘DOMAIN_KERNELCLIENT’ undeclared (first use in this function)
  modify_domain(DOMAIN_KERNEL, fs ? DOMAIN_KERNELCLIENT : DOMAIN_MANAGER);
                                    ^

I think may have been a mistake here?

Code: Select all

--- linux-4.3.3/arch/arm/include/asm/domain.h   2015-12-27 01:50:09.544023866 -0600
+++ /mnt/scratch/linux-4.3.3/arch/arm/include/asm/domain.h   2015-12-26 16:09:01.330864530 -0600
@@ -69,9 +69,9 @@
 #define DOMAIN_USERCLIENT   1
 #define DOMAIN_VECTORS      DOMAIN_USER
 #endif
-#define DOMAIN_KERNELCLIENT   1
 
 #endif
+#define DOMAIN_KERNELCLIENT   1
 
 #define domain_mask(dom)   ((3) << (2 * (dom)))
 #define domain_val(dom,type)   ((type) << (2 * (dom)))

I've only just started running a kernel with the above changes, it hasn't blown up yet. I hope this is helpful. And thanks!

thanks, i fixed these and other issues that i ran across on armv5, if you have configs that expose more breakage, let me know.

Just a couple more (unrelated) things I hit. Immediately on any network activity, I ran into overflows in skb_headers_offset_update (net/core/skbuff.c) and tcp_parse_options (net/ipv4/tcp_input.c). I played monkey-see-monkey-do and made it "work" for me by patching in attributes to the functions, __intentional_overflow(0) for the first, and -1 for the second. I should be completely clear that I know I don't know enough to understand the attributes beyond common sense, and what I did is just 'make it "work."' I'm using both bonding and tagged vlan interfaces on top if that's significant (and I believe that's why the .._offset_update), let me know what I can provide that's useful. I didn't grab any output from the panics these generated unfortunately, but I can always unpatch and re-create them if it helps. skb_headers_update failed right away on ->transport_header, I don't remember anything about what tcp_parse_options barfed on.

you should get the size overflow reports so that we can take a look (boot with pax_size_overflow_report_only to not panic the system on a size overflow event). skb related code is known to have some false positives (introduced by gcc too early that the plugin cannot tell them apart) but also some real problems (e.g., i just fixed/worked around one in skb_headers_offset_update) so it'd be nice to sort more of them out.

that it panics immediately actually makes it easier for me to capture. here are backtraces of each, first skb_headers_offset_update and then after i add __intentional_overflow there i hit tcp_parse_options. let me know if there's anything else useful i can provide.

Code: Select all

[   26.419682] PAX: size overflow detected in function skb_headers_offset_update /mnt/scratch/linux-4.3.3/net/core/skbuff.c:972 cicus.542_39 max, count: 5, decl: transport_header; num: 0; context: sk_buff;
[   26.437780] CPU: 0 PID: 0 Comm: swapper Not tainted 4.3.3-grsec #15
[   26.444078] Hardware name: Marvell Kirkwood (Flattened Device Tree)
[   26.450383] Backtrace:
[   26.452868] [<c01054fc>] (dump_backtrace+0x0/0x118) from [<c01056a4>] (show_stack+0x18/0x1c)
[   26.461364]  r7:0000001f r6:00000020 r5:00000020 r4:df5e3b60
[   26.467130] [<c010568c>] (show_stack+0x0/0x1c) from [<c0318b78>] (dump_stack+0x20/0x28)
[   26.475188] [<c0318b58>] (dump_stack+0x0/0x28) from [<c01bcc70>] (report_size_overflow+0x30/0x3c)
[   26.484117] [<c01bcc40>] (report_size_overflow+0x0/0x3c) from [<c046ce68>] (skb_headers_offset_update+0xf0/0x19c)
[   26.494440] [<c046cd78>] (skb_headers_offset_update+0x0/0x19c) from [<c046f350>] (pskb_expand_head+0x200/0x2e0)
[   26.504585]  r7:df462a4c r6:df462b40 r5:df5e3b60 r4:00000020
[   26.510356] [<c046f150>] (pskb_expand_head+0x0/0x2e0) from [<c04844f0>] (validate_xmit_skb.part.10+0x2e4/0x414)
[   26.520505]  r10:00000000 r9:00000000 r8:00000081 r7:00000000 r6:00010003 r5:00001026
[   26.528477]  r4:df5e3b60 r3:00000020
[   26.532107] [<c048420c>] (validate_xmit_skb.part.10+0x0/0x414) from [<c048496c>] (validate_xmit_skb_list+0x30/0x60)
[   26.542604]  r9:00000000 r8:df5e3b60 r7:00000000 r6:00000000 r5:df7b7000 r4:00000000
[   26.550471] [<c048493c>] (validate_xmit_skb_list+0x0/0x60) from [<c04a7b94>] (sch_direct_xmit+0x130/0x228)
[   26.560184]  r7:df7b7000 r6:df589000 r5:df7b7000 r4:df501300
[   26.565939] [<c04a7a64>] (sch_direct_xmit+0x0/0x228) from [<c0484d0c>] (__dev_queue_xmit+0x2a8/0x554)
[   26.575209]  r8:00000000 r7:df7b7000 r6:00000000 r5:df501300 r4:df5e3b60
[   26.582037] [<c0484a64>] (__dev_queue_xmit+0x0/0x554) from [<c0484fd0>] (dev_queue_xmit_sk+0x18/0x1c)
[   26.591316]  r10:df5e3b60 r9:00000100 r8:0000002a r7:00000001 r6:df5e3b60 r5:df4ec920
[   26.599275]  r4:df560000
[   26.601845] [<c0484fb8>] (dev_queue_xmit_sk+0x0/0x1c) from [<c03d607c>] (bond_start_xmit+0x244/0x4a4)
[   26.611122] [<c03d5e38>] (bond_start_xmit+0x0/0x4a4) from [<c0484868>] (dev_hard_start_xmit+0x248/0x31c)
[   26.620662]  r9:00000100 r8:0000002a r7:df560000 r6:df560000 r5:c0c27e58 r4:df5e3b60
[   26.628535] [<c0484620>] (dev_hard_start_xmit+0x0/0x31c) from [<c0484f50>] (__dev_queue_xmit+0x4ec/0x554)
[   26.638148]  r10:df5e3b60 r9:00000100 r8:0000002a r7:df74f800 r6:df560000 r5:c0c27e58
[   26.646116]  r4:df5e3b60
[   26.648702] [<c0484a64>] (__dev_queue_xmit+0x0/0x554) from [<c0484fd0>] (dev_queue_xmit_sk+0x18/0x1c)
[   26.657968]  r10:df5e3b60 r9:00000100 r8:0000002a r7:df6c7000 r6:df6c7000 r5:c0c27e58
[   26.665938]  r4:df6c7000
[   26.668516] [<c0484fb8>] (dev_queue_xmit_sk+0x0/0x1c) from [<c060a340>] (vlan_dev_hard_start_xmit+0x94/0xf8)
[   26.678404] [<c060a2ac>] (vlan_dev_hard_start_xmit+0x0/0xf8) from [<c0484868>] (dev_hard_start_xmit+0x248/0x31c)
[   26.688639]  r9:00000100 r8:dedcc034 r7:df6c7000 r6:df6c7000 r5:c0c27e58 r4:df5e3b60
[   26.696495] [<c0484620>] (dev_hard_start_xmit+0x0/0x31c) from [<c0484f50>] (__dev_queue_xmit+0x4ec/0x554)
[   26.706123]  r10:00000006 r9:00000100 r8:dedcc034 r7:df69da00 r6:df6c7000 r5:c0c27e58
[   26.714090]  r4:df5e3b60
[   26.716656] [<c0484a64>] (__dev_queue_xmit+0x0/0x554) from [<c0484fd0>] (dev_queue_xmit_sk+0x18/0x1c)
[   26.725929]  r10:00000006 r9:00000100 r8:dedcc034 r7:df500c20 r6:df6c7000 r5:26783c0a
[   26.733901]  r4:df5e3b60
[   26.736469] [<c0484fb8>] (dev_queue_xmit_sk+0x0/0x1c) from [<c0543438>] (arp_xmit+0x88/0x94)
[   26.744962] [<c05433b0>] (arp_xmit+0x0/0x94) from [<c05434b0>] (arp_send_dst.part.1+0x6c/0x74)
[   26.753629]  r7:df500c20 r6:df6c7000 r5:26783c0a r4:00000000
[   26.759401] [<c0543444>] (arp_send_dst.part.1+0x0/0x74) from [<c0543b28>] (arp_process+0x670/0x714)
[   26.768496]  r5:00000000 r4:df6d4b60
[   26.772123] [<c05434b8>] (arp_process+0x0/0x714) from [<c0543d18>] (arp_rcv+0x14c/0x21c)
[   26.780260]  r10:c0c02660 r9:df6c705c r8:c0c097f8 r7:00000608 r6:df6c7000 r5:c0543bcc
[   26.788217]  r4:df6d4b60
[   26.790802] [<c0543bcc>] (arp_rcv+0x0/0x21c) from [<c04800d4>] (__netif_receive_skb_core+0x650/0xb6c)
[   26.800084]  r9:df6c705c r8:c0c097f8 r6:df6c7048 r5:c0543bcc r4:df6d4b60
[   26.806907] [<c047fa84>] (__netif_receive_skb_core+0x0/0xb6c) from [<c0480fe8>] (__netif_receive_skb+0x2c/0x8c)
[   26.817049]  r10:00000001 r9:df5893e0 r8:c0108acc r7:2ccffffe r6:df589000 r5:df6d4b60
[   26.825016]  r4:df6d4b60
[   26.827592] [<c0480fbc>] (__netif_receive_skb+0x0/0x8c) from [<c0481074>] (netif_receive_skb_internal+0x2c/0x74)
[   26.837830]  r5:df6d4b60 r4:df6d4b60
[   26.841466] [<c0481048>] (netif_receive_skb_internal+0x0/0x74) from [<c04853a8>] (napi_gro_receive+0x80/0xac)
[   26.851441]  r4:00000003
[   26.854014] [<c0485328>] (napi_gro_receive+0x0/0xac) from [<c03eebd4>] (mv643xx_eth_poll+0x4cc/0x664)
[   26.863292]  r5:0000003c r4:df589554
[   26.866922] [<c03ee708>] (mv643xx_eth_poll+0x0/0x664) from [<c0485cc4>] (net_rx_action+0x3b0/0x6a8)
[   26.876025]  r10:c0c01dd0 r9:c085300c r8:c0853000 r7:98776218 r6:c5bbad1a r5:c03ee708
[   26.883989]  r4:df5894a8
[   26.886564] [<c0485914>] (net_rx_action+0x0/0x6a8) from [<c01131c4>] (__do_softirq+0xe4/0x244)
[   26.895226]  r10:00000100 r9:c085300c r8:c0853000 r7:00000003 r6:40000003 r5:c0c00000
[   26.903182]  r4:00000000
[   26.905750] [<c01130e0>] (__do_softirq+0x0/0x244) from [<c01135e8>] (irq_exit+0xbc/0xd4)
[   26.913894]  r10:df404400 r9:00000000 r8:df406000 r7:00000001 r6:00000000 r5:c0c18f18
[   26.921860]  r4:00000000
[   26.924429] [<c011352c>] (irq_exit+0x0/0xd4) from [<c013db14>] (__handle_domain_irq+0x58/0xb0)
[   26.933099]  r5:c0c18f18 r4:00000000
[   26.936730] [<c013dabc>] (__handle_domain_irq+0x0/0xb0) from [<c01014ac>] (orion_handle_irq+0x74/0xa0)
[   26.946089]  r9:00000000 r8:c0c01ec8 r7:00000001 r6:c0c7ba90 r5:df40441c r4:00000000
[   26.953954] [<c0101438>] (orion_handle_irq+0x0/0xa0) from [<c01061b0>] (__irq_svc+0x50/0x64)
[   26.962442] Exception stack(0xc0c01ec8 to 0xc0c01f10)
[   26.967519] 1ec0:                   00000000 00000006 263b5386 c04364c8 2628cd6a 00000006
[   26.975744] 1ee0: 263b5386 00000006 c0c24540 00000001 00000001 c0c01f5c c0c01f18 c0c01f18
[   26.983963] 1f00: c043637c c043638c 00000013 ffffffff
[   26.989050]  r10:00000001 r9:00000001 r8:c0c24540 r7:c0c01efc r6:ffffffff r5:00000013
[   26.997009]  r4:c043638c r3:c043637c
[   27.000655] [<c0436204>] (cpuidle_enter_state+0x0/0x2d4) from [<c0436518>] (cpuidle_enter+0x1c/0x20)
[   27.009841]  r10:c0c0206c r9:00000000 r8:c0c24538 r7:c0c24950 r6:c0c0206c r5:c0c24540
[   27.017798]  r4:c0c00000
[   27.020389] [<c04364fc>] (cpuidle_enter+0x0/0x20) from [<c0136edc>] (cpu_startup_entry+0x138/0x1a0)
[   27.029494] [<c0136da4>] (cpu_startup_entry+0x0/0x1a0) from [<c060fdbc>] (rest_init+0x64/0x7c)
[   27.038154]  r7:ffffffff
[   27.040734] [<c060fd58>] (rest_init+0x0/0x7c) from [<c0901990>] (0xc0901990)
[   27.047810] [<c0901314>] (0xc0901314) from [<00008048>] (0x8048)
[   27.053855] Kernel panic - not syncing: Aiee, killing interrupt handler!
[   27.060583] ---[ end Kernel panic - not syncing: Aiee, killing interrupt handler!

Code: Select all

[   17.283872] PAX: size overflow detected in function tcp_parse_options /mnt/scratch/linux-4.3.3/include/linux/unaligned/be_byteshift.h:13 cicus.1806_179 max, count: 225, decl: rcv_tsecr; num: 0; context: tcp_options_received;
[   17.303891] CPU: 0 PID: 0 Comm: swapper Not tainted 4.3.3-grsec #16
[   17.310196] Hardware name: Marvell Kirkwood (Flattened Device Tree)
[   17.316488] Backtrace:
[   17.318974] [<c01054fc>] (dump_backtrace+0x0/0x118) from [<c01056a4>] (show_stack+0x18/0x1c)
[   17.327471]  r7:df6d69e0 r6:dedd4060 r5:fffff989 r4:0000000a
[   17.333245] [<c010568c>] (show_stack+0x0/0x1c) from [<c0318b78>] (dump_stack+0x20/0x28)
[   17.341303] [<c0318b58>] (dump_stack+0x0/0x28) from [<c01bcc70>] (report_size_overflow+0x30/0x3c)
[   17.350237] [<c01bcc40>] (report_size_overflow+0x0/0x3c) from [<c05283bc>] (tcp_parse_options+0x4b0/0x4d4)
[   17.359957] [<c0527f0c>] (tcp_parse_options+0x0/0x4d4) from [<c0536718>] (tcp_check_req+0x118/0x494)
[   17.369141]  r10:00001000 r9:00000000 r8:defc1a60 r7:e5001080 r6:df631cc0 r5:dedd4048
[   17.377099]  r4:df6d69e0
[   17.379684] [<c0536600>] (tcp_check_req+0x0/0x494) from [<c0533038>] (tcp_v4_do_rcv+0x174/0x36c)
[   17.388524]  r10:00000048 r9:dedd4048 r8:df631cc0 r7:dedd4034 r6:dedd4000 r5:defc1a60
[   17.396483]  r4:df6d69e0
[   17.399068] [<c0532ec4>] (tcp_v4_do_rcv+0x0/0x36c) from [<c05358f8>] (tcp_v4_rcv+0x94c/0x9d8)
[   17.407637]  r10:0000fe92 r9:00001131 r8:defc1a60 r7:dedd4034 r6:c0c270a0 r5:00000000
[   17.415606]  r4:df6d69e0
[   17.418172] [<c0534fac>] (tcp_v4_rcv+0x0/0x9d8) from [<c050c4b0>] (ip_local_deliver_finish+0xcc/0x298)
[   17.427529]  r10:c0c02660 r9:00000020 r8:c0c270a0 r7:00000000 r6:c0c08ea0 r5:c0736194
[   17.435497]  r4:df6d69e0
[   17.438063] [<c050c3e4>] (ip_local_deliver_finish+0x0/0x298) from [<c050cbe8>] (ip_local_deliver+0xc0/0xcc)
[   17.447865]  r9:00000000 r8:00000002 r7:00000001 r6:80000000 r5:00000002 r4:df6d69e0
[   17.455737] [<c050cb28>] (ip_local_deliver+0x0/0xcc) from [<c050c7c8>] (ip_rcv_finish+0x14c/0x388)
[   17.464746]  r7:dedd4000 r6:80000000 r5:dedd4034 r4:df6d69e0
[   17.470509] [<c050c67c>] (ip_rcv_finish+0x0/0x388) from [<c050cf2c>] (ip_rcv+0x338/0x428)
[   17.478733]  r7:dedd4000 r6:80000000 r5:df68e000 r4:df6d69e0
[   17.484490] [<c050cbf4>] (ip_rcv+0x0/0x428) from [<c047ffc4>] (__netif_receive_skb_core+0x650/0xb6c)
[   17.493679]  r9:df68e05c r8:c0c09850 r7:00000008 r6:df68e048 r5:c050cbf4 r4:df6d69e0
[   17.501563] [<c047f974>] (__netif_receive_skb_core+0x0/0xb6c) from [<c0480ed8>] (__netif_receive_skb+0x2c/0x8c)
[   17.511712]  r10:00000001 r9:df5893e0 r8:c0108acc r7:6f8929c6 r6:df589000 r5:df6d69e0
[   17.519679]  r4:df6d69e0
[   17.522245] [<c0480eac>] (__netif_receive_skb+0x0/0x8c) from [<c0480f64>] (netif_receive_skb_internal+0x2c/0x74)
[   17.532478]  r5:df6d69e0 r4:df6d69e0
[   17.536114] [<c0480f38>] (netif_receive_skb_internal+0x0/0x74) from [<c0485298>] (napi_gro_receive+0x80/0xac)
[   17.546088]  r4:00000003
[   17.548668] [<c0485218>] (napi_gro_receive+0x0/0xac) from [<c03eebd4>] (mv643xx_eth_poll+0x4cc/0x664)
[   17.557935]  r5:00000046 r4:df589554
[   17.561576] [<c03ee708>] (mv643xx_eth_poll+0x0/0x664) from [<c0485bb4>] (net_rx_action+0x3b0/0x6a8)
[   17.570678]  r10:c0c01db8 r9:c085300c r8:c0853000 r7:98776218 r6:c5bbad1a r5:c03ee708
[   17.578649]  r4:df5894a8
[   17.581217] [<c0485804>] (net_rx_action+0x0/0x6a8) from [<c01131c4>] (__do_softirq+0xe4/0x244)
[   17.589880]  r10:00000100 r9:c085300c r8:c0853000 r7:00000003 r6:40000003 r5:c0c00000
[   17.597836]  r4:00000000
[   17.600422] [<c01130e0>] (__do_softirq+0x0/0x244) from [<c01135e8>] (irq_exit+0xbc/0xd4)
[   17.608566]  r10:df404400 r9:00000000 r8:df406000 r7:00000001 r6:00000000 r5:c0c18f18
[   17.616515]  r4:00000000
[   17.619102] [<c011352c>] (irq_exit+0x0/0xd4) from [<c013db14>] (__handle_domain_irq+0x58/0xb0)
[   17.627756]  r5:c0c18f18 r4:00000000
[   17.631394] [<c013dabc>] (__handle_domain_irq+0x0/0xb0) from [<c01014ac>] (orion_handle_irq+0x74/0xa0)
[   17.640755]  r9:00000000 r8:c0c01eb0 r7:00000001 r6:c0c7ba90 r5:df40441c r4:00000000
[   17.648627] [<c0101438>] (orion_handle_irq+0x0/0xa0) from [<c01061b0>] (__irq_svc+0x50/0x64)
[   17.657102] Exception stack(0xc0c01eb0 to 0xc0c01ef8)
[   17.662191] 1ea0:                                     df76bc00 00000002 00000000 c0c01f00
[   17.670415] 1ec0: df5cc620 df76bc00 00000001 00000000 df5cc620 00000000 c0c0206c c0c01f1c
[   17.678637] 1ee0: c0c01f20 c0c01f00 c0610874 c012efbc 60000013 ffffffff
[   17.685282]  r10:c0c0206c r9:00000000 r8:df5cc620 r7:c0c01ee4 r6:ffffffff r5:60000013
[   17.693251]  r4:c012efbc r3:c0610874
[   17.696892] [<c012ef88>] (finish_task_switch+0x0/0x110) from [<c0610874>] (__schedule+0x19c/0x3cc)
[   17.705909]  r7:00000000 r6:c0c0e6d8 r5:c0c0b828 r4:df76bc00
[   17.711671] [<c06106d8>] (__schedule+0x0/0x3cc) from [<c0610b30>] (schedule+0x44/0x9c)
[   17.719641]  r9:00000000 r8:c0c24538 r7:c0c24950 r6:c0c0206c r5:c0c24540 r4:c0c00000
[   17.727516] [<c0610aec>] (schedule+0x0/0x9c) from [<c0610b98>] (schedule_preempt_disabled+0x10/0x14)
[   17.736706]  r5:c0c24540 r4:c0c00000
[   17.740349] [<c0610b88>] (schedule_preempt_disabled+0x0/0x14) from [<c0136e4c>] (cpu_startup_entry+0xa8/0x1a0)
[   17.750413] [<c0136da4>] (cpu_startup_entry+0x0/0x1a0) from [<c060fcac>] (rest_init+0x64/0x7c)
[   17.759083]  r7:ffffffff
[   17.761644] [<c060fc48>] (rest_init+0x0/0x7c) from [<c0901990>] (0xc0901990)
[   17.768732] [<c0901314>] (0xc0901314) from [<00008048>] (0x8048)
[   17.774762] Kernel panic - not syncing: Aiee, killing interrupt handler!
[   17.781493] ---[ end Kernel panic - not syncing: Aiee, killing interrupt handler!

thanks, the first one should be fixed (well, worked around) in the latest patch but it's still worth a question upstream as i think there may be a logic bug at a higher level here. the second one looks like a false positive in that some of the fields in tcp_options_received are clearly not size related, we'll remove them from the hash table.