Kernel panic when compiled using gcc-4.7

Grzegorz Sikorski

Other Parts Discussed in Thread: AM3874

Hi,

I am trying to run new openembedded ditro on AM3874 board. I have found that 2.6.37 kernel needs no-unaligned-access switch and I was able to run it. However I still observe issue with notyfik module (pointed here: https://lists.yoctoproject.org/pipermail/meta-ti/2012-May/001144.html). Did anyone solved it already? Could we expect any TI patch for this issue?

Here is printout from serial console:

[    1.850000] usbhid: USB HID core driver
[    1.850000] Unable to handle kernel NULL pointer dereference at virtual address 00000002
[    1.860000] pgd = c0004000
[    1.860000] [00000002] *pgd=00000000
[    1.870000] Internal error: Oops: 5 [#1]
[    1.870000] last sysfs file:
[    1.870000] Modules linked in:
[    1.870000] CPU: 0    Not tainted (2.6.37 #1)
[    1.870000] PC is at strcmp+0x10/0x48
[    1.870000] LR is at omap_mbox_get+0x3c/0x1d0
[    1.870000] pc : [<c0226da0>]    lr : [<c005cecc>]    psr: a0000013
[    1.870000] sp : ef83df20 ip : ef83df30 fp : ef83df2c
[    1.870000] r10: 00000000 r9 : 00000000 r8 : 00000000
[    1.870000] r7 : c0026b58 r6 : c0553800 r5 : c04a3f58 r4 : c052d9e8
[    1.870000] r3 : 00000000 r2 : 00000064 r1 : c04a3f58 r0 : 00000002
[    1.870000] Flags: NzCv IRQs on FIQs on Mode SVC_32 ISA ARM Segment kernel
[    1.870000] Control: 10c5387d Table: 80004019 DAC: 00000017
[    1.870000] Process swapper (pid: 1, stack limit = 0xef83c2e8)
[    1.870000] Stack: (0xef83df20 to 0xef83e000)
[    1.870000] df20: ef83df54 ef83df30 c005cecc c0226d9c c033e138 c0226d9c c05536d8 c05537a4
[    1.870000] df40: c05537a0 c0067eb0 ef83df74 ef83df58 c0341f0c c005ce9c 00000000 00000000
[    1.870000] df60: c002c1c8 c002c24c ef83df9c ef83df78 c0026b94 c0341d84 ef83c000 c002c1c8
[    1.870000] df80: c002c24c c0067eb0 c0026b58 00000000 ef83dfd4 ef83dfa0 c00383bc c0026b64
[    1.870000] dfa0: ef83dfbc 00000176 c053b1a8 c0067eb0 ef83dfd4 c002c1c8 c002c24c c0067eb0
[    1.870000] dfc0: 00000013 00000000 ef83dff4 ef83dfd8 c0008a90 c003830c 00000000 00000000
[    1.870000] dfe0: 00000000 c00089d8 00000000 ef83dff8 c0067eb0 c00089e4 28258100 80741106
[    1.870000] Backtrace:
[    1.870000] [<c0226d90>] (strcmp+0x0/0x48) from [<c005cecc>] (omap_mbox_get+0x3c/0x1d0)
[    1.870000] [<c005ce90>] (omap_mbox_get+0x0/0x1d0) from [<c0341f0c>] (notify_shm_drv_setup+0x194/0x26c)
[    1.870000] r6:c0067eb0 r5:c05537a0 r4:c05537a4
[    1.870000] [<c0341d78>] (notify_shm_drv_setup+0x0/0x26c) from [<c0026b94>] (notify_init+0x3c/0x2b0)
[    1.870000] r5:c002c24c r4:c002c1c8
[    1.870000] [<c0026b58>] (notify_init+0x0/0x2b0) from [<c00383bc>] (do_one_initcall+0xbc/0x190)
[    1.870000] r8:00000000 r7:c0026b58 r6:c0067eb0 r5:c002c24c r4:c002c1c8
[    1.870000] r3:ef83c000
[    1.870000] [<c0038300>] (do_one_initcall+0x0/0x190) from [<c0008a90>] (kernel_init+0xb8/0x168)
[    1.870000] r8:00000000 r7:00000013 r6:c0067eb0 r5:c002c24c r4:c002c1c8
[    1.870000] [<c00089d8>] (kernel_init+0x0/0x168) from [<c0067eb0>] (do_exit+0x0/0x624)
[    1.870000] r5:c00089d8 r4:00000000
[    1.870000] Code: e1a0c00d e92dd800 e24cb004 e3a03000 (e7d02003)
[    2.100000] ---[ end trace f1865b47150691cb ]---
[    2.100000] Kernel panic - not syncing: Attempted to kill init!
[    2.110000] Backtrace:
[    2.110000] [<c0040cb0>] (dump_backtrace+0x0/0x10c) from [<c03ded10>] (dump_stack+0x18/0x1c)
[    2.120000] r6:c0531c6c r5:c0531c6c r4:c05593e0 r3:60000113
[    2.130000] [<c03decf8>] (dump_stack+0x0/0x1c) from [<c03def00>] (panic+0x5c/0x17c)
[    2.130000] [<c03deea4>] (panic+0x0/0x17c) from [<c0067f20>] (do_exit+0x70/0x624)
[    2.140000] r3:ef83816c r2:ef838000 r1:00000000 r0:c04a5d8f
[    2.150000] r7:00000001
[    2.150000] [<c0067eb0>] (do_exit+0x0/0x624) from [<c0041108>] (die+0x298/0x2d4)
[    2.160000] r7:00000000
[    2.160000] [<c0040e70>] (die+0x0/0x2d4) from [<c03ded70>] (__do_kernel_fault.part.2+0x5c/0x7c)
[    2.170000] [<c03ded14>] (__do_kernel_fault.part.2+0x0/0x7c) from [<c03e66c0>] (do_page_fault.part.3+0x270/0x288)
[    2.180000] r7:ef838000 r3:ef83ded8
[    2.180000] [<c03e6450>] (do_page_fault.part.3+0x0/0x288) from [<c03e672c>] (do_page_fault+0x54/0x60)
[    2.190000] [<c03e66d8>] (do_page_fault+0x0/0x60) from [<c03e675c>] (do_translation_fault+0x24/0xa4)
[    2.200000] r6:00000005 r5:c051a600 r4:00000005 r3:00000002
[    2.210000] [<c03e6738>] (do_translation_fault+0x0/0xa4) from [<c00381fc>] (do_DataAbort+0x3c/0xa0)
[    2.220000] r7:ef83ded8 r6:00000002 r5:c051a600 r4:00000005
[    2.230000] [<c00381c0>] (do_DataAbort+0x0/0xa0) from [<c03e472c>] (__dabt_svc+0x4c/0x60)
[    2.230000] Exception stack(0xef83ded8 to 0xef83df20)
[    2.240000] dec0:                                                       00000002 c04a3f58
[    2.250000] dee0: 00000064 00000000 c052d9e8 c04a3f58 c0553800 c0026b58 00000000 00000000
[    2.260000] df00: 00000000 ef83df2c ef83df30 ef83df20 c005cecc c0226da0 a0000013 ffffffff
[    2.260000] r7:c0026b58 r6:c0553800 r5:ef83df0c r4:ffffffff
[    2.270000] [<c0226d90>] (strcmp+0x0/0x48) from [<c005cecc>] (omap_mbox_get+0x3c/0x1d0)
[    2.280000] [<c005ce90>] (omap_mbox_get+0x0/0x1d0) from [<c0341f0c>] (notify_shm_drv_setup+0x194/0x26c)
[    2.290000] r6:c0067eb0 r5:c05537a0 r4:c05537a4
[    2.290000] [<c0341d78>] (notify_shm_drv_setup+0x0/0x26c) from [<c0026b94>] (notify_init+0x3c/0x2b0)
[    2.300000] r5:c002c24c r4:c002c1c8
[    2.310000] [<c0026b58>] (notify_init+0x0/0x2b0) from [<c00383bc>] (do_one_initcall+0xbc/0x190)
[    2.320000] r8:00000000 r7:c0026b58 r6:c0067eb0 r5:c002c24c r4:c002c1c8
[    2.320000] r3:ef83c000
[    2.330000] [<c0038300>] (do_one_initcall+0x0/0x190) from [<c0008a90>] (kernel_init+0xb8/0x168)
[    2.340000] r8:00000000 r7:00000013 r6:c0067eb0 r5:c002c24c r4:c002c1c8
[    2.340000] [<c00089d8>] (kernel_init+0x0/0x168) from [<c0067eb0>] (do_exit+0x0/0x624)
[    2.350000] r5:c00089d8 r4:00000000

over 12 years ago

0 Scott Dunnington over 12 years ago

Prodigy 90 points

There is an error in the omap_mbox_get with the loop using pointer math for each mbox structure location. It should be accessed as an array of pointers; the rest of the routines in this file are written this way.

Fullscreen 3872.mailbox.diff Download

--- a/arch/arm/plat-omap/mailbox.c
+++ b/arch/arm/plat-omap/mailbox.c
@@ -337,14 +337,16 @@ static void omap_mbox_fini(struct omap_mbox *mbox)
 struct omap_mbox *omap_mbox_get(const char *name, struct notifier_block *nb)
 {
 	struct omap_mbox *mbox;
-	int ret;
+	int ret, i;
 
 	if (!mboxes)
 		return ERR_PTR(-EINVAL);
 
-	for (mbox = *mboxes; mbox; mbox++)
+	for (i = 0; mboxes[i]; i++) {
+		mbox = mboxes[i];
 		if (!strcmp(mbox->name, name))
 			break;
+	}
 
 	if (!mbox)
 		return ERR_PTR(-ENOENT);

0 Grzegorz Sikorski over 12 years ago in reply to Scott Dunnington

Intellectual 320 points

Hi,

Thank you Scott, that was very helpful.

Kernel booted after this fix. However I cannot run eth port. Actually, system does not report any errors, but cannot mount NFS. If I run filesystem from memory (NAND jffs2), board works fine and ifconfig shows me correct (static) IP, but the board is inaccessible. Did you have similar issue?

If I set ip in bootargs to DHCP, system cannot discover proper IP:

[    1.920000] Detected MACID=0:17:eb:62:73:ac
[    2.440000]
[    2.440000] CPSW phy found : id is : 0x221612
[    3.460000] Sending DHCP requests ..
[    7.440000] PHY: 0:07 - Link is Up - 1000/Full
[    7.850000] DHCP/BOOTP: Reply not for us, op[2] xid[7c786298]
[   10.440000] ..
[   24.160000] DHCP/BOOTP: Reply not for us, op[2] xid[7d786298]
[   31.800000] .
[   41.700000] DHCP/BOOTP: Reply not for us, op[2] xid[7e786298]
[   41.890000] DHCP/BOOTP: Reply not for us, op[2] xid[eec9c8cc]
[   55.580000] .
[   57.910000] DHCP/BOOTP: Reply not for us, op[2] xid[7f786298]
[   64.150000] DHCP/BOOTP: Reply not for us, op[2] xid[36f02393]
[   74.000000] DHCP/BOOTP: Reply not for us, op[2] xid[80786298]
[   85.580000] timed out!
[   85.580000] IP-Config: Reopening network devices...
[   86.100000]
[   86.100000] CPSW phy found : id is : 0x221612
[   87.120000] Sending DHCP requests .
[   88.100000] PHY: 0:07 - Link is Up - 1000/Full
[   89.400000] .
[   90.280000] DHCP/BOOTP: Reply not for us, op[2] xid[81786298]
[   93.390000] ..
[ 106.480000] DHCP/BOOTP: Reply not for us, op[2] xid[82786298]
[ 112.580000] .
[ 117.730000] DHCP/BOOTP: Reply not for us, op[2] xid[66ae42ae]
[ 124.140000] DHCP/BOOTP: Reply not for us, op[2] xid[83786298]
[ 125.290000] DHCP/BOOTP: Reply not for us, op[2] xid[304ed245]
[ 127.030000] DHCP/BOOTP: Reply not for us, op[2] xid[304ed245]
[ 128.830000] DHCP/BOOTP: Reply not for us, op[2] xid[304ed245]
[ 129.820000] DHCP/BOOTP: Reply not for us, op[2] xid[304ed245]
[ 131.820000] DHCP/BOOTP: Reply not for us, op[2] xid[304ed245]
[ 133.940000] .
[ 134.530000] DHCP/BOOTP: Reply not for us, op[2] xid[16bd20a7]
[ 135.820000] DHCP/BOOTP: Reply not for us, op[2] xid[304ed245]
[ 137.020000] DHCP/BOOTP: Reply not for us, op[2] xid[4021b2ba]
[ 137.020000] DHCP/BOOTP: Reply not for us, op[2] xid[4021b2ba]
[ 137.030000] DHCP/BOOTP: Reply not for us, op[2] xid[4021b2ba]
[ 137.040000] DHCP/BOOTP: Reply not for us, op[2] xid[4021b2ba]
[ 140.500000] DHCP/BOOTP: Reply not for us, op[2] xid[84786298]
[ 157.760000] DHCP/BOOTP: Reply not for us, op[2] xid[85786298]
[ 163.940000] timed out!
[ 163.940000] IP-Config: Auto-configuration of network failed.

0 Grzegorz Sikorski over 12 years ago in reply to Grzegorz Sikorski

Intellectual 320 points

OK, never mind, I have found that issue is caused by my build system. For some reason it did not rebuild the kernel properly. Thanks once again!

Processors

Processors forum

Kernel panic when compiled using gcc-4.7