[3.15-rc3] BUG: null ptr dereference in ichx_gpio_request_regions()

[Peter Hurley]

Booting 3.15-rc3, I get this BUG when loading gpio_ich:

 BUG: unable to handle kernel NULL pointer dereference at           (null)
 IP: [<ffffffffa042339c>] ichx_gpio_probe+0x28c/0x3d0 [gpio_ich]
 usbcore: registered new interface driver btusb
 PGD 2b04aa067 PUD 2af912067 PMD 0
 Oops: 0000 [#1] PREEMPT SMP
 Modules linked in: gpio_ich(+) btusb bluetooth psmouse snd i5400_edac ....
 CPU: 3 PID: 1217 Comm: modprobe Not tainted 3.15.0-rc3+wip-xeon #rc3+wip
 Hardware name: Dell Inc. Precision WorkStation T5400  /0RW203, BIOS A11 04/30/2012
 task: ffff8802ae8448f0 ti: ffff8802b0d74000 task.ti: ffff8802b0d74000
 RIP: 0010:[<ffffffffa042339c>]  [<ffffffffa042339c>] ichx_gpio_probe+0x28c/0x3d0 [gpio_ich]
 RSP: 0018:ffff8802b0d75b78  EFLAGS: 00010246
 RAX: 0000000000000000 RBX: 0000000000000000 RCX: 0000000000000000
 RDX: 0000000000000000 RSI: 0000000000000100 RDI: ffffffff81c378a0
 RBP: ffff8802b0d75bb8 R08: 0000000000000000 R09: ffff880036a0e2c8
 R10: 0000000000005dc0 R11: 8000000000000000 R12: ffff880036a0e000
 R13: ffff8800bad62bc0 R14: 0000000000000003 R15: 0000000000000000
 FS:  00007fb9d38fa700(0000) GS:ffff8802bfcc0000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 0000000000000000 CR3: 00000002af445000 CR4: 00000000000007e0
 Stack:
  ffff8802b0d75b98 ffff880036a0e010 ffff880036a0e020 ffff880036a0e010
  ffffffffa0425028 ffffffffa0425028 0000000000000000 0000000000000001
  ffff8802b0d75be8 ffffffff814793f2 ffff8802b0d75ca8 ffff880036a0e010
 Call Trace:
  [<ffffffff814793f2>] platform_drv_probe+0x32/0x80
  [<ffffffff8147784b>] driver_probe_device+0x8b/0x3a0
  [<ffffffff81477c0b>] __driver_attach+0xab/0xb0
  [<ffffffff81477b60>] ? driver_probe_device+0x3a0/0x3a0
  [<ffffffff8147586d>] bus_for_each_dev+0x5d/0xa0
  [<ffffffff8147727e>] driver_attach+0x1e/0x20
  [<ffffffff81476dd4>] bus_add_driver+0x124/0x250
  [<ffffffffa029a000>] ? 0xffffffffa0299fff
  [<ffffffff81478314>] driver_register+0x64/0xf0
  [<ffffffffa029a000>] ? 0xffffffffa0299fff
  [<ffffffff8147926a>] __platform_driver_register+0x4a/0x50
  [<ffffffffa029a017>] ichx_gpio_driver_init+0x17/0x1000 [gpio_ich]
  [<ffffffff8100032a>] do_one_initcall+0xda/0x180
  [<ffffffff8103e733>] ? set_memory_nx+0x43/0x50
  [<ffffffff816ffeec>] ? set_section_ro_nx+0x6d/0x75
  [<ffffffff810cc9f9>] load_module+0x1d79/0x2770
  [<ffffffff810c8690>] ? unset_module_init_ro_nx+0x80/0x80
  [<ffffffff81172f80>] ? __vmalloc_node_range+0x170/0x250
  [<ffffffff810cd479>] ? SyS_init_module+0x89/0x100
  [<ffffffff810cd4a2>] SyS_init_module+0xb2/0x100
  [<ffffffff81719ad2>] system_call_fastpath+0x16/0x1b
 Code: c7 05 fd 1f 00 00 40 51 42 a0 e9 00 fe ff ff 48 8b 05 f1 1f 00 00 45 31 c0 48 c7 c7 a0 78 c3 81 48 8b 48 08 48 8b 50 10 48 63 c3 <0f> b6 34 01 4c 89 c9 0f b6 14 1a 49 03 75 00 4c 89 4d c8 e8 ec
 RIP  [<ffffffffa042339c>] ichx_gpio_probe+0x28c/0x3d0 [gpio_ich]
  RSP <ffff8802b0d75b78>
 CR2: 0000000000000000


This is almost certainly caused by the uninitialized regs ptr
in the ich6_desc struct (i3100_desc struct has the same problem)
introduced in this commit:

commit bb62a35bd5d96e506af0ea8dd145480b9172a2a6
Author: Vincent Donnefort <vdonnefort@xxxxxxxxx>
Date:   Fri Feb 14 15:01:56 2014 +0100

    gpio: ich: Add support for multiple register addresses

    This patch introduces regs and reglen pointers which allow a chipset to have
    register addresses differing from ICH ones.

    Acked-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
    Signed-off-by: Vincent Donnefort <vdonnefort@xxxxxxxxx>
    Signed-off-by: Lee Jones <lee.jones@xxxxxxxxxx>


The relevant excerpts from the mixed listing are:

0000000000000110 <ichx_gpio_probe>:

<...snip...>

	for (i = 0; i < ARRAY_SIZE(ichx_priv.desc->regs[0]); i++) {
		if (!(use_gpio & (1 << i)))
			continue;
		if (!request_region(
 380:	48 8b 05 00 00 00 00 	mov    0x0(%rip),%rax        # 387 <ichx_gpio_probe+0x277>
			383: R_X86_64_PC32	.bss+0xb4
 387:	45 31 c0             	xor    %r8d,%r8d
 38a:	48 c7 c7 00 00 00 00 	mov    $0x0,%rdi
			38d: R_X86_64_32S	ioport_resource
 391:	48 8b 48 08          	mov    0x8(%rax),%rcx
 395:	48 8b 50 10          	mov    0x10(%rax),%rdx
 399:	48 63 c3             	movslq %ebx,%rax
 39c:	0f b6 34 01          	movzbl (%rcx,%rax,1),%esi       <===== FAULTING INSTN
 3a0:	4c 89 c9             	mov    %r9,%rcx
 3a3:	0f b6 14 1a          	movzbl (%rdx,%rbx,1),%edx
 3a7:	49 03 75 00          	add    0x0(%r13),%rsi
 3ab:	4c 89 4d c8          	mov    %r9,-0x38(%rbp)
 3af:	e8 00 00 00 00       	callq  3b4 <ichx_gpio_probe+0x2a4>
			3b0: R_X86_64_PC32	__request_region-0x4
 3b4:	4c 8b 4d c8          	mov    -0x38(%rbp),%r9
 3b8:	48 85 c0             	test   %rax,%rax
 3bb:	0f 85 17 fe ff ff    	jne    1d8 <ichx_gpio_probe+0xc8>
	}
	return 0;

request_err:
	/* Clean up: release already requested regions, if any */
	for (i--; i >= 0; i--) {
 3c1:	41 83 ef 01          	sub    $0x1,%r15d
 3c5:	41 83 ff ff          	cmp    $0xffffffff,%r15d
 3c9:	0f 84 d1 00 00 00    	je     4a0 <ichx_gpio_probe+0x390>
		if (!(use_gpio & (1 << i)))
 3cf:	45 0f a3 fe          	bt     %r15d,%r14d
 3d3:	73 ec                	jae    3c1 <ichx_gpio_probe+0x2b1>


Regards,
Peter Hurley
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/