2019-03-20 18:30:21

by Tom Zanussi

[permalink] [raw]
Subject: dell_smbios KASAN bug

Hi,

While looking into an unrelated problem, I hit this KASAN use-after-
free warning, so thought I'd let you know.

I have no idea how to fix it, but let me know if you need more info.

Thanks

Tom

[ 23.330893] ==================================================================
[ 23.330987] BUG: KASAN: use-after-free in dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 23.330999] Read of size 2 at addr ffff88840c2bc1a8 by task systemd-udevd/479

[ 23.331020] CPU: 0 PID: 479 Comm: systemd-udevd Not tainted 5.1.0-rc1+ #9
[ 23.331025] Hardware name: Dell Inc. XPS 13 9360/02PG84, BIOS 2.3.1 10/03/2017
[ 23.331030] Call Trace:
[ 23.331043] dump_stack+0x7c/0xbb
[ 23.331059] ? dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 23.331068] print_address_description+0xc7/0x280
[ 23.331080] ? dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 23.331090] ? dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 23.331101] kasan_report+0x14e/0x192
[ 23.331121] ? dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 23.331139] dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 23.331161] kbd_led_init+0x2e7/0x473 [dell_laptop]
[ 23.331178] ? dmi_matched+0x2a/0x2a [dell_laptop]
[ 23.331188] ? get_device_parent.isra.28+0x2a0/0x2a0
[ 23.331202] ? lockdep_init_map+0x98/0x2c0
[ 23.331229] ? platform_device_add+0x1b5/0x3a0
[ 23.331256] dell_init+0x4ad/0xb63 [dell_laptop]
[ 23.331271] ? kbd_led_init+0x473/0x473 [dell_laptop]
[ 23.331290] ? ___slab_alloc+0x61f/0x700
[ 23.331298] ? ___slab_alloc+0x61f/0x700
[ 23.331318] ? preempt_count_sub+0x15/0x100
[ 23.331339] ? kbd_led_init+0x473/0x473 [dell_laptop]
[ 23.331348] do_one_initcall+0xbd/0x3fd
[ 23.331359] ? perf_trace_initcall_level+0x280/0x280
[ 23.331369] ? kasan_unpoison_shadow+0x30/0x40
[ 23.331380] ? __kasan_kmalloc.constprop.8+0xa0/0xd0
[ 23.331397] ? kmem_cache_alloc_trace+0x163/0x390
[ 23.331405] ? kasan_unpoison_shadow+0x30/0x40
[ 23.331428] do_init_module+0xe3/0x341
[ 23.331447] load_module+0x2fc5/0x3ad0
[ 23.331528] ? layout_and_allocate+0x1170/0x1170
[ 23.331541] ? vfs_read+0xd4/0x1b0
[ 23.331558] ? kernel_read+0x74/0xa0
[ 23.331577] ? kernel_read_file+0x148/0x320
[ 23.331614] ? seccomp_notify_release+0x110/0x110
[ 23.331652] ? __do_sys_finit_module+0x192/0x1c0
[ 23.331660] __do_sys_finit_module+0x192/0x1c0
[ 23.331670] ? __ia32_sys_init_module+0x40/0x40
[ 23.331697] ? syscall_trace_enter+0x184/0x5e0
[ 23.331739] ? mark_held_locks+0x1a/0x90
[ 23.331760] do_syscall_64+0x72/0x220
[ 23.331773] entry_SYSCALL_64_after_hwframe+0x49/0xbe
[ 23.331781] RIP: 0033:0x7fcb4f5f5a49
[ 23.331789] Code: 00 f3 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 0f b4 2c 00 f7 d8 64 89 01 48
[ 23.331794] RSP: 002b:00007ffc73e340b8 EFLAGS: 00000246 ORIG_RAX: 0000000000000139
[ 23.331802] RAX: ffffffffffffffda RBX: 00005599992bb850 RCX: 00007fcb4f5f5a49
[ 23.331808] RDX: 0000000000000000 RSI: 00007fcb4f2e11c5 RDI: 0000000000000010
[ 23.331813] RBP: 00007fcb4f2e11c5 R08: 0000000000000000 R09: 00005599992bb850
[ 23.331819] R10: 0000000000000010 R11: 0000000000000246 R12: 0000000000000000
[ 23.331824] R13: 0000559999298f40 R14: 0000000000020000 R15: 00005599992bb850

[ 23.331873] Allocated by task 479:
[ 23.331886] __kasan_kmalloc.constprop.8+0xa0/0xd0
[ 23.331893] krealloc+0xa0/0xc0
[ 23.331900] 0xffffffffc0cc0075
[ 23.331909] dmi_decode_table+0xf6/0x140
[ 23.331915] dmi_walk+0x46/0x70
[ 23.331922] 0xffffffffc0cc0109
[ 23.331928] do_one_initcall+0xbd/0x3fd
[ 23.331935] do_init_module+0xe3/0x341
[ 23.331941] load_module+0x2fc5/0x3ad0
[ 23.331948] __do_sys_finit_module+0x192/0x1c0
[ 23.331954] do_syscall_64+0x72/0x220
[ 23.331961] entry_SYSCALL_64_after_hwframe+0x49/0xbe

[ 23.331975] Freed by task 479:
[ 23.331985] __kasan_slab_free+0x111/0x150
[ 23.331990] kfree+0xf5/0x350
[ 23.331996] 0xffffffffc0cc01d4
[ 23.332002] do_one_initcall+0xbd/0x3fd
[ 23.332009] do_init_module+0xe3/0x341
[ 23.332015] load_module+0x2fc5/0x3ad0
[ 23.332022] __do_sys_finit_module+0x192/0x1c0
[ 23.332028] do_syscall_64+0x72/0x220
[ 23.332035] entry_SYSCALL_64_after_hwframe+0x49/0xbe

[ 23.332050] The buggy address belongs to the object at ffff88840c2bc1a8
which belongs to the cache kmalloc-2k of size 2048
[ 23.332061] The buggy address is located 0 bytes inside of
2048-byte region [ffff88840c2bc1a8, ffff88840c2bc9a8)
[ 23.332070] The buggy address belongs to the page:
[ 23.332081] page:ffffea001030ae00 count:1 mapcount:0 mapping:ffff8884204113c0 index:0x0 compound_mapcount: 0
[ 23.332091] flags: 0x17ffffc0010200(slab|head)
[ 23.332100] raw: 0017ffffc0010200 ffffea0010367608 ffffea000ea31808 ffff8884204113c0
[ 23.332106] raw: 0000000000000000 00000000000d000d 00000001ffffffff 0000000000000000
[ 23.332111] page dumped because: kasan: bad access detected

[ 23.332124] Memory state around the buggy address:
[ 23.332134] ffff88840c2bc080: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 23.332145] ffff88840c2bc100: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 23.332155] >ffff88840c2bc180: fc fc fc fc fc fb fb fb fb fb fb fb fb fb fb fb
[ 23.332164] ^
[ 23.332175] ffff88840c2bc200: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[ 23.332185] ffff88840c2bc280: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[ 23.332194] ==================================================================


2019-03-20 18:42:43

by Steven Rostedt

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

On Wed, 20 Mar 2019 13:29:20 -0500
Tom Zanussi <[email protected]> wrote:

> Hi,
>
> While looking into an unrelated problem, I hit this KASAN use-after-
> free warning, so thought I'd let you know.
>
> I have no idea how to fix it, but let me know if you need more info.
>

Could you run with debug in the kernel command line, and see if you hit
any failed messages from the dell_smbios_init() call?


-- Steve

2019-03-20 19:07:30

by Tom Zanussi

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

On Wed, 2019-03-20 at 14:41 -0400, Steven Rostedt wrote:
> On Wed, 20 Mar 2019 13:29:20 -0500
> Tom Zanussi <[email protected]> wrote:
>
> > Hi,
> >
> > While looking into an unrelated problem, I hit this KASAN use-
> > after-
> > free warning, so thought I'd let you know.
> >
> > I have no idea how to fix it, but let me know if you need more
> > info.
> >
>
> Could you run with debug in the kernel command line, and see if you
> hit
> any failed messages from the dell_smbios_init() call?
>

Not much, but this looks relevant:

[ 26.783749] dell_smbios: No SMBIOS backends available (wmi: -19,
smm: -19)
[ 26.963648] dell_smbios: No dell-smbios drivers are loaded

Tom

>
> -- Steve

2019-03-20 19:14:57

by Steven Rostedt

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

On Wed, 20 Mar 2019 14:05:49 -0500
Tom Zanussi <[email protected]> wrote:

> On Wed, 2019-03-20 at 14:41 -0400, Steven Rostedt wrote:
> > On Wed, 20 Mar 2019 13:29:20 -0500
> > Tom Zanussi <[email protected]> wrote:
> >
> > > Hi,
> > >
> > > While looking into an unrelated problem, I hit this KASAN use-
> > > after-
> > > free warning, so thought I'd let you know.
> > >
> > > I have no idea how to fix it, but let me know if you need more
> > > info.
> > >
> >
> > Could you run with debug in the kernel command line, and see if you
> > hit
> > any failed messages from the dell_smbios_init() call?
> >
>
> Not much, but this looks relevant:
>
> [ 26.783749] dell_smbios: No SMBIOS backends available (wmi: -19,
> smm: -19)
> [ 26.963648] dell_smbios: No dell-smbios drivers are loaded
>

And does this fix you problem?

-- Steve

diff --git a/drivers/platform/x86/dell-smbios-base.c b/drivers/platform/x86/dell-smbios-base.c
index 9dc282ed5a9e..c3825c674522 100644
--- a/drivers/platform/x86/dell-smbios-base.c
+++ b/drivers/platform/x86/dell-smbios-base.c
@@ -619,6 +619,7 @@ static int __init dell_smbios_init(void)

fail_platform_driver:
kfree(da_tokens);
+ da_num_tokens = 0;
return ret;
}


2019-03-20 19:31:32

by Tom Zanussi

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

On Wed, 2019-03-20 at 15:13 -0400, Steven Rostedt wrote:
> On Wed, 20 Mar 2019 14:05:49 -0500
> Tom Zanussi <[email protected]> wrote:
>
> > On Wed, 2019-03-20 at 14:41 -0400, Steven Rostedt wrote:
> > > On Wed, 20 Mar 2019 13:29:20 -0500
> > > Tom Zanussi <[email protected]> wrote:
> > >
> > > > Hi,
> > > >
> > > > While looking into an unrelated problem, I hit this KASAN use-
> > > > after-
> > > > free warning, so thought I'd let you know.
> > > >
> > > > I have no idea how to fix it, but let me know if you need more
> > > > info.
> > > >
> > >
> > > Could you run with debug in the kernel command line, and see if
> > > you
> > > hit
> > > any failed messages from the dell_smbios_init() call?
> > >
> >
> > Not much, but this looks relevant:
> >
> > [ 26.783749] dell_smbios: No SMBIOS backends available (wmi: -19,
> > smm: -19)
> > [ 26.963648] dell_smbios: No dell-smbios drivers are loaded
> >
>
> And does this fix you problem?
>
> -- Steve
>
> diff --git a/drivers/platform/x86/dell-smbios-base.c
> b/drivers/platform/x86/dell-smbios-base.c
> index 9dc282ed5a9e..c3825c674522 100644
> --- a/drivers/platform/x86/dell-smbios-base.c
> +++ b/drivers/platform/x86/dell-smbios-base.c
> @@ -619,6 +619,7 @@ static int __init dell_smbios_init(void)
>
> fail_platform_driver:
> kfree(da_tokens);
> + da_num_tokens = 0;
> return ret;
> }
>


Unfortunately, no.

[ 26.125995] dell_smbios: No SMBIOS backends available (wmi: -19, smm: -19)
[ 26.232716] systemd-journald[407]: Successfully sent stream file descriptor to service manager.
[ 26.242860] dell_smbios: No dell-smbios drivers are loaded
[ 26.243142] ==================================================================
[ 26.243241] BUG: KASAN: use-after-free in dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 26.243254] Read of size 2 at addr ffff8883bdf941a8 by task systemd-udevd/458

[ 26.243277] CPU: 1 PID: 458 Comm: systemd-udevd Not tainted 5.1.0-rc1+ #10
[ 26.243283] Hardware name: Dell Inc. XPS 13 9360/02PG84, BIOS 2.3.1 10/03/2017
[ 26.243288] Call Trace:
[ 26.243303] dump_stack+0x7c/0xbb
[ 26.243317] ? dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 26.243327] print_address_description+0xc7/0x280
[ 26.243339] ? dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 26.243350] ? dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 26.243359] kasan_report+0x14e/0x192
[ 26.243379] ? dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 26.243399] dell_smbios_find_token+0x2e/0x80 [dell_smbios]
[ 26.243421] kbd_led_init+0x2e7/0x473 [dell_laptop]
[ 26.243440] ? dmi_matched+0x2a/0x2a [dell_laptop]
[ 26.243451] ? get_device_parent.isra.28+0x2a0/0x2a0
[ 26.243466] ? lockdep_init_map+0x98/0x2c0
[ 26.243494] ? platform_device_add+0x1b5/0x3a0
[ 26.243525] dell_init+0x4ad/0xb63 [dell_laptop]
[ 26.243542] ? kbd_led_init+0x473/0x473 [dell_laptop]
[ 26.243563] ? ___slab_alloc+0x61f/0x700
[ 26.243572] ? ___slab_alloc+0x61f/0x700
[ 26.243594] ? preempt_count_sub+0x15/0x100
[ 26.243616] ? kbd_led_init+0x473/0x473 [dell_laptop]
[ 26.243626] do_one_initcall+0xbd/0x3fd
[ 26.243638] ? perf_trace_initcall_level+0x280/0x280
[ 26.243650] ? kasan_unpoison_shadow+0x30/0x40
[ 26.243662] ? __kasan_kmalloc.constprop.8+0xa0/0xd0
[ 26.243681] ? kmem_cache_alloc_trace+0x163/0x390
[ 26.243691] ? kasan_unpoison_shadow+0x30/0x40
[ 26.243716] do_init_module+0xe3/0x341
[ 26.243736] load_module+0x2fc5/0x3ad0
[ 26.243824] ? layout_and_allocate+0x1170/0x1170
[ 26.243837] ? vfs_read+0xd4/0x1b0
[ 26.243855] ? kernel_read+0x74/0xa0
[ 26.243877] ? kernel_read_file+0x148/0x320
[ 26.243917] ? seccomp_notify_release+0x110/0x110
[ 26.243958] ? __do_sys_finit_module+0x192/0x1c0
[ 26.243964] __do_sys_finit_module+0x192/0x1c0
[ 26.243975] ? __ia32_sys_init_module+0x40/0x40
[ 26.244000] ? syscall_trace_enter+0x184/0x5e0
[ 26.244046] ? mark_held_locks+0x1a/0x90
[ 26.244068] do_syscall_64+0x72/0x220
[ 26.244083] entry_SYSCALL_64_after_hwframe+0x49/0xbe
[ 26.244091] RIP: 0033:0x7f7ceda3aa49
[ 26.244100] Code: 00 f3 c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 40 00 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 0f b4 2c 00 f7 d8 64 89 01 48
[ 26.244105] RSP: 002b:00007ffe6ca1cbf8 EFLAGS: 00000246 ORIG_RAX: 0000000000000139
[ 26.244114] RAX: ffffffffffffffda RBX: 00005635100838f0 RCX: 00007f7ceda3aa49
[ 26.244121] RDX: 0000000000000000 RSI: 00007f7ced7261c5 RDI: 0000000000000010
[ 26.244127] RBP: 00007f7ced7261c5 R08: 0000000000000000 R09: 00005635100838f0
[ 26.244133] R10: 0000000000000010 R11: 0000000000000246 R12: 0000000000000000
[ 26.244139] R13: 0000563510089e90 R14: 0000000000020000 R15: 00005635100838f0

[ 26.244193] Allocated by task 458:
[ 26.244206] __kasan_kmalloc.constprop.8+0xa0/0xd0
[ 26.244214] krealloc+0xa0/0xc0
[ 26.244220] 0xffffffffc0d60075
[ 26.244228] dmi_decode_table+0xf6/0x140
[ 26.244235] dmi_walk+0x46/0x70
[ 26.244241] 0xffffffffc0d60109
[ 26.244248] do_one_initcall+0xbd/0x3fd
[ 26.244255] do_init_module+0xe3/0x341
[ 26.244261] load_module+0x2fc5/0x3ad0
[ 26.244269] __do_sys_finit_module+0x192/0x1c0
[ 26.244276] do_syscall_64+0x72/0x220
[ 26.244283] entry_SYSCALL_64_after_hwframe+0x49/0xbe

[ 26.244297] Freed by task 458:
[ 26.244309] __kasan_slab_free+0x111/0x150
[ 26.244316] kfree+0xf5/0x350
[ 26.244323] 0xffffffffc0d601d4
[ 26.244330] do_one_initcall+0xbd/0x3fd
[ 26.244337] do_init_module+0xe3/0x341
[ 26.244344] load_module+0x2fc5/0x3ad0
[ 26.244352] __do_sys_finit_module+0x192/0x1c0
[ 26.244358] do_syscall_64+0x72/0x220
[ 26.244366] entry_SYSCALL_64_after_hwframe+0x49/0xbe

[ 26.244381] The buggy address belongs to the object at ffff8883bdf941a8
which belongs to the cache kmalloc-2k of size 2048
[ 26.244393] The buggy address is located 0 bytes inside of
2048-byte region [ffff8883bdf941a8, ffff8883bdf949a8)
[ 26.244402] The buggy address belongs to the page:
[ 26.244413] page:ffffea000ef7e400 count:1 mapcount:0 mapping:ffff88841c0113c0 index:0xffff8883bdf90968 compound_mapcount: 0
[ 26.244423] flags: 0x17ffffc0010200(slab|head)
[ 26.244433] raw: 0017ffffc0010200 ffffea000eff8208 ffff88841c003200 ffff88841c0113c0
[ 26.244442] raw: ffff8883bdf90968 00000000000d0009 00000001ffffffff 0000000000000000
[ 26.244447] page dumped because: kasan: bad access detected

[ 26.244460] Memory state around the buggy address:
[ 26.244472] ffff8883bdf94080: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 26.244483] ffff8883bdf94100: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
[ 26.244494] >ffff8883bdf94180: fc fc fc fc fc fb fb fb fb fb fb fb fb fb fb fb
[ 26.244504] ^
[ 26.244515] ffff8883bdf94200: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[ 26.244526] ffff8883bdf94280: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
[ 26.244535] ==================================================================


2019-03-20 19:47:11

by Steven Rostedt

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

On Wed, 20 Mar 2019 14:30:21 -0500
Tom Zanussi <[email protected]> wrote:

> > diff --git a/drivers/platform/x86/dell-smbios-base.c
> > b/drivers/platform/x86/dell-smbios-base.c
> > index 9dc282ed5a9e..c3825c674522 100644
> > --- a/drivers/platform/x86/dell-smbios-base.c
> > +++ b/drivers/platform/x86/dell-smbios-base.c
> > @@ -619,6 +619,7 @@ static int __init dell_smbios_init(void)
> >
> > fail_platform_driver:
> > kfree(da_tokens);
> > + da_num_tokens = 0;
> > return ret;
> > }
> >
>
>
> Unfortunately, no.

That's really weird. I would recommend throwing in a bunch of printks
to see if things are being freed and set, to make sure this code is hit
and something else isn't just corrupting it.

-- Steve

2019-03-20 20:29:51

by Tom Zanussi

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

On Wed, 2019-03-20 at 15:45 -0400, Steven Rostedt wrote:
> On Wed, 20 Mar 2019 14:30:21 -0500
> Tom Zanussi <[email protected]> wrote:
>
> > > diff --git a/drivers/platform/x86/dell-smbios-base.c
> > > b/drivers/platform/x86/dell-smbios-base.c
> > > index 9dc282ed5a9e..c3825c674522 100644
> > > --- a/drivers/platform/x86/dell-smbios-base.c
> > > +++ b/drivers/platform/x86/dell-smbios-base.c
> > > @@ -619,6 +619,7 @@ static int __init dell_smbios_init(void)
> > >
> > > fail_platform_driver:
> > > kfree(da_tokens);
> > > + da_num_tokens = 0;
> > > return ret;
> > > }
> > >
> >
> >
> > Unfortunately, no.
>
> That's really weird. I would recommend throwing in a bunch of printks
> to see if things are being freed and set, to make sure this code is
> hit
> and something else isn't just corrupting it.
>

Actually the patch does make the KASAN bug go away - I'm so used to
building without modules that I forgot to rebuild the module after
applying the patch, sigh.

Tom

> -- Steve

2019-03-20 20:34:29

by Steven Rostedt

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

On Wed, 20 Mar 2019 15:28:56 -0500
Tom Zanussi <[email protected]> wrote:

> > > Unfortunately, no.
> >
> > That's really weird. I would recommend throwing in a bunch of printks
> > to see if things are being freed and set, to make sure this code is
> > hit
> > and something else isn't just corrupting it.
> >
>
> Actually the patch does make the KASAN bug go away - I'm so used to
> building without modules that I forgot to rebuild the module after
> applying the patch, sigh.

I wasn't about to say anything. That's why I suggested adding printks
and trying it again ;-)

-- Steve

2019-03-20 20:45:41

by Tom Zanussi

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

On Wed, 2019-03-20 at 16:32 -0400, Steven Rostedt wrote:
> On Wed, 20 Mar 2019 15:28:56 -0500
> Tom Zanussi <[email protected]> wrote:
>
> > > > Unfortunately, no.
> > >
> > > That's really weird. I would recommend throwing in a bunch of
> > > printks
> > > to see if things are being freed and set, to make sure this code
> > > is
> > > hit
> > > and something else isn't just corrupting it.
> > >
> >
> > Actually the patch does make the KASAN bug go away - I'm so used to
> > building without modules that I forgot to rebuild the module after
> > applying the patch, sigh.
>
> I wasn't about to say anything. That's why I suggested adding printks
> and trying it again ;-)
>

Yeah, when a simple printk doesn't work, you know there's something
idiotic going on. Thanks for the tip ;-)

Tom

> -- Steve

2019-03-25 13:32:43

by Mario Limonciello

[permalink] [raw]
Subject: RE: dell_smbios KASAN bug

> -----Original Message-----
> From: Steven Rostedt <[email protected]>
> Sent: Wednesday, March 20, 2019 2:14 PM
> To: Tom Zanussi
> Cc: Pali Roh?r; Limonciello, Mario; [email protected]
> Subject: Re: dell_smbios KASAN bug
>
>
> [EXTERNAL EMAIL]
>
> On Wed, 20 Mar 2019 14:05:49 -0500
> Tom Zanussi <[email protected]> wrote:
>
> > On Wed, 2019-03-20 at 14:41 -0400, Steven Rostedt wrote:
> > > On Wed, 20 Mar 2019 13:29:20 -0500
> > > Tom Zanussi <[email protected]> wrote:
> > >
> > > > Hi,
> > > >
> > > > While looking into an unrelated problem, I hit this KASAN use-
> > > > after-
> > > > free warning, so thought I'd let you know.
> > > >
> > > > I have no idea how to fix it, but let me know if you need more
> > > > info.
> > > >
> > >
> > > Could you run with debug in the kernel command line, and see if you
> > > hit
> > > any failed messages from the dell_smbios_init() call?
> > >
> >
> > Not much, but this looks relevant:
> >
> > [ 26.783749] dell_smbios: No SMBIOS backends available (wmi: -19,
> > smm: -19)
> > [ 26.963648] dell_smbios: No dell-smbios drivers are loaded
> >
>
> And does this fix you problem?
>
> -- Steve
>
> diff --git a/drivers/platform/x86/dell-smbios-base.c b/drivers/platform/x86/dell-
> smbios-base.c
> index 9dc282ed5a9e..c3825c674522 100644
> --- a/drivers/platform/x86/dell-smbios-base.c
> +++ b/drivers/platform/x86/dell-smbios-base.c
> @@ -619,6 +619,7 @@ static int __init dell_smbios_init(void)
>
> fail_platform_driver:
> kfree(da_tokens);
> + da_num_tokens = 0;
> return ret;
> }
>

Steven,

Since Tom confirmed this works will you send a formal patch to the ML with this fix?

Thanks,

2019-03-25 14:18:22

by Steven Rostedt

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

On Mon, 25 Mar 2019 13:22:35 +0000
<[email protected]> wrote:

> > diff --git a/drivers/platform/x86/dell-smbios-base.c b/drivers/platform/x86/dell-
> > smbios-base.c
> > index 9dc282ed5a9e..c3825c674522 100644
> > --- a/drivers/platform/x86/dell-smbios-base.c
> > +++ b/drivers/platform/x86/dell-smbios-base.c
> > @@ -619,6 +619,7 @@ static int __init dell_smbios_init(void)
> >
> > fail_platform_driver:
> > kfree(da_tokens);
> > + da_num_tokens = 0;
> > return ret;
> > }
> >
>
> Steven,
>
> Since Tom confirmed this works will you send a formal patch to the ML with this fix?

Will do.

Tom,

Is it OK if I add the following tags to that patch?

Reported-by: Tom Zanussi <[email protected]>
Tested-by: Tom Zanussi <[email protected]>

-- Steve

2019-03-25 16:22:07

by Tom Zanussi

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

Hi Steve,

On Mon, 2019-03-25 at 10:17 -0400, Steven Rostedt wrote:
> On Mon, 25 Mar 2019 13:22:35 +0000
> <[email protected]> wrote:
>
> > > diff --git a/drivers/platform/x86/dell-smbios-base.c
> > > b/drivers/platform/x86/dell-
> > > smbios-base.c
> > > index 9dc282ed5a9e..c3825c674522 100644
> > > --- a/drivers/platform/x86/dell-smbios-base.c
> > > +++ b/drivers/platform/x86/dell-smbios-base.c
> > > @@ -619,6 +619,7 @@ static int __init dell_smbios_init(void)
> > >
> > > fail_platform_driver:
> > > kfree(da_tokens);
> > > + da_num_tokens = 0;
> > > return ret;
> > > }
> > >
> >
> > Steven,
> >
> > Since Tom confirmed this works will you send a formal patch to the
> > ML with this fix?
>
> Will do.
>
> Tom,
>
> Is it OK if I add the following tags to that patch?
>
> Reported-by: Tom Zanussi <[email protected]>
> Tested-by: Tom Zanussi <[email protected]>
>

Yep, fine with me.

Tom

> -- Steve

2019-04-03 17:19:17

by Steven Rostedt

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

On Mon, 25 Mar 2019 11:20:47 -0500
Tom Zanussi <[email protected]> wrote:

> > Is it OK if I add the following tags to that patch?
> >
> > Reported-by: Tom Zanussi <[email protected]>
> > Tested-by: Tom Zanussi <[email protected]>
> >
>
> Yep, fine with me.

I modified the patch slightly. Can you test this too. There should be
no reason that it wont work as this is "safer" but I don't want to put
your tested-by if you haven't tested it ;-)

-- Steve

diff --git a/drivers/platform/x86/dell-smbios-base.c b/drivers/platform/x86/dell-smbios-base.c
index 0537d44d45a6..a74c0df25b15 100644
--- a/drivers/platform/x86/dell-smbios-base.c
+++ b/drivers/platform/x86/dell-smbios-base.c
@@ -625,6 +625,8 @@ static int __init dell_smbios_init(void)

fail_platform_driver:
kfree(da_tokens);
+ da_tokens = NULL;
+ da_num_tokens = 0;
return ret;
}

2019-04-03 19:15:26

by Tom Zanussi

[permalink] [raw]
Subject: Re: dell_smbios KASAN bug

Hi Steve,

On Wed, 2019-04-03 at 13:16 -0400, Steven Rostedt wrote:
> On Mon, 25 Mar 2019 11:20:47 -0500
> Tom Zanussi <[email protected]> wrote:
>
> > > Is it OK if I add the following tags to that patch?
> > >
> > > Reported-by: Tom Zanussi <[email protected]>
> > > Tested-by: Tom Zanussi <[email protected]>
> > >
> >
> > Yep, fine with me.
>
> I modified the patch slightly. Can you test this too. There should be
> no reason that it wont work as this is "safer" but I don't want to
> put
> your tested-by if you haven't tested it ;-)
>

This one works fine too, so my Tested-by still applies. ;-)

Tom

> -- Steve
>
> diff --git a/drivers/platform/x86/dell-smbios-base.c
> b/drivers/platform/x86/dell-smbios-base.c
> index 0537d44d45a6..a74c0df25b15 100644
> --- a/drivers/platform/x86/dell-smbios-base.c
> +++ b/drivers/platform/x86/dell-smbios-base.c
> @@ -625,6 +625,8 @@ static int __init dell_smbios_init(void)
>
> fail_platform_driver:
> kfree(da_tokens);
> + da_tokens = NULL;
> + da_num_tokens = 0;
> return ret;
> }
>