linux-kernel - Re: parallel panel display stopped working

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Fri, 9 Jul 2021 18:45:06 +0300
From:   Andy Shevchenko <andy.shevchenko@...il.com>
To:     Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
        Miguel Ojeda <ojeda@...nel.org>,
        Lars Poeschel <poeschel@...onage.de>
Cc:     Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Geert Uytterhoeven <geert.uytterhoeven@...il.com>
Subject: Re: parallel panel display stopped working

On Fri, Jul 9, 2021 at 5:54 PM Andy Shevchenko
<andy.shevchenko@...il.com> wrote:
>
> On Fri, Jul 9, 2021 at 5:40 PM Andy Shevchenko
> <andy.shevchenko@...il.com> wrote:
> >
> > Parallel (4 bits)  panel display stopped working.
> > I have no time right now to bisect but I have noticed there was quite
> > a pile of patches from Lars to the driver recently (yeah, it seems for
> > a few months I haven't run my tests against the AUX display driver).
> >
> > I believe the issue is there somewhere.
> >
> > Moreover on the removal I have got an oops
> >
> > [  218.825445] kernel tried to execute NX-protected page - exploit
> > attempt? (uid: 0)
> > [  218.832965] BUG: unable to handle page fault for address: ffff8f8f06559dc0
> > [  218.839863] #PF: supervisor instruction fetch in kernel mode
> > [  218.845540] #PF: error_code(0x0011) - permissions violation
> > [  218.851132] PGD e601067 P4D e601067 PUD e602067 PMD 645a063 PTE
> > 8000000006559063
> > [  218.858587] Oops: 0011 [#1] SMP PTI
> > [  218.862099] CPU: 0 PID: 7 Comm: kworker/u4:0 Tainted: G         C
> >      5.13.0+ #23
> > [  218.869870] Hardware name: Intel Corporation Merrifield/BODEGA BAY,
> > BIOS 542 2015.01.21:18.19.48
> > [  218.878681] Workqueue: kacpi_hotplug acpi_device_del_work_fn
> > [  218.884380] RIP: 0010:0xffff8f8f06559dc0
> > [  218.888328] Code: ff ff d2 2b 21 8c ff ff ff ff 08 00 00 00 00 00
> > 00 00 78 34 fa 02 8f 8f ff ff 00 00 00 00 00
> > 00 00 00 00 00 00 00 00 00 00 00 <72> 65 67 75 6c 61 74 6f 72 3a 72 65
> > 67 75 6c 61 74 6f 72 2e 30 2d
> > [  218.907138] RSP: 0000:ffffad36c0043c90 EFLAGS: 00010246
> > [  218.912387] RAX: ffff8f8f06559dc0 RBX: ffff8f8f062cbc00 RCX: ffff8f8f01239fc8
> > [  218.919542] RDX: 000000002a3cccf8 RSI: 0000000000000001 RDI: ffff8f8f06559480
> > [  218.926701] RBP: ffffffffc037e279 R08: 00000000d5832520 R09: 0000000000000001
> > [  218.933856] R10: 0000000000000000 R11: 0000000000000001 R12: ffff8f8f062cbc00
> > [  218.941010] R13: ffffffffc038e028 R14: ffffffff8c5e0b60 R15: 00000000fffffffd
> > [  218.948166] FS:  0000000000000000(0000) GS:ffff8f8f3e200000(0000)
> > knlGS:0000000000000000
> > [  218.956286] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > [  218.962053] CR2: ffff8f8f06559dc0 CR3: 0000000002ef4000 CR4: 00000000001006f0
> > [  218.969208] Call Trace:
> > [  218.971678]  ? hd44780_common_clear_display+0x17/0x30 [hd44780_common]
> > [  218.978252]  ? charlcd_write_char+0x21a/0x810 [charlcd]
> > [  218.983519]  ? charlcd_puts+0x30/0x60 [charlcd]
> > [  218.988083]  ? charlcd_unregister+0x24/0x70 [charlcd]
> > [  218.993167]  ? hd44780_remove+0x1e/0x30 [hd44780]
> > [  218.997901]  ? platform_remove+0x1f/0x40
> > ...
> >
> > Just for your information and with hope that somebody may test and/or
> > reproduce this and fix it. Later on I might find time to bisect the
> > other issue I mentioned above.
>
> For time being confirmed the following:
> 1. a2060f288e15 ("auxdisplay: Call charlcd_backlight in place") has a
> remove bug (above), while panel works
> 2. 66ce7d5c1e12 ("auxdisplay: Use an enum for charlcd backlight on/off
> ops") everything works as expected
>
> So, the oops introduced in between those two.

The culprit (*) is 01ec46dfa633 ("auxdisplay: Move init_display to
hd44780_common").

*) Obviously it's not. It reveals the design issue on how the modules
were split. How this had been tested if tested at all?

P.S. If we don't get a fix in time, I will ask Miguel to revert the entire pile.

-- 
With Best Regards,
Andy Shevchenko