linux-kernel - Re: [RESEND PATCH v2] efi: Only print errors about failing to get certs if EFI vars are found

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]

Message-ID: <30819cad-1a00-1ea7-13cf-d1d15c0fa96c@redhat.com>
Date:   Fri, 28 Feb 2020 10:35:18 +0100
From:   David Hildenbrand <david@...hat.com>
To:     Javier Martinez Canillas <javierm@...hat.com>,
        linux-kernel@...r.kernel.org
Cc:     linux-efi@...r.kernel.org, Hans de Goede <hdegoede@...hat.com>,
        Eric Richter <erichte@...ux.ibm.com>,
        James Morris <jmorris@...ei.org>,
        Michael Ellerman <mpe@...erman.id.au>,
        Mimi Zohar <zohar@...ux.ibm.com>,
        Nayna Jain <nayna@...ux.ibm.com>,
        "Serge E. Hallyn" <serge@...lyn.com>,
        YueHaibing <yuehaibing@...wei.com>,
        linux-security-module@...r.kernel.org,
        Ard Biesheuvel <ardb@...nel.org>
Subject: Re: [RESEND PATCH v2] efi: Only print errors about failing to get
 certs if EFI vars are found

On 28.02.20 10:31, David Hildenbrand wrote:
> On 28.02.20 10:19, David Hildenbrand wrote:
>> On 17.02.20 12:39, Javier Martinez Canillas wrote:
>>> If CONFIG_LOAD_UEFI_KEYS is enabled, the kernel attempts to load the certs
>>> from the db, dbx and MokListRT EFI variables into the appropriate keyrings.
>>>
>>> But it just assumes that the variables will be present and prints an error
>>> if the certs can't be loaded, even when is possible that the variables may
>>> not exist. For example the MokListRT variable will only be present if shim
>>> is used.
>>>
>>> So only print an error message about failing to get the certs list from an
>>> EFI variable if this is found. Otherwise these printed errors just pollute
>>> the kernel log ring buffer with confusing messages like the following:
>>>
>>> [    5.427251] Couldn't get size: 0x800000000000000e
>>> [    5.427261] MODSIGN: Couldn't get UEFI db list
>>> [    5.428012] Couldn't get size: 0x800000000000000e
>>> [    5.428023] Couldn't get UEFI MokListRT
>>>
>>> Reported-by: Hans de Goede <hdegoede@...hat.com>
>>> Signed-off-by: Javier Martinez Canillas <javierm@...hat.com>
>>> Tested-by: Hans de Goede <hdegoede@...hat.com>
>>
>> This patch seems to break a very basic x86-64 QEMU setup (booting
>> upstream kernel with a F31 initrd - are you running basic boot tests?).
>> Luckily, it only took me 5 minutes to identify this patch. Reverting
>> this patch from linux-next fixes it for me.
>>
>>
>> [    1.042766] Loaded X.509 cert 'Build time autogenerated kernel key: 6625d6e34255935276d2c9851e2458909a4bcd69'
>> [    1.044314] zswap: loaded using pool lzo/zbud
>> [    1.045663] Key type ._fscrypt registered
>> [    1.046154] Key type .fscrypt registered
>> [    1.046524] Key type fscrypt-provisioning registered
>> [    1.051178] Key type big_key registered
>> [    1.055108] Key type encrypted registered
>> [    1.055513] BUG: kernel NULL pointer dereference, address: 0000000000000000
>> [    1.056172] #PF: supervisor instruction fetch in kernel mode
>> [    1.056706] #PF: error_code(0x0010) - not-present page
>> [    1.057367] PGD 0 P4D 0 
>> [    1.057729] Oops: 0010 [#1] SMP NOPTI
>> [    1.058249] CPU: 1 PID: 1 Comm: swapper/0 Not tainted 5.6.0-rc3-next-20200228+ #79
>> [    1.059167] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.13.0-0-gf21b5a4aeb02-prebuilt.qemu.4
>> [    1.060230] RIP: 0010:0x0
>> [    1.060478] Code: Bad RIP value.
>> [    1.060786] RSP: 0018:ffffbc7880637d98 EFLAGS: 00010246
>> [    1.061281] RAX: 0000000000000000 RBX: 0000000000000000 RCX: ffffbc7880637dc8
>> [    1.061954] RDX: 0000000000000000 RSI: ffffbc7880637df0 RDI: ffffffffa73c40be
>> [    1.062611] RBP: ffffbc7880637e20 R08: ffffbc7880637dac R09: ffffa0238f4ba6c0
>> [    1.063278] R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000000
>> [    1.063956] R13: ffffa024bdd6f660 R14: 0000000000000000 R15: 0000000000000000
>> [    1.064609] FS:  0000000000000000(0000) GS:ffffa023fdd00000(0000) knlGS:0000000000000000
>> [    1.065360] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
>> [    1.065900] CR2: ffffffffffffffd6 CR3: 00000000b1610000 CR4: 00000000000006e0
>> [    1.066562] Call Trace:
>> [    1.066803]  load_uefi_certs+0xc8/0x2bb
>> [    1.067171]  ? get_cert_list+0xfb/0xfb
>> [    1.067523]  do_one_initcall+0x5d/0x2f0
>> [    1.067894]  ? rcu_read_lock_sched_held+0x52/0x80
>> [    1.068337]  kernel_init_freeable+0x243/0x2c2
>> [    1.068751]  ? rest_init+0x23a/0x23a
>> [    1.069095]  kernel_init+0xa/0x106
>> [    1.069416]  ret_from_fork+0x27/0x50
>> [    1.069759] Modules linked in:
>> [    1.070050] CR2: 0000000000000000
>> [    1.070361] ---[ end trace fcce9bb4feb21d99 ]---
>>
>>
> 
> Sorry, wrong mail identified, the patch is actually
> 
> commit 6b75d54d5258ccd655387a00bbe1b00f92f4d965
> Author: Ard Biesheuvel <ardb@...nel.org>
> Date:   Sun Feb 16 19:46:25 2020 +0100
> 
>     integrity: Check properly whether EFI GetVariable() is available
> 
>     Testing the value of the efi.get_variable function pointer is not
> 
> which made it work. (not even able to find that patch on lkml ...)

To clarify for Ard, your patch breaks a basic QEMU setup (see above,
NULL pointer dereference). Reverting your patch from linux-next makes it
work again.


-- 
Thanks,

David / dhildenb