linux-kernel - Re:Re: [PATCH] firmware: arm_scpi: Fix error handle when scpi probe failed

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: <1752bac0.5d05.181b93873da.Coremail.15815827059@163.com>
Date:   Fri, 1 Jul 2022 18:04:53 +0800 (CST)
From:   huhai <15815827059@....com>
To:     "Sudeep Holla" <sudeep.holla@....com>
Cc:     cristian.marussi@....com, linux-arm-kernel@...ts.infradead.org,
        linux-kernel@...r.kernel.org, luriwen@...inos.cn,
        liuyun01@...inos.cn, huhai <huhai@...inos.cn>,
        k2ci <kernel-bot@...inos.cn>
Subject: Re:Re: [PATCH] firmware: arm_scpi: Fix error handle when scpi probe
 failed


>On Fri, Jul 01, 2022 at 02:16:06PM +0800, huhai wrote:
>> From: huhai <huhai@...inos.cn>
>>
>> When scpi probe fails, do not just return the error code, but also reset
>> the global scpi_info to NULL, otherwise scpi_hwmon_probe() may get a UAF
>> and cause panic:
>>
>
>Interesting, which platform are you using this on ? SCPI is deprecated for
>a while, please switch to SCMI which is well maintained both in terms of
>specification and support in the kernel. I also assume this is 64-bit
>platform, so I don't want you to get stuck in future because of lack

>of some feature in SCPI. Please see if you can migrate to SCMI.


Thinks for sharing this massage, and my platform is FT2000+(ARM64).

>
>>   scpi_protocol FTSC0001:00: incorrect or no SCP firmware found
>>   ... ...
>>   Unable to handle kernel NULL pointer dereference at virtual address 0000000000000000
>>   Mem abort info:
>>
>
>[...]
>

>I think we don't see to make it complex. Can't it be as simple as:

scpi_info is a global variable， we must set it to NULL when it is not valid

>Regards,
>Sudeep
>
>-->8
>
>diff --git i/drivers/firmware/arm_scpi.c w/drivers/firmware/arm_scpi.c
>index ddf0b9ff9e15..6fa1a5b193b8 100644
>--- i/drivers/firmware/arm_scpi.c
>+++ w/drivers/firmware/arm_scpi.c
>@@ -799,7 +799,7 @@ static struct scpi_ops scpi_ops = {
>
> struct scpi_ops *get_scpi_ops(void)
> {
>-       return scpi_info ? scpi_info->scpi_ops : NULL;

>+       return scpi_info && scpi_info->scpi_ops ? scpi_info->scpi_ops : NULL;


I don't think it work well, because it's a UAF and scpi_info->scpi_ops could be any value set by others.

> }
> EXPORT_SYMBOL_GPL(get_scpi_ops);