linux-kernel - Re: [PATCH] x86: eas should not be NULL when it is referenced

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: <7d4862ae-7ac6-11e8-5c8d-74610eabd5b5@linux.intel.com>
Date:   Fri, 25 Jun 2021 10:11:32 -0400
From:   "Liang, Kan" <kan.liang@...ux.intel.com>
To:     Alexander Antonov <alexander.antonov@...ux.intel.com>,
        13145886936@....com, tglx@...utronix.de, bp@...en8.de,
        x86@...nel.org
Cc:     linux-perf-users@...r.kernel.org, linux-kernel@...r.kernel.org,
        gushengxian <gushengxian@...ong.com>,
        Peter Zijlstra <peterz@...radead.org>
Subject: Re: [PATCH] x86: eas should not be NULL when it is referenced



On 6/25/2021 9:33 AM, Alexander Antonov wrote:
> Hello Kan,
>> On 6/24/2021 3:03 PM, Liang, Kan wrote:
>>> I think the NULL pointer dereference of eas should not happen, 
>>> because die is -1 if eas is NULL. But the whole error handling path 
>>> looks fragile.
>>>
>>> We already fixed one issue caused by it in commit ID f797f05d917f 
>>> ("perf/x86/intel/uncore: Fix for iio mapping on Skylake Server")
>>> https://lore.kernel.org/lkml/160149233331.7002.10919231011379055356.tip-bot2@tip-bot2/ 
>>>
>>>
>>> Maybe something as below?
>>>
>>>  From 3de81ba3b04262ef3346297d82f6c4ffb4af7029 Mon Sep 17 00:00:00 2001
>>> From: Kan Liang <kan.liang@...ux.intel.com>
>>> Date: Thu, 24 Jun 2021 11:17:57 -0700
>>> Subject: [PATCH] perf/x86/intel/uncore: Clean up error handling path 
>>> of iio mapping
>>>
>>> The error handling path of iio mapping looks fragile. We already fixed
>>> one issue caused by it, commit ID f797f05d917f ("perf/x86/intel/uncore:
>>> Fix for iio mapping on Skylake Server"). Clean up the error handling
>>> path and make the code robust.
> I didn't catch why does the current error handling path look fragile?
> Are there cases when it works incorrect?
>


I don't think it causes any severe problem for now, e.g., crash, because 
current code checks die before the dereference.
But I think it violates the Linux kernel coding style (one err bug) and 
may bring potential issues.

https://www.kernel.org/doc/html/v4.10/process/coding-style.html

Thanks,
Kan

> Thanks,
> Alexander
>>>
>>> Reported-by: gushengxian <gushengxian@...ong.com>
>>> Signed-off-by: Kan Liang <kan.liang@...ux.intel.com>
>>> ---
>>>   arch/x86/events/intel/uncore_snbep.c | 6 ++++--
>>>   1 file changed, 4 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/arch/x86/events/intel/uncore_snbep.c 
>>> b/arch/x86/events/intel/uncore_snbep.c
>>> index 7622762..6d4a5a9 100644
>>> --- a/arch/x86/events/intel/uncore_snbep.c
>>> +++ b/arch/x86/events/intel/uncore_snbep.c
>>> @@ -3802,11 +3802,11 @@ pmu_iio_set_mapping(struct intel_uncore_type 
>>> *type, struct attribute_group *ag)
>>>       /* One more for NULL. */
>>>       attrs = kcalloc((uncore_max_dies() + 1), sizeof(*attrs), 
>>> GFP_KERNEL);
>>>       if (!attrs)
>>> -        goto err;
>>> +        goto clear_topology;
>>>
>>>       eas = kcalloc(uncore_max_dies(), sizeof(*eas), GFP_KERNEL);
>>>       if (!eas)
>>> -        goto err;
>>> +        goto clear_attrs;
>>>
>>>       for (die = 0; die < uncore_max_dies(); die++) {
>>>           sprintf(buf, "die%ld", die);
>>> @@ -3827,7 +3827,9 @@ pmu_iio_set_mapping(struct intel_uncore_type 
>>> *type, struct attribute_group *ag)
>>>       for (; die >= 0; die--)
>>>           kfree(eas[die].attr.attr.name);
>>>       kfree(eas);
>>> +clear_attrs:
>>>       kfree(attrs);
>>> +clear_topology:
>>>       kfree(type->topology);
>>>   clear_attr_update:
>>>       type->attr_update = NULL;