linux-kernel - Re: [PATCH] lkdtm: use the return value of strim()

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: <953b0290-a703-2bd8-179b-01442f736768@huaweicloud.com>
Date:   Fri, 18 Aug 2023 10:39:06 +0800
From:   "GONG, Ruiqi" <gongruiqi@...weicloud.com>
To:     Kees Cook <keescook@...omium.org>
Cc:     Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        linux-kernel@...r.kernel.org, gongruiqi1@...wei.com
Subject: Re: [PATCH] lkdtm: use the return value of strim()


On 2023/08/18 0:55, Kees Cook wrote:
> On Thu, Aug 17, 2023 at 10:21:17PM +0800, GONG, Ruiqi wrote:
>> From: "GONG, Ruiqi" <gongruiqi1@...wei.com>
>>
>> Make use of the return value of strim() to achieve left-trim as well as
>> right-trim, which prevents the following unusual fail case:
>>
>>  # echo " EXCEPTION" > /sys/kernel/debug/provoke-crash/DIRECT
>>  sh: write error: Invalid argument
>>
>> Link: https://github.com/KSPP/linux/issues/337
>> Signed-off-by: GONG, Ruiqi <gongruiqi1@...wei.com>
>> ---
>>  drivers/misc/lkdtm/core.c | 4 ++--
>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/misc/lkdtm/core.c b/drivers/misc/lkdtm/core.c
>> index 0772e4a4757e..812c96461ab2 100644
>> --- a/drivers/misc/lkdtm/core.c
>> +++ b/drivers/misc/lkdtm/core.c
>> @@ -242,7 +242,7 @@ static ssize_t lkdtm_debugfs_entry(struct file *f,
>>  	}
>>  	/* NULL-terminate and remove enter */
>>  	buf[count] = '\0';
>> -	strim(buf);
>> +	buf = strim(buf);
>>  
>>  	crashtype = find_crashtype(buf);
>>  	free_page((unsigned long)buf);
> 
> Will free_page() still work in this case, though? The address won't
> match the allocation any more...

Yes I noticed that, but I was under the impression that it's fine to do
free_page(paddr + offset_in_page) since the offset is within the page,
and its corresponding struct page can still be found when being freed.
Please let me know if this thought is wrong and I will submit another
version.

> 
> -Kees
> 
>> @@ -318,7 +318,7 @@ static ssize_t direct_entry(struct file *f, const char __user *user_buf,
>>  	}
>>  	/* NULL-terminate and remove enter */
>>  	buf[count] = '\0';
>> -	strim(buf);
>> +	buf = strim(buf);
>>  
>>  	crashtype = find_crashtype(buf);
>>  	free_page((unsigned long) buf);
>> -- 
>> 2.25.1
>>
>