linux-kernel - Re: [PATCH] nvdimm/pmem: Fix an error handling path in 'pmem_attach

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: <d791466d-56db-7dc3-342b-73db0de48b68@wanadoo.fr>
Date:   Sun, 7 Nov 2021 19:14:37 +0100
From:   Marion & Christophe JAILLET <christophe.jaillet@...adoo.fr>
To:     Ira Weiny <ira.weiny@...el.com>
Cc:     dan.j.williams@...el.com, vishal.l.verma@...el.com,
        dave.jiang@...el.com, nvdimm@...ts.linux.dev,
        linux-kernel@...r.kernel.org, kernel-janitors@...r.kernel.org
Subject: Re: [PATCH] nvdimm/pmem: Fix an error handling path in
 'pmem_attach_disk()'



Le 07/11/2021 à 18:25, Marion & Christophe JAILLET a écrit :
> 
> 
> Le 07/11/2021 à 18:20, Christophe JAILLET a écrit :
>> Le 07/11/2021 à 18:11, Ira Weiny a écrit :
>>> On Sat, Nov 06, 2021 at 06:27:11PM +0100, Christophe JAILLET wrote:
>>>> If 'devm_init_badblocks()' fails, a previous 'blk_alloc_disk()' call 
>>>> must
>>>> be undone.
>>>
>>> I think this is a problem...
>>>
>>>>
>>>> Signed-off-by: Christophe JAILLET <christophe.jaillet@...adoo.fr>
>>>> ---
>>>> This patch is speculative. Several fixes on error handling paths 
>>>> have been
>>>> done recently, but this one has been left as-is. There was maybe a good
>>>> reason that I have missed for that. So review with care!
>>>>
>>>> I've not been able to identify a Fixes tag that please me :(
>>>> ---
>>>>   drivers/nvdimm/pmem.c | 5 +++--
>>>>   1 file changed, 3 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/nvdimm/pmem.c b/drivers/nvdimm/pmem.c
>>>> index fe7ece1534e1..c37a1e6750b3 100644
>>>> --- a/drivers/nvdimm/pmem.c
>>>> +++ b/drivers/nvdimm/pmem.c
>>>> @@ -490,8 +490,9 @@ static int pmem_attach_disk(struct device *dev,
>>>>       nvdimm_namespace_disk_name(ndns, disk->disk_name);
>>>>       set_capacity(disk, (pmem->size - pmem->pfn_pad - 
>>>> pmem->data_offset)
>>>>               / 512);
>>>> -    if (devm_init_badblocks(dev, &pmem->bb))
>>>> -        return -ENOMEM;
>>>> +    rc = devm_init_badblocks(dev, &pmem->bb);
>>>> +    if (rc)
>>>> +        goto out;
>>>
>>> But I don't see this 'out' label in the function currently?  Was that 
>>> part of
>>> your patch missing?
>>
>> Hi,
>> the patch is based on the latest linux-next.
>> See [1]. The 'out' label exists there and is already used.
>>
>> In fact, I run an own-made coccinelle script which tries to spot 
>> mix-up between return and goto.
>> In this case, we have a 'return -ENOMEM' after a 'goto out' which 
>> looks spurious. Hence, my patch.
>>
>> [1]:https://git.kernel.org/pub/scm/linux/kernel/git/next/linux-next.git/tree/drivers/nvdimm/pmem.c#n512 
>>
> 
> Lol, the #n512 above is in fact another place that should be updated as 
> well. I missed it and only fixed #n494!

In fact, no, line 512 should be left as-is. The clean-up wilol be made 
by 'pmem_release_disk()'.

The patch attached at the very first mail of this thread looks good to me.

CJ

> 
> CJ
> 
>>
>> CJ
>>
>>>
>>> Ira
>>>
>>>>       nvdimm_badblocks_populate(nd_region, &pmem->bb, &bb_range);
>>>>       disk->bb = &pmem->bb;
>>>> -- 
>>>> 2.30.2
>>>>
>>>
>>
>>
>