Received-SPF: pass (google.com: domain of linux-kernel+bounces-139658-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45e3:2400::1 as permitted sender) client-ip=2604:1380:45e3:2400::1;
Subject: Re: [PATCH 2/2] ubi: ubi_init: Fix missed ubiblock cleanup in error
 handling path
To: Daniel Golle <daniel@makrotopia.org>
CC: <richard@nod.at>, <linux-mtd@lists.infradead.org>,
	<linux-kernel@vger.kernel.org>
References: <20240410074033.2523399-1-chengzhihao1@huawei.com>
 <20240410074033.2523399-3-chengzhihao1@huawei.com>
 <Zhccfw4HbC_MKj65@makrotopia.org>
From: Zhihao Cheng <chengzhihao1@huawei.com>
Message-ID: <5d858fbc-11cb-8d02-a4b6-a640625c7e1d@huawei.com>
Date: Thu, 11 Apr 2024 10:39:44 +0800
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:68.0) Gecko/20100101
 Thunderbird/68.5.0
Precedence: bulk
MIME-Version: 1.0
In-Reply-To: <Zhccfw4HbC_MKj65@makrotopia.org>
Content-Type: text/plain; charset="gbk"; format=flowed
Content-Transfer-Encoding: 8bit

?? 2024/4/11 7:10, Daniel Golle д??:
> Hi!
> 
> On Wed, Apr 10, 2024 at 03:40:33PM +0800, Zhihao Cheng wrote:
>> The ubiblock_init called by ubi_init will register device number, but
>> device number is not released in error handling path of ubi_init when
>> ubi is loaded by inserting module (eg. attaching failure), which leads
>> to subsequent ubi_init calls failed by running out of device number
>> (dmesg shows that "__register_blkdev: failed to get major for ubiblock").
>> Fix it by invoking ubiblock_exit() in corresponding error handling path.
> 
> Thank you for taking care of this issue.
> 
> See my comment inline below:
> 
>>
>> Fixes: 927c145208b0 ("mtd: ubi: attach from device tree")
>> Signed-off-by: Zhihao Cheng <chengzhihao1@huawei.com>
>> ---
>>   drivers/mtd/ubi/build.c | 5 +++--
>>   1 file changed, 3 insertions(+), 2 deletions(-)
>>
>> diff --git a/drivers/mtd/ubi/build.c b/drivers/mtd/ubi/build.c
>> index 7f95fd7968a8..354517194099 100644
>> --- a/drivers/mtd/ubi/build.c
>> +++ b/drivers/mtd/ubi/build.c
>> @@ -1380,12 +1380,13 @@ static int __init ubi_init(void)
>>   	if (ubi_is_module()) {
>>   		err = ubi_init_attach();
>>   		if (err)
>> -			goto out_mtd_notifier;
>> +			goto out_block_exit;
>>   	}
>>   
>>   	return 0;
>>   
>> -out_mtd_notifier:
>> +out_block_exit:
>> +	ubiblock_exit();
> 
> I believe that this call is the reason for the section mismatch we
> are seeing on Intel's kernel test builds:
> 
> https://lore.kernel.org/oe-kbuild-all/202404110656.wLLc5mHR-lkp@intel.com/
> 
> Also note that Ben Hutchings has supplied a more complete and imho
> better solution for this problem, which yet still suffers from the
> same problem (calling __exit function from __init function which
> results in section mismatch).
> 

My mistake, I always forget to check make W=1, I will send a new version 
to combine Ben's modifications. Thanks to point that.
> 
>>   	unregister_mtd_user(&ubi_mtd_notifier);
>>   out_debugfs:
>>   	ubi_debugfs_exit();
>> -- 
>> 2.39.2
>>
> .
>