linux-kernel - Re: [PATCH] crypto: qat - Fix missing destroy_workqueue in adf_init

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: <1094e5d9-208c-4c41-83cd-5a77081aa5c1@huawei.com>
Date: Fri, 25 Oct 2024 09:24:52 +0800
From: Wang Hai <wanghai38@...wei.com>
To: "Cabiddu, Giovanni" <giovanni.cabiddu@...el.com>
CC: <herbert@...dor.apana.org.au>, <davem@...emloft.net>,
	<ahsan.atta@...el.com>, <mun.chun.yep@...el.com>,
	<markas.rapoportas@...el.com>, <damian.muszynski@...el.com>,
	<furong.zhou@...el.com>, <adam.guerin@...el.com>, <zhangxiaoxu5@...wei.com>,
	<qat-linux@...el.com>, <linux-crypto@...r.kernel.org>,
	<linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] crypto: qat - Fix missing destroy_workqueue in
 adf_init_aer()



On 2024/10/24 23:04, Cabiddu, Giovanni wrote:
> On Thu, Oct 24, 2024 at 10:00:57PM +0800, Wang Hai wrote:
>> The adf_init_aer() won't destroy device_reset_wq when alloc_workqueue()
>> for device_sriov_wq failed. Add destroy_workqueue for device_reset_wq to
>> fix this issue.
>>
>> Fixes: 4469f9b23468 ("crypto: qat - re-enable sriov after pf reset")
>> Signed-off-by: Wang Hai <wanghai38@...wei.com>
>> ---
>>   drivers/crypto/intel/qat/qat_common/adf_aer.c | 4 +++-
>>   1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/crypto/intel/qat/qat_common/adf_aer.c b/drivers/crypto/intel/qat/qat_common/adf_aer.c
>> index ec7913ab00a2..907144ec7e65 100644
>> --- a/drivers/crypto/intel/qat/qat_common/adf_aer.c
>> +++ b/drivers/crypto/intel/qat/qat_common/adf_aer.c
>> @@ -281,8 +281,10 @@ int adf_init_aer(void)
>>   		return -EFAULT;
>>   
>>   	device_sriov_wq = alloc_workqueue("qat_device_sriov_wq", 0, 0);
>> -	if (!device_sriov_wq)
>> +	if (!device_sriov_wq) {
>> +		destroy_workqueue(device_reset_wq);
> The missing destroy_workqueue() here is intentional as the device_reset_wq
> is destroyed in adf_exit_aer() which is called if adf_init_aer() fails,
> see [1].
> 
Hi, Giovanni.

Thanks for the review.

If adf_init_aer() fails, it will goto err_aer and then call
adf_exit_misc_wq() instead of goto err_pf_wq and then call
adf_exit_aer(). So this patch is needed.

static long adf_ctl_ioctl(struct file *fp, unsigned int cmd, unsigned 
long arg)
{
...
	if (adf_init_aer())
		goto err_aer;
...		
err_pf_wq:
	adf_exit_aer(); // will not be called when adf_init_aer() failed
err_aer:
	adf_exit_misc_wq();
err_misc_wq:
...
}
> With this change, destroy_workqueue() is called twice.
> 
> Regards,
>