[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <30e9c7d4-75c6-8cbc-7a27-d406eae01dad@huawei.com>
Date: Wed, 19 May 2021 11:36:22 +0800
From: Yang Yingliang <yangyingliang@...wei.com>
To: John Garry <john.garry@...wei.com>, <linux-kernel@...r.kernel.org>,
<linux-scsi@...r.kernel.org>
CC: <jejb@...ux.ibm.com>, <martin.petersen@...cle.com>,
chenxiang <chenxiang66@...ilicon.com>,
"luojiaxing@...wei.com" <luojiaxing@...wei.com>
Subject: Re: [PATCH -next] scsi: hisi_sas: drop free_irq of devm_request_irq
allocated irq
On 2021/5/18 23:34, John Garry wrote:
> On 18/05/2021 14:09, Yang Yingliang wrote:
>> irq allocated with devm_request_irq should not be freed using
>> free_irq, because doing so causes a dangling pointer, and a
>> subsequent double free.
>>
>> Reported-by: Hulk Robot <hulkci@...wei.com>
>> Signed-off-by: Yang Yingliang <yangyingliang@...wei.com>
>> ---
>> drivers/scsi/hisi_sas/hisi_sas_v3_hw.c | 6 +++---
>> 1 file changed, 3 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
>> b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
>> index 499c770d405c..684f762bcfb3 100644
>> --- a/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
>> +++ b/drivers/scsi/hisi_sas/hisi_sas_v3_hw.c
>> @@ -4811,9 +4811,9 @@ hisi_sas_v3_destroy_irqs(struct pci_dev *pdev,
>> struct hisi_hba *hisi_hba)
>> {
>> int i;
>> - free_irq(pci_irq_vector(pdev, 1), hisi_hba);
>> - free_irq(pci_irq_vector(pdev, 2), hisi_hba);
>> - free_irq(pci_irq_vector(pdev, 11), hisi_hba);
>> + devm_free_irq(&pdev->dev, pci_irq_vector(pdev, 1), hisi_hba);
>> + devm_free_irq(&pdev->dev, pci_irq_vector(pdev, 2), hisi_hba);
>> + devm_free_irq(&pdev->dev, pci_irq_vector(pdev, 11), hisi_hba);
>> for (i = 0; i < hisi_hba->cq_nvecs; i++) {
>> struct hisi_sas_cq *cq = &hisi_hba->cq[i];
>> int nr = hisi_sas_intr_conv ? 16 : 16 + i;
>>
>
> Does the free_irq(pci_irq_vector(pdev, nr, cq)) call also need to
> change (not shown)?
Yes, I missed that, it should be changed too.
>
> Having said that, why have these at all if we use devm_request_irq()?
> devm_irq_release() calls free_irq().
I keep the original logic here, only avoid double free.
>
> Thanks,
> John
>
> .
Powered by blists - more mailing lists