[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <56570432.40500@redhat.com>
Date: Thu, 26 Nov 2015 14:08:02 +0100
From: Tomas Henzl <thenzl@...hat.com>
To: Ching Huang <ching2048@...ca.com.tw>
Cc: hch@...radead.org, jbottomley@...allels.com,
dan.carpenter@...cle.com, agordeev@...hat.com,
linux-scsi@...r.kernel.org, linux-kernel@...r.kernel.org,
hare@...e.de, jthumshirn@...e.de, hch@....de
Subject: Re: [PATCH v3 2/5] arcmsr: fixes not release allocated resource
On 26.11.2015 08:32, Ching Huang wrote:
> On Wed, 2015-11-25 at 14:55 +0100, Tomas Henzl wrote:
>> On 25.11.2015 12:41, Ching Huang wrote:
>>> From: Ching Huang <ching2048@...ca.com.tw>
>>>
>>> Releasing allocated resource if get configuration data failed.
>>>
>>> Signed-of-by: Ching Huang <ching2048@...ca.com.tw>
>>>
>>> ---
>>>
>>> diff -uprN a/drivers/scsi/arcmsr/arcmsr_hba.c b/drivers/scsi/arcmsr/arcmsr_hba.c
>>> --- a/drivers/scsi/arcmsr/arcmsr_hba.c 2015-11-24 11:35:26.000000000 +0800
>>> +++ b/drivers/scsi/arcmsr/arcmsr_hba.c 2015-11-25 19:04:44.590970000 +0800
>>> @@ -2664,7 +2664,7 @@ static bool arcmsr_hbaB_get_config(struc
>>> if (!arcmsr_hbaB_wait_msgint_ready(acb)) {
>>> printk(KERN_NOTICE "arcmsr%d: wait 'get adapter firmware \
>>> miscellaneous data' timeout \n", acb->host->host_no);
>>> - return false;
>>> + goto err_free_dma;
>>> }
>>> count = 8;
>>> while (count){
>>> @@ -2707,6 +2707,10 @@ static bool arcmsr_hbaB_get_config(struc
>>> acb->firm_cfg_version = readl(®->message_rwbuffer[25]); /*firm_cfg_version,25,100-103*/
>>> /*firm_ide_channels,4,16-19*/
>>> return true;
>>> +err_free_dma:
>>> + dma_free_coherent(&acb->pdev->dev, acb->roundup_ccbsize,
>>> + acb->dma_coherent2, acb->dma_coherent_handle2);
>>> + return false;
>>> }
>>>
>> The resource should be released here, that is okay.
>>
>> How works the resource management when the eh_bus_reset_handler
>> is called ? It looks to me that you allocate a new
>> acb->dma_coherent2 without releasing it before.
>> Btw. is it needed in that handler to re-read the firmware-spec?
>>
>> Tomas
>>
> You are right. In eh_bus_reset_handler, there may re-allocate a new dma resource if call get_config again.
> I will initiate a new patches for this bug after current 5 patches.
> Thanks Tomas.
ok, so for this patch -
Reviewed-by: Tomas Henzl <thenzl@...hat.com>
Tomas
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists