[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <a2735b36-b64c-41d5-80ef-4d0cb47df055@fujitsu.com>
Date: Thu, 9 May 2024 03:11:16 +0000
From: "Zhijian Li (Fujitsu)" <lizhijian@...itsu.com>
To: Dave Jiang <dave.jiang@...el.com>, Ira Weiny <ira.weiny@...el.com>,
"dave@...olabs.net" <dave@...olabs.net>, "jonathan.cameron@...wei.com"
<jonathan.cameron@...wei.com>, "alison.schofield@...el.com"
<alison.schofield@...el.com>, "vishal.l.verma@...el.com"
<vishal.l.verma@...el.com>, "dan.j.williams@...el.com"
<dan.j.williams@...el.com>, "linux-cxl@...r.kernel.org"
<linux-cxl@...r.kernel.org>
CC: "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] cxl/region: Fix memregion leaks in devm_cxl_add_region()
On 07/05/2024 23:52, Dave Jiang wrote:
>
>
> On 5/6/24 7:42 PM, Zhijian Li (Fujitsu) wrote:
>>
>>
>> On 30/04/2024 00:23, Ira Weiny wrote:
>>> Li Zhijian wrote:
>>>> Move memregion_free(id) out of cxl_region_alloc() and
>>>> explicately free memregion in devm_cxl_add_region() error path.
>>> ^^^^
>>> explicitly
>>>
>>> BTW you should run checkpatch.pl which will check spelling.
>>
>>
>> I remember I've done this check, but it seems the it doesn't work for me.
>> Did I miss something?
>>
>> $ ./scripts/checkpatch.pl 0001-cxl-region-Fix-memregion-leaks-in-devm_cxl_add_regio.patch
>> total: 0 errors, 0 warnings, 23 lines checked
>>
>> 0001-cxl-region-Fix-memregion-leaks-in-devm_cxl_add_regio.patch has no obvious style problems and is ready for submission.
>
> Pass in --codespell parameter. And make sure you have the codespell dictionary installed (i.e. /usr/share/codespell/dictionary.txt).
Good to know this, many thanks!
Thanks
>
> DJ
>
>>
>>
>>
>>>
>>> I'm not following what the problem is you are trying to fix. This seems
>>> like it just moves the memregion_free() around a bit but the logic is the
>>> same.
>>>
>>
>> Sorry, I think my commit log may be misleading. In fact, this patch mainly fixes
>> the memregion leak in devm_cxl_add_region() when it gets an invalid mode.
>>
>>>> default:
>>>> dev_err(&cxlrd->cxlsd.cxld.dev, "unsupported mode %d\n", mode);
>>>> + memregion_free(id);
>>>> return ERR_PTR(-EINVAL);
>>>> }
>>
>> Additionally, to maintain consistent error handling, I also moved memregion_free(id) from
>> cxl_region_alloc() to devm_cxl_add_region() so that devm_cxl_add_region() can
>> free memregion explicitly in error path.
>>
>>
>> Thanks
>> Zhijian
>>
>>
>>> Ira
>>>
>>>>
>>>> After cxl_region_alloc() succeed, memregion will be freed by
>>>> cxl_region_type.release()
>>>>
>>>> Fixes: 6e099264185d ("cxl/region: Add volatile region creation support")
>>>> Signed-off-by: Li Zhijian <lizhijian@...itsu.com>
>>>> ---
>>>> drivers/cxl/core/region.c | 9 +++++----
>>>> 1 file changed, 5 insertions(+), 4 deletions(-)
>>>>
>>>> diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c
>>>> index 812b2948b6c6..8535718a20f0 100644
>>>> --- a/drivers/cxl/core/region.c
>>>> +++ b/drivers/cxl/core/region.c
>>>> @@ -2250,10 +2250,8 @@ static struct cxl_region *cxl_region_alloc(struct cxl_root_decoder *cxlrd, int i
>>>> struct device *dev;
>>>>
>>>> cxlr = kzalloc(sizeof(*cxlr), GFP_KERNEL);
>>>> - if (!cxlr) {
>>>> - memregion_free(id);
>>>> + if (!cxlr)
>>>> return ERR_PTR(-ENOMEM);
>>>> - }
>>>>
>>>> dev = &cxlr->dev;
>>>> device_initialize(dev);
>>>> @@ -2358,12 +2356,15 @@ static struct cxl_region *devm_cxl_add_region(struct cxl_root_decoder *cxlrd,
>>>> break;
>>>> default:
>>>> dev_err(&cxlrd->cxlsd.cxld.dev, "unsupported mode %d\n", mode);
>>>> + memregion_free(id);
>>>> return ERR_PTR(-EINVAL);
>>>> }
>>>>
>>>> cxlr = cxl_region_alloc(cxlrd, id);
>>>> - if (IS_ERR(cxlr))
>>>> + if (IS_ERR(cxlr)) {
>>>> + memregion_free(id);
>>>> return cxlr;
>>>> + }
>>>> cxlr->mode = mode;
>>>> cxlr->type = type;
>>>>
>>>> --
>>>> 2.29.2
>>>>
>>>
>>>
Powered by blists - more mailing lists