Re: [PATCH] cxl/region: Fix leakage in __construct_region()

[Davidlohr Bueso <dave@stgolabs.net>]

On Mon, 09 Feb 2026, Ira Weiny wrote:

>Davidlohr Bueso wrote:
>> Failing the first sysfs_update_group() needs to explicitly
>> kfree the resource as it is too early for cxl_region_iomem_release()
>> to do so.
>>
>> Signed-off-by: Davidlohr Bueso <dave@stgolabs.net>
>> ---
>>  drivers/cxl/core/region.c | 4 +++-
>>  1 file changed, 3 insertions(+), 1 deletion(-)
>>
>> diff --git a/drivers/cxl/core/region.c b/drivers/cxl/core/region.c
>> index 96888d87a8df..28d70497283d 100644
>> --- a/drivers/cxl/core/region.c
>> +++ b/drivers/cxl/core/region.c
>> @@ -3795,8 +3795,10 @@ static int __construct_region(struct cxl_region *cxlr,
>>	}
>>
>>	rc = sysfs_update_group(&cxlr->dev.kobj, &cxl_region_group);
>> -	if (rc)
>> +	if (rc) {
>> +		kfree(res);
>>		return rc;
>> +	}
>>
>>	rc = insert_resource(cxlrd->res, res);
>>	if (rc) {
>
>There is another call to sysfs_update_group() with potential failure after
>the insert_resource().  Does that need to be handled?  Or is the resource
>turned over to the root decoder at that point?
>
>Assuming it is ok.

Yes, the second call failure is after the region param p->res = res
so taken care of in cxl_region_iomem_release().

>Reviewed-by: Ira Weiny <ira.weiny@intel.com>

Thanks!