linux-kernel - Re: [PATCH V4 16/17] cxl/pmem_region: Create pmem region using information parsed from LSA

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] [day] [month] [year] [list]

Message-ID: <700072760.81768145283611.JavaMail.epsvc@epcpadp2new>
Date: Fri, 9 Jan 2026 18:07:18 +0530
From: Neeraj Kumar <s.neeraj@...sung.com>
To: Dave Jiang <dave.jiang@...el.com>
Cc: linux-cxl@...r.kernel.org, nvdimm@...ts.linux.dev,
	linux-kernel@...r.kernel.org, gost.dev@...sung.com,
	a.manzanares@...sung.com, vishak.g@...sung.com, neeraj.kernel@...il.com,
	cpgs@...sung.com
Subject: Re: [PATCH V4 16/17] cxl/pmem_region: Create pmem region using
 information parsed from LSA

On 19/11/25 04:37PM, Dave Jiang wrote:
>
>
>On 11/19/25 12:52 AM, Neeraj Kumar wrote:
>> create_pmem_region() creates cxl region based on region information
>> parsed from LSA. This routine required cxl root decoder and endpoint
>> decoder. Therefore added cxl_find_root_decoder_by_port() and
>> cxl_find_free_ep_decoder(). These routines find cxl root decoder and
>> free endpoint decoder on cxl bus using cxl port
>
>Please consider:
>create_pmem_region() creates CXL region based on region information
>parsed from the Label Storage Area (LSA). This routine requires cxl root
>decoder and endpoint decoder. Add cxl_find_root_decoder_by_port()
>and cxl_find_free_ep_decoder() to find the root decoder and a free
>endpoint decoder respectively.
>

Fixed it accordingly in V5

>>
>> Signed-off-by: Neeraj Kumar <s.neeraj@...sung.com>
>> ---
>>  drivers/cxl/core/core.h        |  4 ++
>>  drivers/cxl/core/pmem_region.c | 97 ++++++++++++++++++++++++++++++++++
>>  drivers/cxl/core/region.c      | 13 +++--
>>  drivers/cxl/cxl.h              |  5 ++
>>  4 files changed, 115 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/cxl/core/core.h b/drivers/cxl/core/core.h
>> index beeb9b7527b8..dd2efd3deb5e 100644
>> --- a/drivers/cxl/core/core.h
>> +++ b/drivers/cxl/core/core.h
>> @@ -35,6 +35,7 @@ int cxl_decoder_detach(struct cxl_region *cxlr,
>>  #define CXL_REGION_TYPE(x) (&cxl_region_type)
>>  #define SET_CXL_REGION_ATTR(x) (&dev_attr_##x.attr),
>>  #define CXL_DAX_REGION_TYPE(x) (&cxl_dax_region_type)
>> +int verify_free_decoder(struct device *dev);
>>  int cxl_region_init(void);
>>  void cxl_region_exit(void);
>>  int cxl_get_poison_by_endpoint(struct cxl_port *port);
>> @@ -88,6 +89,9 @@ static inline struct cxl_region *to_cxl_region(struct device *dev)
>>  {
>>  	return NULL;
>>  }
>> +static inline int verify_free_decoder(struct device *dev)
>> +{
>
>this function needs to return something

Thanks for catching this. Fixed it in V5

>
>> +}
>>  #define CXL_REGION_ATTR(x) NULL
>>  #define CXL_REGION_TYPE(x) NULL
>>  #define SET_CXL_REGION_ATTR(x)
>> diff --git a/drivers/cxl/core/pmem_region.c b/drivers/cxl/core/pmem_region.c
>> index be4feb73aafc..06665937c180 100644
>> --- a/drivers/cxl/core/pmem_region.c
>> +++ b/drivers/cxl/core/pmem_region.c
>> @@ -291,3 +291,100 @@ int devm_cxl_add_pmem_region(struct cxl_region *cxlr)
>>  	cxlr->cxl_nvb = NULL;
>>  	return rc;
>>  }
>> +
>> +static int match_root_decoder(struct device *dev, const void *data)
>> +{
>> +	return is_root_decoder(dev);
>
>Is it suppose to just grab the first root decoder? If so the function should be match_first_root_decoder(). However, should the root decoder cover the region it's trying to match to? Should there be some checks to see if the region fits under the root decoder range? Also, should it not check the root decoder flags to see if it has CXL_DECODER_F_PMEM set so the CFMWS can cover PMEM?
>

Yes Dave, Here we should check as you suggested. Also currently its only
considering only first decoder but it should return the root decoder
associated with particular endpoint decoder. I have fixed this in V5.



Regards,
Neeraj