linux-kernel - Re: [PATCH v6 05/14] remoteproc: pas: Replace metadata context with PAS context structure

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <17c9c47f-22f5-42c4-81ce-a5fae9bfff37@oss.qualcomm.com>
Date: Fri, 7 Nov 2025 11:04:56 +0100
From: Konrad Dybcio <konrad.dybcio@....qualcomm.com>
To: Mukesh Ojha <mukesh.ojha@....qualcomm.com>
Cc: Bjorn Andersson <andersson@...nel.org>,
        Mathieu Poirier <mathieu.poirier@...aro.org>,
        Rob Herring <robh@...nel.org>,
        Krzysztof Kozlowski <krzk+dt@...nel.org>,
        Conor Dooley
 <conor+dt@...nel.org>,
        Manivannan Sadhasivam <mani@...nel.org>,
        Konrad Dybcio <konradybcio@...nel.org>, linux-arm-msm@...r.kernel.org,
        linux-remoteproc@...r.kernel.org, devicetree@...r.kernel.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v6 05/14] remoteproc: pas: Replace metadata context with
 PAS context structure

On 11/5/25 7:42 AM, Mukesh Ojha wrote:
> On Tue, Nov 04, 2025 at 06:33:49PM +0100, Konrad Dybcio wrote:
>> On 11/4/25 8:35 AM, Mukesh Ojha wrote:
>>> As a superset of the existing metadata context, the PAS context
>>> structure enables both remoteproc and non-remoteproc subsystems to
>>> better support scenarios where the SoC runs with or without the Gunyah
>>> hypervisor. To reflect this, relevant SCM and metadata functions are
>>> updated to incorporate PAS context awareness.
>>>
>>> Signed-off-by: Mukesh Ojha <mukesh.ojha@....qualcomm.com>
>>> ---
>>>  drivers/firmware/qcom/qcom_scm.c       | 25 +++++++++++++---------
>>>  drivers/remoteproc/qcom_q6v5_pas.c     | 38 ++++++++++++++++++++++++----------
>>>  drivers/soc/qcom/mdt_loader.c          |  4 ++--
>>>  include/linux/firmware/qcom/qcom_scm.h |  4 ++--
>>>  include/linux/soc/qcom/mdt_loader.h    |  6 +++---
>>>  5 files changed, 49 insertions(+), 28 deletions(-)
>>>
>>> diff --git a/drivers/firmware/qcom/qcom_scm.c b/drivers/firmware/qcom/qcom_scm.c
>>> index 5a525dbd0a2e..9cdd152da592 100644
>>> --- a/drivers/firmware/qcom/qcom_scm.c
>>> +++ b/drivers/firmware/qcom/qcom_scm.c
>>> @@ -603,7 +603,7 @@ EXPORT_SYMBOL_GPL(devm_qcom_scm_pas_context_init);
>>>   *		and optional blob of data used for authenticating the metadata
>>>   *		and the rest of the firmware
>>>   * @size:	size of the metadata
>>> - * @ctx:	optional metadata context
>>> + * @ctx:	optional pas context
>>>   *
>>>   * Return: 0 on success.
>>>   *
>>> @@ -612,8 +612,9 @@ EXPORT_SYMBOL_GPL(devm_qcom_scm_pas_context_init);
>>>   * qcom_scm_pas_metadata_release() by the caller.
>>>   */
>>>  int qcom_scm_pas_init_image(u32 pas_id, const void *metadata, size_t size,
>>> -			    struct qcom_scm_pas_metadata *ctx)
>>> +			    struct qcom_scm_pas_context *ctx)
>>>  {
>>> +	struct qcom_scm_pas_metadata *mdt_ctx;
>>
>> This is never initialized
>>
>>>  	dma_addr_t mdata_phys;
>>>  	void *mdata_buf;
>>>  	int ret;
>>> @@ -665,9 +666,10 @@ int qcom_scm_pas_init_image(u32 pas_id, const void *metadata, size_t size,
>>>  	if (ret < 0 || !ctx) {
>>>  		dma_free_coherent(__scm->dev, size, mdata_buf, mdata_phys);
>>>  	} else if (ctx) {
>>> -		ctx->ptr = mdata_buf;
>>> -		ctx->phys = mdata_phys;
>>> -		ctx->size = size;
>>> +		mdt_ctx = ctx->metadata;
>>> +		mdt_ctx->ptr = mdata_buf;
>>> +		mdt_ctx->phys = mdata_phys;
>>> +		mdt_ctx->size = size;
>>
>> So this will always cause stack corruption
>>
>>>  	}
>>>  
>>>  	return ret ? : res.result[0];
>>> @@ -676,16 +678,19 @@ EXPORT_SYMBOL_GPL(qcom_scm_pas_init_image);
>>>  
>>>  /**
>>>   * qcom_scm_pas_metadata_release() - release metadata context
>>> - * @ctx:	metadata context
>>> + * @ctx:	pas context
>>>   */
>>> -void qcom_scm_pas_metadata_release(struct qcom_scm_pas_metadata *ctx)
>>> +void qcom_scm_pas_metadata_release(struct qcom_scm_pas_context *ctx)
>>>  {
>>> -	if (!ctx->ptr)
>>> +	struct qcom_scm_pas_metadata *mdt_ctx;
>>
>> Is the existence of this struct any useful after you introduced
>> pas_context?
> 
> Yes, it is still useful, mdt_ctx is only relevant for remoteproc based
> subsystem like adsp, cdsp, modem while they are not required for video,
> ipa, gpu etc. but the superset which is pas_context is needed by
> whosoever need to support secure PAS method Linux at EL2.

$ b4 shazam 20251104-kvm_rproc_v6-v6-0-7017b0adc24e@....qualcomm.com
$ rg 'struct qcom_scm_pas_metadata'

include/linux/firmware/qcom/qcom_scm.h
69:struct qcom_scm_pas_metadata {
80:     struct qcom_scm_pas_metadata *metadata;

drivers/firmware/qcom/qcom_scm.c
636:    struct qcom_scm_pas_metadata *mdt_ctx;
680:    struct qcom_scm_pas_metadata *mdt_ctx;
728:    struct qcom_scm_pas_metadata *mdt_ctx;

So really it seems like it always exists as part of the pas_context..
should we just make the larger struct integrate the smaller one and drop
the unnecessary layer?

TBF I don't really insist on this, but it surely looks a little odd

Konrad