Re: [PATCH v3 11/11] cxl: Enable AMD Zen5 address translation using ACPI PRMT

[Jonathan Cameron <jonathan.cameron@huawei.com>]

On Wed, 17 Sep 2025 11:43:39 +0200
Robert Richter <rrichter@amd.com> wrote:

> On 15.09.25 11:59:48, Jonathan Cameron wrote:
> > On Fri, 12 Sep 2025 16:45:13 +0200
> > Robert Richter <rrichter@amd.com> wrote:
> >   
> > > Add AMD Zen5 support for address translation.
> > > 
> > > Zen5 systems may be configured to use 'Normalized addresses'. Then,
> > > CXL endpoints use their own physical address space and are programmed
> > > passthrough (DPA == HPA), the number of interleaving ways for the
> > > endpoint is set to one. The Host Physical Addresses (HPAs) need to be
> > > translated from the endpoint to its CXL host bridge. The HPA of a CXL
> > > host bridge is equivalent to the System Physical Address (SPA).
> > > 
> > > ACPI Platform Runtime Mechanism (PRM) is used to translate the CXL
> > > Device Physical Address (DPA) to its System Physical Address. This is
> > > documented in:
> > > 
> > >  AMD Family 1Ah Models 00h–0Fh and Models 10h–1Fh
> > >  ACPI v6.5 Porting Guide, Publication # 58088
> > >  https://www.amd.com/en/search/documentation/hub.html
> > > 
> > > To implement AMD Zen5 address translation the following steps are
> > > needed:
> > > 
> > > AMD Zen5 systems support the ACPI PRM CXL Address Translation firmware
> > > call (Address Translation - CXL DPA to System Physical Address, see
> > > ACPI v6.5 Porting Guide above) when address translation is enabled.
> > > The existence of the callback can be identified using a specific GUID
> > > as documented. The initialization code checks firmware and kernel
> > > support of ACPI PRM.
> > > 
> > > Introduce a new file core/atl.c to handle ACPI PRM specific address
> > > translation code. Naming is loosely related to the kernel's AMD
> > > Address Translation Library (CONFIG_AMD_ATL) but implementation does
> > > not dependent on it, nor it is vendor specific. Use Kbuild and Kconfig
> > > options respectively to enable the code depending on architecture and
> > > platform options.
> > > 
> > > Implement an ACPI PRM firmware call for CXL address translation in the
> > > new function cxl_prm_to_hpa(). This includes sanity checks. Enable the
> > > callback for applicable CXL host bridges using the new cxl_atl_init()
> > > function.
> > > 
> > > Signed-off-by: Robert Richter <rrichter@amd.com>  
> > A few minor additions inline.  
> > 
> > J  
> > > ---
> > >  drivers/cxl/Kconfig       |   4 ++
> > >  drivers/cxl/core/Makefile |   1 +
> > >  drivers/cxl/core/atl.c    | 138 ++++++++++++++++++++++++++++++++++++++
> > >  drivers/cxl/core/core.h   |   1 +
> > >  drivers/cxl/core/port.c   |   8 +++
> > >  5 files changed, 152 insertions(+)
> > >  create mode 100644 drivers/cxl/core/atl.c
> > > 
> > > diff --git a/drivers/cxl/Kconfig b/drivers/cxl/Kconfig
> > > index 48b7314afdb8..31f9c96ef908 100644
> > > --- a/drivers/cxl/Kconfig
> > > +++ b/drivers/cxl/Kconfig
> > > @@ -233,4 +233,8 @@ config CXL_MCE
> > >  	def_bool y
> > >  	depends on X86_MCE && MEMORY_FAILURE
> > >  
> > > +config CXL_ATL
> > > +       def_bool y  
> > 
> > Given no help we can't turn this off manually and it's down to
> > whether ACPI_PRMT is configured or not.
> > 
> > To me this feels like something we should be able to control.
> > Not a huge amount of code, but none the less 'so far' it only
> > applies to particular AMD platforms yet ACPI_PRMT gets built
> > on ARM platforms and other stuff even on AMD (CONFIG_AMD_ATL_PRM)  
> 
> How about default y where possible but have a menu entry to disable
> address translation?
> 
> config CXL_ATL
> 	bool "CXL Address Translation support"
> 	default y
> 	depends on ACPI_PRMT
> 
> I don't want to make it specific to AMD.
Can we drop the default y?

That feels to me like a defconfig thing rather than a thing we should
apply generally.

Also remember to add some detailed help text given it's now visible in
menuconfig etc.

> 
> > 
> > 
> >   
> > > +       depends on ACPI_PRMT
> > > +
> > >  endif  
> >   

> > > +static void cxl_prm_init(struct cxl_port *port)
> > > +{
> > > +	u64 spa;
> > > +	struct prm_cxl_dpa_spa_data data = { .out = &spa, };
> > > +	int rc;
> > > +
> > > +	if (!check_prm_address_translation(port))
> > > +		return;
> > > +
> > > +	/* Check kernel (-EOPNOTSUPP) and firmware support (-ENODEV) */
> > > +	rc = acpi_call_prm_handler(prm_cxl_dpa_spa_guid, &data);
> > > +	if (rc == -EOPNOTSUPP || rc == -ENODEV)
> > > +		return;  
> > 
> > So other error values are fine?  IF they don't occur no need to be explicit
> > just check rc < 0 and return.  
> 
> This is just to check the existence of the PRM, but it will fail (if
> exists) here as parameters are a stub only. Both error codes are
> reserved for firmware or kernel support respectively. Else, it returns
> the PRM's error code, which is ignored here.

Why is the PRM error code not something we want to pay attention to?
Perhaps that's the comment that is missing.  I guess because we will
see any consistent failure later?


> 
> >   
> > > +
> > > +	port->to_hpa = cxl_prm_to_hpa;
> > > +
> > > +	dev_dbg(port->host_bridge, "PRM address translation enabled for %s.\n",
> > > +		dev_name(&port->dev));
> > > +}
> > > +
> > > +void cxl_atl_init(struct cxl_port *port)
> > > +{
> > > +	cxl_prm_init(port);  
> > Why not just rename cxl_prm_init() to cxl_atl_init() and get rid of this wrapper?  
> 
> cxl_prm_init() handles the PRM specifics, while cxl_atl_init() is used
> as an entry for the core module to enable address translation. I
> thought it would be misleading to name cxl_prm_init() different. The
> compiler result should be the same for both.
Ok.  I'm not sure it's worth the separation but as you say it'll almost
certainly get flattened anyway!

Jonathan

> 
> -Robert
> 
> >   
> > > +}  
> > 
> >