[Intel-xe] [PATCH 0/6] drm/xe: Meteorlake fixes.

[Intel-xe] [PATCH 0/6] drm/xe: Meteorlake fixes.

[Maarten Lankhorst]

I've been trying to make MTL-P work on Xe, and noticed a few things that needed fixing.
First of all, intel_dram needed updating for MTL. But instead of trying to keep track of
intel_pch.c and intel_dram.c, we compile the copy from i915. Ideally we would also use the
headers from i915, but I haven't found a clean way to make it work.

Maybe change the compat headers to #include "../../i915/soc/FILENAME?", or allow a symlink?

Maarten Lankhorst (6):
  drm/xe: Fix meteorlake stolen memory
  drm/xe: Use full_gt batchbuffer allocation for media tiles.
  drm/i915: Fix comparison in intel_dram.
  drm/xe: Fix XE_LPDP and meteorlake display info.
  drm/xe: Build soc files directly
  drm/xe: Fixup __intel_de_wait_for_register macro.

 drivers/gpu/drm/i915/soc/intel_dram.c         |  32 +-
 drivers/gpu/drm/xe/Makefile                   |   8 +-
 .../gpu/drm/xe/compat-i915-headers/i915_drv.h |   2 +-
 .../soc}/intel_dram.h                         |   1 -
 .../soc}/intel_pch.h                          |   0
 drivers/gpu/drm/xe/display/ext/intel_dram.c   | 495 ------------------
 drivers/gpu/drm/xe/display/ext/intel_pch.c    | 157 ------
 drivers/gpu/drm/xe/display/xe_de.h            |   4 +
 drivers/gpu/drm/xe/xe_bo.c                    |   3 +-
 drivers/gpu/drm/xe/xe_device_types.h          |   2 +-
 drivers/gpu/drm/xe/xe_display.c               |  11 +-
 drivers/gpu/drm/xe/xe_gt.c                    |  10 +-
 drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c        |   3 +-
 13 files changed, 50 insertions(+), 678 deletions(-)
 rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_dram.h (80%)
 rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_pch.h (100%)
 delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_dram.c
 delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_pch.c

-- 
2.34.1

[Intel-xe] [PATCH 1/6] drm/xe: Fix meteorlake stolen memory

[Maarten Lankhorst]

MTL should use bar2 to access stolen memory, not the physical addresses
directly. Correct the GPU address to point to stolen base, and set the
VRAM flag to indicate it's in device memory.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
---
 drivers/gpu/drm/xe/xe_bo.c             | 3 ++-
 drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c | 3 ++-
 2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
index e4d079b61d52..9b2f95f3ce49 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -47,7 +47,8 @@ bool mem_type_is_vram(u32 mem_type)
 
 static bool resource_is_stolen_vram(struct xe_device *xe, struct ttm_resource *res)
 {
-	return res->mem_type == XE_PL_STOLEN && IS_DGFX(xe);
+	return res->mem_type == XE_PL_STOLEN &&
+	       (IS_DGFX(xe) || !xe_ttm_stolen_cpu_access_needs_ggtt(xe));
 }
 
 static bool resource_is_vram(struct ttm_resource *res)
diff --git a/drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c b/drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c
index 31887fec1073..bd1b73fa84a5 100644
--- a/drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c
+++ b/drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c
@@ -94,7 +94,8 @@ static u32 detect_bar2_integrated(struct xe_device *xe, struct xe_ttm_stolen_mgr
 	if (drm_WARN_ON(&xe->drm, (ggc & GGMS_MASK) != GGMS_MASK))
 		return 0;
 
-	mgr->stolen_base = mgr->io_base = pci_resource_start(pdev, 2) + SZ_8M;
+	mgr->io_base = pci_resource_start(pdev, 2) + SZ_8M;
+	mgr->stolen_base = SZ_8M;
 
 	/* return valid GMS value, -EIO if invalid */
 	gms = REG_FIELD_GET(GMS_MASK, ggc);
-- 
2.34.1

Re: [Intel-xe] [PATCH 1/6] drm/xe: Fix meteorlake stolen memory

[Lucas De Marchi]

On Fri, Mar 31, 2023 at 12:24:14PM +0200, Maarten Lankhorst wrote:
>MTL should use bar2 to access stolen memory, not the physical addresses
>directly. Correct the GPU address to point to stolen base, and set the
>VRAM flag to indicate it's in device memory.

humn... shouldn't the commit message point that for integrated the
address for stolen is an offset from the base rather than the physical
address?  It was already using bar2, but with the wrong address.

>
>Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>


Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>

thanks
Lucas De Marchi

>---
> drivers/gpu/drm/xe/xe_bo.c             | 3 ++-
> drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c | 3 ++-
> 2 files changed, 4 insertions(+), 2 deletions(-)
>
>diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
>index e4d079b61d52..9b2f95f3ce49 100644
>--- a/drivers/gpu/drm/xe/xe_bo.c
>+++ b/drivers/gpu/drm/xe/xe_bo.c
>@@ -47,7 +47,8 @@ bool mem_type_is_vram(u32 mem_type)
>
> static bool resource_is_stolen_vram(struct xe_device *xe, struct ttm_resource *res)
> {
>-	return res->mem_type == XE_PL_STOLEN && IS_DGFX(xe);
>+	return res->mem_type == XE_PL_STOLEN &&
>+	       (IS_DGFX(xe) || !xe_ttm_stolen_cpu_access_needs_ggtt(xe));
> }
>
> static bool resource_is_vram(struct ttm_resource *res)
>diff --git a/drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c b/drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c
>index 31887fec1073..bd1b73fa84a5 100644
>--- a/drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c
>+++ b/drivers/gpu/drm/xe/xe_ttm_stolen_mgr.c
>@@ -94,7 +94,8 @@ static u32 detect_bar2_integrated(struct xe_device *xe, struct xe_ttm_stolen_mgr
> 	if (drm_WARN_ON(&xe->drm, (ggc & GGMS_MASK) != GGMS_MASK))
> 		return 0;
>
>-	mgr->stolen_base = mgr->io_base = pci_resource_start(pdev, 2) + SZ_8M;
>+	mgr->io_base = pci_resource_start(pdev, 2) + SZ_8M;
>+	mgr->stolen_base = SZ_8M;
>
> 	/* return valid GMS value, -EIO if invalid */
> 	gms = REG_FIELD_GET(GMS_MASK, ggc);
>-- 
>2.34.1
>

[Intel-xe] [PATCH 2/6] drm/xe: Use full_gt batchbuffer allocation for media tiles.

[Maarten Lankhorst]

This fixes an oops on media tiles, because we don't initialise the
bb pool.
---
 drivers/gpu/drm/xe/xe_gt.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c
index bc821f431c45..c7a2e9baabfb 100644
--- a/drivers/gpu/drm/xe/xe_gt.c
+++ b/drivers/gpu/drm/xe/xe_gt.c
@@ -237,6 +237,10 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
 		struct xe_engine *e, *nop_e;
 		struct xe_vm *vm;
 		void *default_lrc;
+		struct xe_gt *full_gt = gt;
+
+		if (xe_gt_is_media_type(gt))
+			full_gt = xe_find_full_gt(gt);
 
 		if (gt->default_lrc[hwe->class])
 			continue;
@@ -262,7 +266,7 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
 		}
 
 		/* Prime golden LRC with known good state */
-		err = emit_wa_job(gt, e);
+		err = emit_wa_job(full_gt, e);
 		if (err) {
 			drm_err(&xe->drm, "gt%d, hwe %s, guc_id=%d, emit_wa_job,e failed=%d",
 				gt->info.id, hwe->name, e->guc->id, err);
@@ -279,7 +283,7 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
 		}
 
 		/* Switch to different LRC */
-		err = emit_nop_job(gt, nop_e);
+		err = emit_nop_job(full_gt, nop_e);
 		if (err) {
 			drm_err(&xe->drm, "gt%d, hwe %s, guc_id=%d, emit_nop_job,nop_e failed=%d",
 				gt->info.id, hwe->name, nop_e->guc->id, err);
@@ -287,7 +291,7 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
 		}
 
 		/* Reload golden LRC to record the effect of any indirect W/A */
-		err = emit_nop_job(gt, e);
+		err = emit_nop_job(full_gt, e);
 		if (err) {
 			drm_err(&xe->drm, "gt%d, hwe %s, guc_id=%d, emit_nop_job,e failed=%d",
 				gt->info.id, hwe->name, e->guc->id, err);
-- 
2.34.1

Re: [Intel-xe] [PATCH 2/6] drm/xe: Use full_gt batchbuffer allocation for media tiles.

[Matt Roper]

On Fri, Mar 31, 2023 at 12:24:15PM +0200, Maarten Lankhorst wrote:
> This fixes an oops on media tiles, because we don't initialise the
> bb pool.
> ---
>  drivers/gpu/drm/xe/xe_gt.c | 10 +++++++---
>  1 file changed, 7 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c
> index bc821f431c45..c7a2e9baabfb 100644
> --- a/drivers/gpu/drm/xe/xe_gt.c
> +++ b/drivers/gpu/drm/xe/xe_gt.c
> @@ -237,6 +237,10 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
>  		struct xe_engine *e, *nop_e;
>  		struct xe_vm *vm;
>  		void *default_lrc;
> +		struct xe_gt *full_gt = gt;
> +
> +		if (xe_gt_is_media_type(gt))
> +			full_gt = xe_find_full_gt(gt);
>  
>  		if (gt->default_lrc[hwe->class])
>  			continue;
> @@ -262,7 +266,7 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
>  		}
>  
>  		/* Prime golden LRC with known good state */
> -		err = emit_wa_job(gt, e);
> +		err = emit_wa_job(full_gt, e);

If we pass the primary GT here, then we'll take the wrong path in
bb_prefetch().  That doesn't actually matter very much (we'll just
overestimate the padding needed for media engines, which is safe and
something we already do for BCS engines), but we should probably just
drop !xe_gt_is_media_type(gt) condition there as part of this patch
since it will always be true now.


Matt

>  		if (err) {
>  			drm_err(&xe->drm, "gt%d, hwe %s, guc_id=%d, emit_wa_job,e failed=%d",
>  				gt->info.id, hwe->name, e->guc->id, err);
> @@ -279,7 +283,7 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
>  		}
>  
>  		/* Switch to different LRC */
> -		err = emit_nop_job(gt, nop_e);
> +		err = emit_nop_job(full_gt, nop_e);
>  		if (err) {
>  			drm_err(&xe->drm, "gt%d, hwe %s, guc_id=%d, emit_nop_job,nop_e failed=%d",
>  				gt->info.id, hwe->name, nop_e->guc->id, err);
> @@ -287,7 +291,7 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
>  		}
>  
>  		/* Reload golden LRC to record the effect of any indirect W/A */
> -		err = emit_nop_job(gt, e);
> +		err = emit_nop_job(full_gt, e);
>  		if (err) {
>  			drm_err(&xe->drm, "gt%d, hwe %s, guc_id=%d, emit_nop_job,e failed=%d",
>  				gt->info.id, hwe->name, e->guc->id, err);
> -- 
> 2.34.1
> 

-- 
Matt Roper
Graphics Software Engineer
Linux GPU Platform Enablement
Intel Corporation

Re: [Intel-xe] [PATCH 2/6] drm/xe: Use full_gt batchbuffer allocation for media tiles.

[Maarten Lankhorst]

On 2023-04-03 22:27, Matt Roper wrote:
> On Fri, Mar 31, 2023 at 12:24:15PM +0200, Maarten Lankhorst wrote:
>> This fixes an oops on media tiles, because we don't initialise the
>> bb pool.
>> ---
>>   drivers/gpu/drm/xe/xe_gt.c | 10 +++++++---
>>   1 file changed, 7 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c
>> index bc821f431c45..c7a2e9baabfb 100644
>> --- a/drivers/gpu/drm/xe/xe_gt.c
>> +++ b/drivers/gpu/drm/xe/xe_gt.c
>> @@ -237,6 +237,10 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
>>   		struct xe_engine *e, *nop_e;
>>   		struct xe_vm *vm;
>>   		void *default_lrc;
>> +		struct xe_gt *full_gt = gt;
>> +
>> +		if (xe_gt_is_media_type(gt))
>> +			full_gt = xe_find_full_gt(gt);
>>   
>>   		if (gt->default_lrc[hwe->class])
>>   			continue;
>> @@ -262,7 +266,7 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
>>   		}
>>   
>>   		/* Prime golden LRC with known good state */
>> -		err = emit_wa_job(gt, e);
>> +		err = emit_wa_job(full_gt, e);
> If we pass the primary GT here, then we'll take the wrong path in
> bb_prefetch().  That doesn't actually matter very much (we'll just
> overestimate the padding needed for media engines, which is safe and
> something we already do for BCS engines), but we should probably just
> drop !xe_gt_is_media_type(gt) condition there as part of this patch
> since it will always be true now.

Yeah makes sense. That check doesn't strictly speaking seem to be 
needed, so I'm fine with dropping it.

We could always request the engine to be specified, but I don't think we 
really use those batchbuffers beyond initialisation outside of 
migration/VM_BIND,

which is done on a BCS engine anyway.

Can I have your r-b with that check removed?

~Maarten

Re: [Intel-xe] [PATCH 2/6] drm/xe: Use full_gt batchbuffer allocation for media tiles.

[Matt Roper]

On Tue, Apr 04, 2023 at 08:59:27AM +0200, Maarten Lankhorst wrote:
> 
> On 2023-04-03 22:27, Matt Roper wrote:
> > On Fri, Mar 31, 2023 at 12:24:15PM +0200, Maarten Lankhorst wrote:
> > > This fixes an oops on media tiles, because we don't initialise the
> > > bb pool.
> > > ---
> > >   drivers/gpu/drm/xe/xe_gt.c | 10 +++++++---
> > >   1 file changed, 7 insertions(+), 3 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/xe/xe_gt.c b/drivers/gpu/drm/xe/xe_gt.c
> > > index bc821f431c45..c7a2e9baabfb 100644
> > > --- a/drivers/gpu/drm/xe/xe_gt.c
> > > +++ b/drivers/gpu/drm/xe/xe_gt.c
> > > @@ -237,6 +237,10 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
> > >   		struct xe_engine *e, *nop_e;
> > >   		struct xe_vm *vm;
> > >   		void *default_lrc;
> > > +		struct xe_gt *full_gt = gt;
> > > +
> > > +		if (xe_gt_is_media_type(gt))
> > > +			full_gt = xe_find_full_gt(gt);
> > >   		if (gt->default_lrc[hwe->class])
> > >   			continue;
> > > @@ -262,7 +266,7 @@ int xe_gt_record_default_lrcs(struct xe_gt *gt)
> > >   		}
> > >   		/* Prime golden LRC with known good state */
> > > -		err = emit_wa_job(gt, e);
> > > +		err = emit_wa_job(full_gt, e);
> > If we pass the primary GT here, then we'll take the wrong path in
> > bb_prefetch().  That doesn't actually matter very much (we'll just
> > overestimate the padding needed for media engines, which is safe and
> > something we already do for BCS engines), but we should probably just
> > drop !xe_gt_is_media_type(gt) condition there as part of this patch
> > since it will always be true now.
> 
> Yeah makes sense. That check doesn't strictly speaking seem to be needed, so
> I'm fine with dropping it.
> 
> We could always request the engine to be specified, but I don't think we
> really use those batchbuffers beyond initialisation outside of
> migration/VM_BIND,
> 
> which is done on a BCS engine anyway.
> 
> Can I have your r-b with that check removed?

Yeah,

Reviewed-by: Matt Roper <matthew.d.roper@intel.com>

Thanks.


Matt

> 
> ~Maarten
> 

-- 
Matt Roper
Graphics Software Engineer
Linux GPU Platform Enablement
Intel Corporation

[Intel-xe] [PATCH 3/6] drm/i915: Fix comparison in intel_dram.

[Maarten Lankhorst]

Gen13 should probably be the same as gen12, not gen11.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: intel-gfx@lists.freedesktop.org
---
 drivers/gpu/drm/i915/soc/intel_dram.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
index 9f0651d48d41..9649051ed8ed 100644
--- a/drivers/gpu/drm/i915/soc/intel_dram.c
+++ b/drivers/gpu/drm/i915/soc/intel_dram.c
@@ -548,7 +548,7 @@ static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
 	if (ret)
 		return ret;
 
-	if (GRAPHICS_VER(dev_priv) == 12) {
+	if (GRAPHICS_VER(dev_priv) >= 12) {
 		switch (val & 0xf) {
 		case 0:
 			dram_info->type = INTEL_DRAM_DDR4;
-- 
2.34.1

Re: [Intel-xe] [PATCH 3/6] drm/i915: Fix comparison in intel_dram.

[Matt Roper]

On Fri, Mar 31, 2023 at 12:24:16PM +0200, Maarten Lankhorst wrote:
> Gen13 should probably be the same as gen12, not gen11.
> 
> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> Cc: intel-gfx@lists.freedesktop.org
> ---
>  drivers/gpu/drm/i915/soc/intel_dram.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
> index 9f0651d48d41..9649051ed8ed 100644
> --- a/drivers/gpu/drm/i915/soc/intel_dram.c
> +++ b/drivers/gpu/drm/i915/soc/intel_dram.c
> @@ -548,7 +548,7 @@ static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)

I don't think we need this change.  We were only reading the this pcode
mailbox for display purposes, and even on the latest graphics version 12
platforms we shouldn't make it into this function anymore.  Instead the
display IP now provides this information itself so there's no need to go
through the GT's pcode mailbox; we read it directly from display
registers in xelpdp_get_dram_info().  It looks like this condition here
would only matter if we had a future platform with graphics version
higher than 12, but display version lower than 14, which seems very
unlikely.


Matt

>  	if (ret)
>  		return ret;
>  
> -	if (GRAPHICS_VER(dev_priv) == 12) {
> +	if (GRAPHICS_VER(dev_priv) >= 12) {
>  		switch (val & 0xf) {
>  		case 0:
>  			dram_info->type = INTEL_DRAM_DDR4;
> -- 
> 2.34.1
> 

-- 
Matt Roper
Graphics Software Engineer
Linux GPU Platform Enablement
Intel Corporation

Re: [Intel-xe] [Intel-gfx] [PATCH 3/6] drm/i915: Fix comparison in intel_dram.

[Ville Syrjälä]

On Mon, Apr 03, 2023 at 01:35:40PM -0700, Matt Roper wrote:
> On Fri, Mar 31, 2023 at 12:24:16PM +0200, Maarten Lankhorst wrote:
> > Gen13 should probably be the same as gen12, not gen11.
> > 
> > Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> > Cc: intel-gfx@lists.freedesktop.org
> > ---
> >  drivers/gpu/drm/i915/soc/intel_dram.c | 2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
> > index 9f0651d48d41..9649051ed8ed 100644
> > --- a/drivers/gpu/drm/i915/soc/intel_dram.c
> > +++ b/drivers/gpu/drm/i915/soc/intel_dram.c
> > @@ -548,7 +548,7 @@ static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
> 
> I don't think we need this change.  We were only reading the this pcode
> mailbox for display purposes, and even on the latest graphics version 12
> platforms we shouldn't make it into this function anymore.  Instead the
> display IP now provides this information itself so there's no need to go
> through the GT's pcode mailbox; we read it directly from display
> registers in xelpdp_get_dram_info().  It looks like this condition here
> would only matter if we had a future platform with graphics version
> higher than 12, but display version lower than 14, which seems very
> unlikely.
> 
> 
> Matt
> 
> >  	if (ret)
> >  		return ret;
> >  
> > -	if (GRAPHICS_VER(dev_priv) == 12) {
> > +	if (GRAPHICS_VER(dev_priv) >= 12) {

We should perhaps change all the checks here to DISPLAY_VER
since it's the display code that needs this information.

Though I suppose technically neither DISPLAY_VER nor
GRAPHICS_VER is entirely correct since it's really SoC
level stuff that we're looking at here. But the current 
mismash of GRAPHICS_VER and DISPLAY_VER is certainly not
helping reduce the confusion.

> >  		switch (val & 0xf) {
> >  		case 0:
> >  			dram_info->type = INTEL_DRAM_DDR4;
> > -- 
> > 2.34.1
> > 
> 
> -- 
> Matt Roper
> Graphics Software Engineer
> Linux GPU Platform Enablement
> Intel Corporation

-- 
Ville Syrjälä
Intel

Re: [Intel-xe] [PATCH 3/6] drm/i915: Fix comparison in intel_dram.

[Maarten Lankhorst]

Hey,

On 2023-04-03 22:35, Matt Roper wrote:
> On Fri, Mar 31, 2023 at 12:24:16PM +0200, Maarten Lankhorst wrote:
>> Gen13 should probably be the same as gen12, not gen11.
>>
>> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
>> Cc: intel-gfx@lists.freedesktop.org
>> ---
>>   drivers/gpu/drm/i915/soc/intel_dram.c | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
>> index 9f0651d48d41..9649051ed8ed 100644
>> --- a/drivers/gpu/drm/i915/soc/intel_dram.c
>> +++ b/drivers/gpu/drm/i915/soc/intel_dram.c
>> @@ -548,7 +548,7 @@ static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
> I don't think we need this change.  We were only reading the this pcode
> mailbox for display purposes, and even on the latest graphics version 12
> platforms we shouldn't make it into this function anymore.  Instead the
> display IP now provides this information itself so there's no need to go
> through the GT's pcode mailbox; we read it directly from display
> registers in xelpdp_get_dram_info().  It looks like this condition here
> would only matter if we had a future platform with graphics version
> higher than 12, but display version lower than 14, which seems very
> unlikely.

Probably not, but it's at least cosmetically more correct. We only need 
the dram code for display, so in theory we could check for display_ver 
 >= 12 instead and have it unified.

If something breaks the pattern, we can fix it then. :)

~Maarten

Re: [Intel-xe] [Intel-gfx] [PATCH 3/6] drm/i915: Fix comparison in intel_dram.

[Lucas De Marchi]

On Fri, Mar 31, 2023 at 12:24:16PM +0200, Maarten Lankhorst wrote:
>Gen13 should probably be the same as gen12, not gen11.

there is no such thing as gen13. Please reword commit message to use
graphics version, although here it's probably more related to the
display engine (and hence DISPLAY_VER()). But then the code you are
adding here is probably dead code since 825477e77912 ("drm/i915/mtl:
Obtain SAGV values from MMIO instead of GT pcode mailbox")

If this is fixing anything what exactly is it? Otherwise it'd be more
"extend to next platforms" rather than "Fix".

Lucas De Marchi

>
>Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
>Cc: intel-gfx@lists.freedesktop.org
>---
> drivers/gpu/drm/i915/soc/intel_dram.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
>diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
>index 9f0651d48d41..9649051ed8ed 100644
>--- a/drivers/gpu/drm/i915/soc/intel_dram.c
>+++ b/drivers/gpu/drm/i915/soc/intel_dram.c
>@@ -548,7 +548,7 @@ static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
> 	if (ret)
> 		return ret;
>
>-	if (GRAPHICS_VER(dev_priv) == 12) {
>+	if (GRAPHICS_VER(dev_priv) >= 12) {
> 		switch (val & 0xf) {
> 		case 0:
> 			dram_info->type = INTEL_DRAM_DDR4;
>-- 
>2.34.1
>

[Intel-xe] [PATCH 4/6] drm/xe: Fix XE_LPDP and meteorlake display info.

[Maarten Lankhorst]

MTL has modular FIA, and LPDP also has FBC on pipe B.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
---
 drivers/gpu/drm/xe/xe_display.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_display.c b/drivers/gpu/drm/xe/xe_display.c
index e9b3997c4028..64e65b29d8de 100644
--- a/drivers/gpu/drm/xe/xe_display.c
+++ b/drivers/gpu/drm/xe/xe_display.c
@@ -475,8 +475,8 @@ __diag_ignore_all("-Woverride-init", "Allow field overrides in table");
 	XE_LPD,								\
 	.ver = 14,							\
 	.has_cdclk_crawl = 1,						\
-	.has_cdclk_squash = 1
-
+	.has_cdclk_squash = 1,						\
+	.fbc_mask = BIT(INTEL_FBC_A) | BIT(INTEL_FBC_B)
 
 void xe_display_info_init(struct xe_device *xe)
 {
@@ -513,7 +513,10 @@ void xe_display_info_init(struct xe_device *xe)
 		};
 		break;
 	case XE_METEORLAKE:
-		xe->info.display = (struct xe_device_display_info) { XE_LPDP };
+		xe->info.display = (struct xe_device_display_info) {
+			XE_LPDP,
+			.has_modular_fia = 1,
+		};
 		break;
 	default:
 		drm_dbg(&xe->drm, "No display IP, skipping\n");
-- 
2.34.1

Re: [Intel-xe] [PATCH 4/6] drm/xe: Fix XE_LPDP and meteorlake display info.

[Lucas De Marchi]

On Fri, Mar 31, 2023 at 12:24:17PM +0200, Maarten Lankhorst wrote:
>MTL has modular FIA, and LPDP also has FBC on pipe B.

commit message needs to be drm/xe/display: Fix XE_LPDP and meteorlake

And here:

Bspec: 65380
>
>Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
>---
> drivers/gpu/drm/xe/xe_display.c | 9 ++++++---
> 1 file changed, 6 insertions(+), 3 deletions(-)
>
>diff --git a/drivers/gpu/drm/xe/xe_display.c b/drivers/gpu/drm/xe/xe_display.c
>index e9b3997c4028..64e65b29d8de 100644
>--- a/drivers/gpu/drm/xe/xe_display.c
>+++ b/drivers/gpu/drm/xe/xe_display.c
>@@ -475,8 +475,8 @@ __diag_ignore_all("-Woverride-init", "Allow field overrides in table");
> 	XE_LPD,								\
> 	.ver = 14,							\
> 	.has_cdclk_crawl = 1,						\
>-	.has_cdclk_squash = 1
>-
>+	.has_cdclk_squash = 1,						\
>+	.fbc_mask = BIT(INTEL_FBC_A) | BIT(INTEL_FBC_B)
>
> void xe_display_info_init(struct xe_device *xe)
> {
>@@ -513,7 +513,10 @@ void xe_display_info_init(struct xe_device *xe)
> 		};
> 		break;
> 	case XE_METEORLAKE:
>-		xe->info.display = (struct xe_device_display_info) { XE_LPDP };
>+		xe->info.display = (struct xe_device_display_info) {
>+			XE_LPDP,
>+			.has_modular_fia = 1,

I think to avoid the extra 3-level inheritance here and based on BSpec 65380,
it's safe to just add this to LPDP.

+Matt Roper on that.

With those changes, feel free to add:

	Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>

thanks
Lucas De Marchi

>+		};
> 		break;
> 	default:
> 		drm_dbg(&xe->drm, "No display IP, skipping\n");
>-- 
>2.34.1
>

[Intel-xe] [PATCH 5/6] drm/xe: Build soc files directly

[Maarten Lankhorst]

Instead of making a copy that is hard to keep up to date, build the
source directly.

Add a rule for soc_* files, and use those to build i915.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
---
 drivers/gpu/drm/i915/soc/intel_dram.c         |  30 +-
 drivers/gpu/drm/xe/Makefile                   |   8 +-
 .../gpu/drm/xe/compat-i915-headers/i915_drv.h |   2 +-
 .../soc}/intel_dram.h                         |   1 -
 .../soc}/intel_pch.h                          |   0
 drivers/gpu/drm/xe/display/ext/intel_dram.c   | 495 ------------------
 drivers/gpu/drm/xe/display/ext/intel_pch.c    | 157 ------
 drivers/gpu/drm/xe/xe_device_types.h          |   2 +-
 drivers/gpu/drm/xe/xe_display.c               |   2 +-
 9 files changed, 28 insertions(+), 669 deletions(-)
 rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_dram.h (80%)
 rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_pch.h (100%)
 delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_dram.c
 delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_pch.c

diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
index 9649051ed8ed..13fe18a9372f 100644
--- a/drivers/gpu/drm/i915/soc/intel_dram.c
+++ b/drivers/gpu/drm/i915/soc/intel_dram.c
@@ -9,8 +9,12 @@
 #include "i915_reg.h"
 #include "intel_dram.h"
 #include "intel_mchbar_regs.h"
-#include "intel_pcode.h"
+#ifdef I915
 #include "vlv_sideband.h"
+#include "display/intel_de.h"
+#else
+#include "xe_de.h"
+#endif
 
 struct dram_dimm_info {
 	u16 size;
@@ -43,6 +47,8 @@ static const char *intel_dram_type_str(enum intel_dram_type type)
 
 #undef DRAM_TYPE_STR
 
+#ifdef I915
+
 static void pnv_detect_mem_freq(struct drm_i915_private *dev_priv)
 {
 	u32 tmp;
@@ -191,6 +197,9 @@ static void detect_mem_freq(struct drm_i915_private *i915)
 	if (i915->mem_freq)
 		drm_dbg(&i915->drm, "DDR speed: %d MHz\n", i915->mem_freq);
 }
+#else
+#define detect_mem_freq(i915) do { } while (0)
+#endif
 
 static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
 {
@@ -339,14 +348,12 @@ skl_dram_get_channels_info(struct drm_i915_private *i915)
 	u32 val;
 	int ret;
 
-	val = intel_uncore_read(&i915->uncore,
-				SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
+	val = intel_de_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
 	ret = skl_dram_get_channel_info(i915, &ch0, 0, val);
 	if (ret == 0)
 		dram_info->num_channels++;
 
-	val = intel_uncore_read(&i915->uncore,
-				SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
+	val = intel_de_read(i915, SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
 	ret = skl_dram_get_channel_info(i915, &ch1, 1, val);
 	if (ret == 0)
 		dram_info->num_channels++;
@@ -376,8 +383,7 @@ skl_get_dram_type(struct drm_i915_private *i915)
 {
 	u32 val;
 
-	val = intel_uncore_read(&i915->uncore,
-				SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
+	val = intel_de_read(i915, SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
 
 	switch (val & SKL_DRAM_DDR_TYPE_MASK) {
 	case SKL_DRAM_DDR_TYPE_DDR3:
@@ -503,7 +509,7 @@ static int bxt_get_dram_info(struct drm_i915_private *i915)
 		struct dram_dimm_info dimm;
 		enum intel_dram_type type;
 
-		val = intel_uncore_read(&i915->uncore, BXT_D_CR_DRP0_DUNIT(i));
+		val = intel_de_read(i915, BXT_D_CR_DRP0_DUNIT(i));
 		if (val == 0xFFFFFFFF)
 			continue;
 
@@ -543,8 +549,8 @@ static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
 	u32 val = 0;
 	int ret;
 
-	ret = snb_pcode_read(&dev_priv->uncore, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
-			     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
+	ret = intel_de_pcode_read(dev_priv, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
+				  ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
 	if (ret)
 		return ret;
 
@@ -618,7 +624,7 @@ static int gen12_get_dram_info(struct drm_i915_private *i915)
 
 static int xelpdp_get_dram_info(struct drm_i915_private *i915)
 {
-	u32 val = intel_uncore_read(&i915->uncore, MTL_MEM_SS_INFO_GLOBAL);
+	u32 val = intel_de_read(i915, MTL_MEM_SS_INFO_GLOBAL);
 	struct dram_info *dram_info = &i915->dram_info;
 
 	switch (REG_FIELD_GET(MTL_DDR_TYPE_MASK, val)) {
@@ -687,6 +693,7 @@ void intel_dram_detect(struct drm_i915_private *i915)
 		    str_yes_no(dram_info->wm_lv_0_adjust_needed));
 }
 
+#ifdef I915
 static u32 gen9_edram_size_mb(struct drm_i915_private *i915, u32 cap)
 {
 	static const u8 ways[8] = { 4, 8, 12, 16, 16, 16, 16, 16 };
@@ -722,3 +729,4 @@ void intel_dram_edram_detect(struct drm_i915_private *i915)
 
 	drm_info(&i915->drm, "Found %uMB of eDRAM\n", i915->edram_size_mb);
 }
+#endif
diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
index 669c7b6f552c..3ebef784445e 100644
--- a/drivers/gpu/drm/xe/Makefile
+++ b/drivers/gpu/drm/xe/Makefile
@@ -124,6 +124,10 @@ $(obj)/display/intel_%.o: $(srctree)/drivers/gpu/drm/i915/display/intel_%.c FORC
 	$(call cmd,force_checksrc)
 	$(call if_changed_rule,cc_o_c)
 
+$(obj)/display/soc_%.o: $(srctree)/drivers/gpu/drm/i915/soc/%.c FORCE
+	$(call cmd,force_checksrc)
+	$(call if_changed_rule,cc_o_c)
+
 # Display..
 xe-$(CONFIG_DRM_XE_DISPLAY) += \
 	xe_display.o \
@@ -204,8 +208,8 @@ xe-$(CONFIG_DRM_XE_DISPLAY) += \
 	display/ext/i915_irq.o \
 	display/ext/i9xx_wm.o \
 	display/ext/intel_device_info.o \
-	display/ext/intel_dram.o \
-	display/ext/intel_pch.o \
+	display/soc_intel_dram.o \
+	display/soc_intel_pch.o \
 	display/ext/intel_pm.o
 
 ifeq ($(CONFIG_ACPI),y)
diff --git a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
index 92dfbb8b3ca4..fa8c53e91bce 100644
--- a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
+++ b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
@@ -15,7 +15,7 @@
 #include "xe_pm.h"
 #include "xe_step.h"
 #include "i915_reg_defs.h"
-#include "intel_pch.h"
+#include "soc/intel_pch.h"
 #include "i915_utils.h"
 #include <linux/pm_runtime.h>
 
diff --git a/drivers/gpu/drm/xe/display/ext/intel_dram.h b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
similarity index 80%
rename from drivers/gpu/drm/xe/display/ext/intel_dram.h
rename to drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
index 4ba13c13162c..7ed5b6e1f805 100644
--- a/drivers/gpu/drm/xe/display/ext/intel_dram.h
+++ b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
@@ -8,7 +8,6 @@
 
 struct drm_i915_private;
 
-void intel_dram_edram_detect(struct drm_i915_private *i915);
 void intel_dram_detect(struct drm_i915_private *i915);
 
 #endif /* __INTEL_DRAM_H__ */
diff --git a/drivers/gpu/drm/xe/display/ext/intel_pch.h b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h
similarity index 100%
rename from drivers/gpu/drm/xe/display/ext/intel_pch.h
rename to drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h
diff --git a/drivers/gpu/drm/xe/display/ext/intel_dram.c b/drivers/gpu/drm/xe/display/ext/intel_dram.c
deleted file mode 100644
index a2df0c502c34..000000000000
--- a/drivers/gpu/drm/xe/display/ext/intel_dram.c
+++ /dev/null
@@ -1,495 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright © 2020 Intel Corporation
- */
-
-#include <linux/string_helpers.h>
-
-#include "i915_drv.h"
-#include "i915_reg.h"
-#include "intel_de.h"
-#include "intel_dram.h"
-#include "intel_mchbar_regs.h"
-
-struct dram_dimm_info {
-	u16 size;
-	u8 width, ranks;
-};
-
-struct dram_channel_info {
-	struct dram_dimm_info dimm_l, dimm_s;
-	u8 ranks;
-	bool is_16gb_dimm;
-};
-
-#define DRAM_TYPE_STR(type) [INTEL_DRAM_ ## type] = #type
-
-static const char *intel_dram_type_str(enum intel_dram_type type)
-{
-	static const char * const str[] = {
-		DRAM_TYPE_STR(UNKNOWN),
-		DRAM_TYPE_STR(DDR3),
-		DRAM_TYPE_STR(DDR4),
-		DRAM_TYPE_STR(LPDDR3),
-		DRAM_TYPE_STR(LPDDR4),
-	};
-
-	if (type >= ARRAY_SIZE(str))
-		type = INTEL_DRAM_UNKNOWN;
-
-	return str[type];
-}
-
-#undef DRAM_TYPE_STR
-
-static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
-{
-	return dimm->ranks * 64 / (dimm->width ?: 1);
-}
-
-/* Returns total Gb for the whole DIMM */
-static int skl_get_dimm_size(u16 val)
-{
-	return (val & SKL_DRAM_SIZE_MASK) * 8;
-}
-
-static int skl_get_dimm_width(u16 val)
-{
-	if (skl_get_dimm_size(val) == 0)
-		return 0;
-
-	switch (val & SKL_DRAM_WIDTH_MASK) {
-	case SKL_DRAM_WIDTH_X8:
-	case SKL_DRAM_WIDTH_X16:
-	case SKL_DRAM_WIDTH_X32:
-		val = (val & SKL_DRAM_WIDTH_MASK) >> SKL_DRAM_WIDTH_SHIFT;
-		return 8 << val;
-	default:
-		MISSING_CASE(val);
-		return 0;
-	}
-}
-
-static int skl_get_dimm_ranks(u16 val)
-{
-	if (skl_get_dimm_size(val) == 0)
-		return 0;
-
-	val = (val & SKL_DRAM_RANK_MASK) >> SKL_DRAM_RANK_SHIFT;
-
-	return val + 1;
-}
-
-/* Returns total Gb for the whole DIMM */
-static int icl_get_dimm_size(u16 val)
-{
-	return (val & ICL_DRAM_SIZE_MASK) * 8 / 2;
-}
-
-static int icl_get_dimm_width(u16 val)
-{
-	if (icl_get_dimm_size(val) == 0)
-		return 0;
-
-	switch (val & ICL_DRAM_WIDTH_MASK) {
-	case ICL_DRAM_WIDTH_X8:
-	case ICL_DRAM_WIDTH_X16:
-	case ICL_DRAM_WIDTH_X32:
-		val = (val & ICL_DRAM_WIDTH_MASK) >> ICL_DRAM_WIDTH_SHIFT;
-		return 8 << val;
-	default:
-		MISSING_CASE(val);
-		return 0;
-	}
-}
-
-static int icl_get_dimm_ranks(u16 val)
-{
-	if (icl_get_dimm_size(val) == 0)
-		return 0;
-
-	val = (val & ICL_DRAM_RANK_MASK) >> ICL_DRAM_RANK_SHIFT;
-
-	return val + 1;
-}
-
-static bool
-skl_is_16gb_dimm(const struct dram_dimm_info *dimm)
-{
-	/* Convert total Gb to Gb per DRAM device */
-	return dimm->size / (intel_dimm_num_devices(dimm) ?: 1) == 16;
-}
-
-static void
-skl_dram_get_dimm_info(struct drm_i915_private *i915,
-		       struct dram_dimm_info *dimm,
-		       int channel, char dimm_name, u16 val)
-{
-	if (GRAPHICS_VER(i915) >= 11) {
-		dimm->size = icl_get_dimm_size(val);
-		dimm->width = icl_get_dimm_width(val);
-		dimm->ranks = icl_get_dimm_ranks(val);
-	} else {
-		dimm->size = skl_get_dimm_size(val);
-		dimm->width = skl_get_dimm_width(val);
-		dimm->ranks = skl_get_dimm_ranks(val);
-	}
-
-	drm_dbg_kms(&i915->drm,
-		    "CH%u DIMM %c size: %u Gb, width: X%u, ranks: %u, 16Gb DIMMs: %s\n",
-		    channel, dimm_name, dimm->size, dimm->width, dimm->ranks,
-		    str_yes_no(skl_is_16gb_dimm(dimm)));
-}
-
-static int
-skl_dram_get_channel_info(struct drm_i915_private *i915,
-			  struct dram_channel_info *ch,
-			  int channel, u32 val)
-{
-	skl_dram_get_dimm_info(i915, &ch->dimm_l,
-			       channel, 'L', val & 0xffff);
-	skl_dram_get_dimm_info(i915, &ch->dimm_s,
-			       channel, 'S', val >> 16);
-
-	if (ch->dimm_l.size == 0 && ch->dimm_s.size == 0) {
-		drm_dbg_kms(&i915->drm, "CH%u not populated\n", channel);
-		return -EINVAL;
-	}
-
-	if (ch->dimm_l.ranks == 2 || ch->dimm_s.ranks == 2)
-		ch->ranks = 2;
-	else if (ch->dimm_l.ranks == 1 && ch->dimm_s.ranks == 1)
-		ch->ranks = 2;
-	else
-		ch->ranks = 1;
-
-	ch->is_16gb_dimm = skl_is_16gb_dimm(&ch->dimm_l) ||
-		skl_is_16gb_dimm(&ch->dimm_s);
-
-	drm_dbg_kms(&i915->drm, "CH%u ranks: %u, 16Gb DIMMs: %s\n",
-		    channel, ch->ranks, str_yes_no(ch->is_16gb_dimm));
-
-	return 0;
-}
-
-static bool
-intel_is_dram_symmetric(const struct dram_channel_info *ch0,
-			const struct dram_channel_info *ch1)
-{
-	return !memcmp(ch0, ch1, sizeof(*ch0)) &&
-		(ch0->dimm_s.size == 0 ||
-		 !memcmp(&ch0->dimm_l, &ch0->dimm_s, sizeof(ch0->dimm_l)));
-}
-
-static int
-skl_dram_get_channels_info(struct drm_i915_private *i915)
-{
-	struct dram_info *dram_info = &i915->dram_info;
-	struct dram_channel_info ch0 = {}, ch1 = {};
-	u32 val;
-	int ret;
-
-	val = intel_de_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
-	ret = skl_dram_get_channel_info(i915, &ch0, 0, val);
-	if (ret == 0)
-		dram_info->num_channels++;
-
-	val = intel_de_read(i915, SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
-	ret = skl_dram_get_channel_info(i915, &ch1, 1, val);
-	if (ret == 0)
-		dram_info->num_channels++;
-
-	if (dram_info->num_channels == 0) {
-		drm_info(&i915->drm, "Number of memory channels is zero\n");
-		return -EINVAL;
-	}
-
-	if (ch0.ranks == 0 && ch1.ranks == 0) {
-		drm_info(&i915->drm, "couldn't get memory rank information\n");
-		return -EINVAL;
-	}
-
-	dram_info->wm_lv_0_adjust_needed = ch0.is_16gb_dimm || ch1.is_16gb_dimm;
-
-	dram_info->symmetric_memory = intel_is_dram_symmetric(&ch0, &ch1);
-
-	drm_dbg_kms(&i915->drm, "Memory configuration is symmetric? %s\n",
-		    str_yes_no(dram_info->symmetric_memory));
-
-	return 0;
-}
-
-static enum intel_dram_type
-skl_get_dram_type(struct drm_i915_private *i915)
-{
-	u32 val;
-
-	val = intel_de_read(i915, SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
-
-	switch (val & SKL_DRAM_DDR_TYPE_MASK) {
-	case SKL_DRAM_DDR_TYPE_DDR3:
-		return INTEL_DRAM_DDR3;
-	case SKL_DRAM_DDR_TYPE_DDR4:
-		return INTEL_DRAM_DDR4;
-	case SKL_DRAM_DDR_TYPE_LPDDR3:
-		return INTEL_DRAM_LPDDR3;
-	case SKL_DRAM_DDR_TYPE_LPDDR4:
-		return INTEL_DRAM_LPDDR4;
-	default:
-		MISSING_CASE(val);
-		return INTEL_DRAM_UNKNOWN;
-	}
-}
-
-static int
-skl_get_dram_info(struct drm_i915_private *i915)
-{
-	struct dram_info *dram_info = &i915->dram_info;
-	int ret;
-
-	dram_info->type = skl_get_dram_type(i915);
-	drm_dbg_kms(&i915->drm, "DRAM type: %s\n",
-		    intel_dram_type_str(dram_info->type));
-
-	ret = skl_dram_get_channels_info(i915);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-/* Returns Gb per DRAM device */
-static int bxt_get_dimm_size(u32 val)
-{
-	switch (val & BXT_DRAM_SIZE_MASK) {
-	case BXT_DRAM_SIZE_4GBIT:
-		return 4;
-	case BXT_DRAM_SIZE_6GBIT:
-		return 6;
-	case BXT_DRAM_SIZE_8GBIT:
-		return 8;
-	case BXT_DRAM_SIZE_12GBIT:
-		return 12;
-	case BXT_DRAM_SIZE_16GBIT:
-		return 16;
-	default:
-		MISSING_CASE(val);
-		return 0;
-	}
-}
-
-static int bxt_get_dimm_width(u32 val)
-{
-	if (!bxt_get_dimm_size(val))
-		return 0;
-
-	val = (val & BXT_DRAM_WIDTH_MASK) >> BXT_DRAM_WIDTH_SHIFT;
-
-	return 8 << val;
-}
-
-static int bxt_get_dimm_ranks(u32 val)
-{
-	if (!bxt_get_dimm_size(val))
-		return 0;
-
-	switch (val & BXT_DRAM_RANK_MASK) {
-	case BXT_DRAM_RANK_SINGLE:
-		return 1;
-	case BXT_DRAM_RANK_DUAL:
-		return 2;
-	default:
-		MISSING_CASE(val);
-		return 0;
-	}
-}
-
-static enum intel_dram_type bxt_get_dimm_type(u32 val)
-{
-	if (!bxt_get_dimm_size(val))
-		return INTEL_DRAM_UNKNOWN;
-
-	switch (val & BXT_DRAM_TYPE_MASK) {
-	case BXT_DRAM_TYPE_DDR3:
-		return INTEL_DRAM_DDR3;
-	case BXT_DRAM_TYPE_LPDDR3:
-		return INTEL_DRAM_LPDDR3;
-	case BXT_DRAM_TYPE_DDR4:
-		return INTEL_DRAM_DDR4;
-	case BXT_DRAM_TYPE_LPDDR4:
-		return INTEL_DRAM_LPDDR4;
-	default:
-		MISSING_CASE(val);
-		return INTEL_DRAM_UNKNOWN;
-	}
-}
-
-static void bxt_get_dimm_info(struct dram_dimm_info *dimm, u32 val)
-{
-	dimm->width = bxt_get_dimm_width(val);
-	dimm->ranks = bxt_get_dimm_ranks(val);
-
-	/*
-	 * Size in register is Gb per DRAM device. Convert to total
-	 * Gb to match the way we report this for non-LP platforms.
-	 */
-	dimm->size = bxt_get_dimm_size(val) * intel_dimm_num_devices(dimm);
-}
-
-static int bxt_get_dram_info(struct drm_i915_private *i915)
-{
-	struct dram_info *dram_info = &i915->dram_info;
-	u32 val;
-	u8 valid_ranks = 0;
-	int i;
-
-	/*
-	 * Now read each DUNIT8/9/10/11 to check the rank of each dimms.
-	 */
-	for (i = BXT_D_CR_DRP0_DUNIT_START; i <= BXT_D_CR_DRP0_DUNIT_END; i++) {
-		struct dram_dimm_info dimm;
-		enum intel_dram_type type;
-
-		val = intel_de_read(i915, BXT_D_CR_DRP0_DUNIT(i));
-		if (val == 0xFFFFFFFF)
-			continue;
-
-		dram_info->num_channels++;
-
-		bxt_get_dimm_info(&dimm, val);
-		type = bxt_get_dimm_type(val);
-
-		drm_WARN_ON(&i915->drm, type != INTEL_DRAM_UNKNOWN &&
-			    dram_info->type != INTEL_DRAM_UNKNOWN &&
-			    dram_info->type != type);
-
-		drm_dbg_kms(&i915->drm,
-			    "CH%u DIMM size: %u Gb, width: X%u, ranks: %u, type: %s\n",
-			    i - BXT_D_CR_DRP0_DUNIT_START,
-			    dimm.size, dimm.width, dimm.ranks,
-			    intel_dram_type_str(type));
-
-		if (valid_ranks == 0)
-			valid_ranks = dimm.ranks;
-
-		if (type != INTEL_DRAM_UNKNOWN)
-			dram_info->type = type;
-	}
-
-	if (dram_info->type == INTEL_DRAM_UNKNOWN || valid_ranks == 0) {
-		drm_info(&i915->drm, "couldn't get memory information\n");
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
-{
-	struct dram_info *dram_info = &dev_priv->dram_info;
-	u32 val = 0;
-	int ret;
-
-	ret = intel_de_pcode_read(dev_priv, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
-			     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
-	if (ret)
-		return ret;
-
-	if (GRAPHICS_VER(dev_priv) >= 12) {
-		switch (val & 0xf) {
-		case 0:
-			dram_info->type = INTEL_DRAM_DDR4;
-			break;
-		case 1:
-			dram_info->type = INTEL_DRAM_DDR5;
-			break;
-		case 2:
-			dram_info->type = INTEL_DRAM_LPDDR5;
-			break;
-		case 3:
-			dram_info->type = INTEL_DRAM_LPDDR4;
-			break;
-		case 4:
-			dram_info->type = INTEL_DRAM_DDR3;
-			break;
-		case 5:
-			dram_info->type = INTEL_DRAM_LPDDR3;
-			break;
-		default:
-			MISSING_CASE(val & 0xf);
-			return -EINVAL;
-		}
-	} else {
-		switch (val & 0xf) {
-		case 0:
-			dram_info->type = INTEL_DRAM_DDR4;
-			break;
-		case 1:
-			dram_info->type = INTEL_DRAM_DDR3;
-			break;
-		case 2:
-			dram_info->type = INTEL_DRAM_LPDDR3;
-			break;
-		case 3:
-			dram_info->type = INTEL_DRAM_LPDDR4;
-			break;
-		default:
-			MISSING_CASE(val & 0xf);
-			return -EINVAL;
-		}
-	}
-
-	dram_info->num_channels = (val & 0xf0) >> 4;
-	dram_info->num_qgv_points = (val & 0xf00) >> 8;
-	dram_info->num_psf_gv_points = (val & 0x3000) >> 12;
-
-	return 0;
-}
-
-static int gen11_get_dram_info(struct drm_i915_private *i915)
-{
-	int ret = skl_get_dram_info(i915);
-
-	if (ret)
-		return ret;
-
-	return icl_pcode_read_mem_global_info(i915);
-}
-
-static int gen12_get_dram_info(struct drm_i915_private *i915)
-{
-	i915->dram_info.wm_lv_0_adjust_needed = false;
-
-	return icl_pcode_read_mem_global_info(i915);
-}
-
-void intel_dram_detect(struct drm_i915_private *i915)
-{
-	struct dram_info *dram_info = &i915->dram_info;
-	int ret;
-
-	if (GRAPHICS_VER(i915) < 9 || IS_DG2(i915) || !HAS_DISPLAY(i915))
-		return;
-
-	/*
-	 * Assume level 0 watermark latency adjustment is needed until proven
-	 * otherwise, this w/a is not needed by bxt/glk.
-	 */
-	dram_info->wm_lv_0_adjust_needed = !IS_GEN9_LP(i915);
-
-	if (GRAPHICS_VER(i915) >= 12)
-		ret = gen12_get_dram_info(i915);
-	else if (GRAPHICS_VER(i915) >= 11)
-		ret = gen11_get_dram_info(i915);
-	else if (IS_GEN9_LP(i915))
-		ret = bxt_get_dram_info(i915);
-	else
-		ret = skl_get_dram_info(i915);
-	if (ret)
-		return;
-
-	drm_dbg_kms(&i915->drm, "DRAM channels: %u\n", dram_info->num_channels);
-
-	drm_dbg_kms(&i915->drm, "Watermark level 0 adjustment needed: %s\n",
-		    str_yes_no(dram_info->wm_lv_0_adjust_needed));
-}
diff --git a/drivers/gpu/drm/xe/display/ext/intel_pch.c b/drivers/gpu/drm/xe/display/ext/intel_pch.c
deleted file mode 100644
index dc2b15b5c4be..000000000000
--- a/drivers/gpu/drm/xe/display/ext/intel_pch.c
+++ /dev/null
@@ -1,157 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright 2019 Intel Corporation.
- */
-
-#include "i915_drv.h"
-#include "i915_utils.h"
-#include "intel_pch.h"
-
-/* Map PCH device id to PCH type, or PCH_NONE if unknown. */
-static enum intel_pch
-intel_pch_type(const struct drm_i915_private *dev_priv, unsigned short id)
-{
-	switch (id) {
-	case INTEL_PCH_MCC_DEVICE_ID_TYPE:
-		drm_dbg_kms(&dev_priv->drm, "Found Mule Creek Canyon PCH\n");
-		drm_WARN_ON(&dev_priv->drm, !IS_JSL_EHL(dev_priv));
-		/* MCC is TGP compatible */
-		return PCH_TGP;
-	case INTEL_PCH_TGP_DEVICE_ID_TYPE:
-	case INTEL_PCH_TGP2_DEVICE_ID_TYPE:
-		drm_dbg_kms(&dev_priv->drm, "Found Tiger Lake LP PCH\n");
-		drm_WARN_ON(&dev_priv->drm, !IS_TIGERLAKE(dev_priv) &&
-			    !IS_ROCKETLAKE(dev_priv) &&
-			    !IS_GEN9_BC(dev_priv));
-		return PCH_TGP;
-	case INTEL_PCH_ADP_DEVICE_ID_TYPE:
-	case INTEL_PCH_ADP2_DEVICE_ID_TYPE:
-	case INTEL_PCH_ADP3_DEVICE_ID_TYPE:
-	case INTEL_PCH_ADP4_DEVICE_ID_TYPE:
-		drm_dbg_kms(&dev_priv->drm, "Found Alder Lake PCH\n");
-		drm_WARN_ON(&dev_priv->drm, !IS_ALDERLAKE_S(dev_priv) &&
-			    !IS_ALDERLAKE_P(dev_priv));
-		return PCH_ADP;
-	case INTEL_PCH_MTP_DEVICE_ID_TYPE:
-	case INTEL_PCH_MTP2_DEVICE_ID_TYPE:
-		drm_dbg_kms(&dev_priv->drm, "Found Meteor Lake PCH\n");
-		drm_WARN_ON(&dev_priv->drm, !IS_METEORLAKE(dev_priv));
-		return PCH_MTP;
-	default:
-		return PCH_NONE;
-	}
-}
-
-static bool intel_is_virt_pch(unsigned short id,
-			      unsigned short svendor, unsigned short sdevice)
-{
-	return (id == INTEL_PCH_P2X_DEVICE_ID_TYPE ||
-		id == INTEL_PCH_P3X_DEVICE_ID_TYPE ||
-		(id == INTEL_PCH_QEMU_DEVICE_ID_TYPE &&
-		 svendor == PCI_SUBVENDOR_ID_REDHAT_QUMRANET &&
-		 sdevice == PCI_SUBDEVICE_ID_QEMU));
-}
-
-static void
-intel_virt_detect_pch(const struct drm_i915_private *dev_priv,
-		      unsigned short *pch_id, enum intel_pch *pch_type)
-{
-	unsigned short id = 0;
-
-	/*
-	 * In a virtualized passthrough environment we can be in a
-	 * setup where the ISA bridge is not able to be passed through.
-	 * In this case, a south bridge can be emulated and we have to
-	 * make an educated guess as to which PCH is really there.
-	 */
-
-	if (IS_METEORLAKE(dev_priv))
-		id = INTEL_PCH_MTP_DEVICE_ID_TYPE;
-	else if (IS_ALDERLAKE_S(dev_priv) || IS_ALDERLAKE_P(dev_priv))
-		id = INTEL_PCH_ADP_DEVICE_ID_TYPE;
-	else if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv))
-		id = INTEL_PCH_TGP_DEVICE_ID_TYPE;
-	else BUG_ON(1);
-
-	if (id)
-		drm_dbg_kms(&dev_priv->drm, "Assuming PCH ID %04x\n", id);
-	else
-		drm_dbg_kms(&dev_priv->drm, "Assuming no PCH\n");
-
-	*pch_type = intel_pch_type(dev_priv, id);
-
-	/* Sanity check virtual PCH id */
-	if (drm_WARN_ON(&dev_priv->drm,
-			id && *pch_type == PCH_NONE))
-		id = 0;
-
-	*pch_id = id;
-}
-
-void intel_detect_pch(struct drm_i915_private *dev_priv)
-{
-	struct pci_dev *pch = NULL;
-	unsigned short id;
-	enum intel_pch pch_type;
-
-	/* DG1 has south engine display on the same PCI device */
-	if (IS_DG1(dev_priv)) {
-		dev_priv->pch_type = PCH_DG1;
-		return;
-	} else if (IS_DG2(dev_priv)) {
-		dev_priv->pch_type = PCH_DG2;
-		return;
-	}
-
-	/*
-	 * The reason to probe ISA bridge instead of Dev31:Fun0 is to
-	 * make graphics device passthrough work easy for VMM, that only
-	 * need to expose ISA bridge to let driver know the real hardware
-	 * underneath. This is a requirement from virtualization team.
-	 *
-	 * In some virtualized environments (e.g. XEN), there is irrelevant
-	 * ISA bridge in the system. To work reliably, we should scan trhough
-	 * all the ISA bridge devices and check for the first match, instead
-	 * of only checking the first one.
-	 */
-	while ((pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, pch))) {
-		if (pch->vendor != PCI_VENDOR_ID_INTEL)
-			continue;
-
-		id = pch->device & INTEL_PCH_DEVICE_ID_MASK;
-
-		pch_type = intel_pch_type(dev_priv, id);
-		if (pch_type != PCH_NONE) {
-			dev_priv->pch_type = pch_type;
-			dev_priv->pch_id = id;
-			break;
-		} else if (intel_is_virt_pch(id, pch->subsystem_vendor,
-					     pch->subsystem_device)) {
-			intel_virt_detect_pch(dev_priv, &id, &pch_type);
-			dev_priv->pch_type = pch_type;
-			dev_priv->pch_id = id;
-			break;
-		}
-	}
-
-	/*
-	 * Use PCH_NOP (PCH but no South Display) for PCH platforms without
-	 * display.
-	 */
-	if (pch && !HAS_DISPLAY(dev_priv)) {
-		drm_dbg_kms(&dev_priv->drm,
-			    "Display disabled, reverting to NOP PCH\n");
-		dev_priv->pch_type = PCH_NOP;
-		dev_priv->pch_id = 0;
-	} else if (!pch) {
-		if (i915_run_as_guest() && HAS_DISPLAY(dev_priv)) {
-			intel_virt_detect_pch(dev_priv, &id, &pch_type);
-			dev_priv->pch_type = pch_type;
-			dev_priv->pch_id = id;
-		} else {
-			drm_dbg_kms(&dev_priv->drm, "No PCH found.\n");
-		}
-	}
-
-	pci_dev_put(pch);
-}
diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index 88f863edc41c..54de3229553b 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -18,7 +18,7 @@
 
 #if IS_ENABLED(CONFIG_DRM_XE_DISPLAY)
 #include "ext/intel_device_info.h"
-#include "ext/intel_pch.h"
+#include "soc/intel_pch.h"
 #include "intel_display_core.h"
 #endif
 
diff --git a/drivers/gpu/drm/xe/xe_display.c b/drivers/gpu/drm/xe/xe_display.c
index 64e65b29d8de..35dee29da832 100644
--- a/drivers/gpu/drm/xe/xe_display.c
+++ b/drivers/gpu/drm/xe/xe_display.c
@@ -16,8 +16,8 @@
 #include <drm/xe_drm.h>
 
 #include "ext/i915_irq.h"
-#include "ext/intel_dram.h"
 #include "ext/intel_pm.h"
+#include "soc/intel_dram.h"
 #include "intel_acpi.h"
 #include "intel_audio.h"
 #include "intel_bw.h"
-- 
2.34.1

Re: [Intel-xe] [PATCH 5/6] drm/xe: Build soc files directly

[Jani Nikula]

On Fri, 31 Mar 2023, Maarten Lankhorst <maarten.lankhorst@linux.intel.com> wrote:
> Instead of making a copy that is hard to keep up to date, build the
> source directly.
>
> Add a rule for soc_* files, and use those to build i915.
>
> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
> ---
>  drivers/gpu/drm/i915/soc/intel_dram.c         |  30 +-
>  drivers/gpu/drm/xe/Makefile                   |   8 +-
>  .../gpu/drm/xe/compat-i915-headers/i915_drv.h |   2 +-
>  .../soc}/intel_dram.h                         |   1 -
>  .../soc}/intel_pch.h                          |   0
>  drivers/gpu/drm/xe/display/ext/intel_dram.c   | 495 ------------------
>  drivers/gpu/drm/xe/display/ext/intel_pch.c    | 157 ------
>  drivers/gpu/drm/xe/xe_device_types.h          |   2 +-
>  drivers/gpu/drm/xe/xe_display.c               |   2 +-
>  9 files changed, 28 insertions(+), 669 deletions(-)
>  rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_dram.h (80%)
>  rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_pch.h (100%)
>  delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_dram.c
>  delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_pch.c
>
> diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
> index 9649051ed8ed..13fe18a9372f 100644
> --- a/drivers/gpu/drm/i915/soc/intel_dram.c
> +++ b/drivers/gpu/drm/i915/soc/intel_dram.c
> @@ -9,8 +9,12 @@
>  #include "i915_reg.h"
>  #include "intel_dram.h"
>  #include "intel_mchbar_regs.h"
> -#include "intel_pcode.h"
> +#ifdef I915
>  #include "vlv_sideband.h"
> +#include "display/intel_de.h"
> +#else
> +#include "xe_de.h"
> +#endif
>  
>  struct dram_dimm_info {
>  	u16 size;
> @@ -43,6 +47,8 @@ static const char *intel_dram_type_str(enum intel_dram_type type)
>  
>  #undef DRAM_TYPE_STR
>  
> +#ifdef I915
> +
>  static void pnv_detect_mem_freq(struct drm_i915_private *dev_priv)
>  {
>  	u32 tmp;
> @@ -191,6 +197,9 @@ static void detect_mem_freq(struct drm_i915_private *i915)
>  	if (i915->mem_freq)
>  		drm_dbg(&i915->drm, "DDR speed: %d MHz\n", i915->mem_freq);
>  }
> +#else
> +#define detect_mem_freq(i915) do { } while (0)
> +#endif
>  
>  static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
>  {
> @@ -339,14 +348,12 @@ skl_dram_get_channels_info(struct drm_i915_private *i915)
>  	u32 val;
>  	int ret;
>  
> -	val = intel_uncore_read(&i915->uncore,
> -				SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
> +	val = intel_de_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);

This isn't great, because those aren't DE registers. So this won't fly
upstream.

BR,
Jani.


>  	ret = skl_dram_get_channel_info(i915, &ch0, 0, val);
>  	if (ret == 0)
>  		dram_info->num_channels++;
>  
> -	val = intel_uncore_read(&i915->uncore,
> -				SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
> +	val = intel_de_read(i915, SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
>  	ret = skl_dram_get_channel_info(i915, &ch1, 1, val);
>  	if (ret == 0)
>  		dram_info->num_channels++;
> @@ -376,8 +383,7 @@ skl_get_dram_type(struct drm_i915_private *i915)
>  {
>  	u32 val;
>  
> -	val = intel_uncore_read(&i915->uncore,
> -				SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
> +	val = intel_de_read(i915, SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
>  
>  	switch (val & SKL_DRAM_DDR_TYPE_MASK) {
>  	case SKL_DRAM_DDR_TYPE_DDR3:
> @@ -503,7 +509,7 @@ static int bxt_get_dram_info(struct drm_i915_private *i915)
>  		struct dram_dimm_info dimm;
>  		enum intel_dram_type type;
>  
> -		val = intel_uncore_read(&i915->uncore, BXT_D_CR_DRP0_DUNIT(i));
> +		val = intel_de_read(i915, BXT_D_CR_DRP0_DUNIT(i));
>  		if (val == 0xFFFFFFFF)
>  			continue;
>  
> @@ -543,8 +549,8 @@ static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
>  	u32 val = 0;
>  	int ret;
>  
> -	ret = snb_pcode_read(&dev_priv->uncore, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
> -			     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
> +	ret = intel_de_pcode_read(dev_priv, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
> +				  ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
>  	if (ret)
>  		return ret;
>  
> @@ -618,7 +624,7 @@ static int gen12_get_dram_info(struct drm_i915_private *i915)
>  
>  static int xelpdp_get_dram_info(struct drm_i915_private *i915)
>  {
> -	u32 val = intel_uncore_read(&i915->uncore, MTL_MEM_SS_INFO_GLOBAL);
> +	u32 val = intel_de_read(i915, MTL_MEM_SS_INFO_GLOBAL);
>  	struct dram_info *dram_info = &i915->dram_info;
>  
>  	switch (REG_FIELD_GET(MTL_DDR_TYPE_MASK, val)) {
> @@ -687,6 +693,7 @@ void intel_dram_detect(struct drm_i915_private *i915)
>  		    str_yes_no(dram_info->wm_lv_0_adjust_needed));
>  }
>  
> +#ifdef I915
>  static u32 gen9_edram_size_mb(struct drm_i915_private *i915, u32 cap)
>  {
>  	static const u8 ways[8] = { 4, 8, 12, 16, 16, 16, 16, 16 };
> @@ -722,3 +729,4 @@ void intel_dram_edram_detect(struct drm_i915_private *i915)
>  
>  	drm_info(&i915->drm, "Found %uMB of eDRAM\n", i915->edram_size_mb);
>  }
> +#endif
> diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
> index 669c7b6f552c..3ebef784445e 100644
> --- a/drivers/gpu/drm/xe/Makefile
> +++ b/drivers/gpu/drm/xe/Makefile
> @@ -124,6 +124,10 @@ $(obj)/display/intel_%.o: $(srctree)/drivers/gpu/drm/i915/display/intel_%.c FORC
>  	$(call cmd,force_checksrc)
>  	$(call if_changed_rule,cc_o_c)
>  
> +$(obj)/display/soc_%.o: $(srctree)/drivers/gpu/drm/i915/soc/%.c FORCE
> +	$(call cmd,force_checksrc)
> +	$(call if_changed_rule,cc_o_c)
> +
>  # Display..
>  xe-$(CONFIG_DRM_XE_DISPLAY) += \
>  	xe_display.o \
> @@ -204,8 +208,8 @@ xe-$(CONFIG_DRM_XE_DISPLAY) += \
>  	display/ext/i915_irq.o \
>  	display/ext/i9xx_wm.o \
>  	display/ext/intel_device_info.o \
> -	display/ext/intel_dram.o \
> -	display/ext/intel_pch.o \
> +	display/soc_intel_dram.o \
> +	display/soc_intel_pch.o \
>  	display/ext/intel_pm.o
>  
>  ifeq ($(CONFIG_ACPI),y)
> diff --git a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> index 92dfbb8b3ca4..fa8c53e91bce 100644
> --- a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> +++ b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> @@ -15,7 +15,7 @@
>  #include "xe_pm.h"
>  #include "xe_step.h"
>  #include "i915_reg_defs.h"
> -#include "intel_pch.h"
> +#include "soc/intel_pch.h"
>  #include "i915_utils.h"
>  #include <linux/pm_runtime.h>
>  
> diff --git a/drivers/gpu/drm/xe/display/ext/intel_dram.h b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
> similarity index 80%
> rename from drivers/gpu/drm/xe/display/ext/intel_dram.h
> rename to drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
> index 4ba13c13162c..7ed5b6e1f805 100644
> --- a/drivers/gpu/drm/xe/display/ext/intel_dram.h
> +++ b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
> @@ -8,7 +8,6 @@
>  
>  struct drm_i915_private;
>  
> -void intel_dram_edram_detect(struct drm_i915_private *i915);
>  void intel_dram_detect(struct drm_i915_private *i915);
>  
>  #endif /* __INTEL_DRAM_H__ */
> diff --git a/drivers/gpu/drm/xe/display/ext/intel_pch.h b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h
> similarity index 100%
> rename from drivers/gpu/drm/xe/display/ext/intel_pch.h
> rename to drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h
> diff --git a/drivers/gpu/drm/xe/display/ext/intel_dram.c b/drivers/gpu/drm/xe/display/ext/intel_dram.c
> deleted file mode 100644
> index a2df0c502c34..000000000000
> --- a/drivers/gpu/drm/xe/display/ext/intel_dram.c
> +++ /dev/null
> @@ -1,495 +0,0 @@
> -// SPDX-License-Identifier: MIT
> -/*
> - * Copyright © 2020 Intel Corporation
> - */
> -
> -#include <linux/string_helpers.h>
> -
> -#include "i915_drv.h"
> -#include "i915_reg.h"
> -#include "intel_de.h"
> -#include "intel_dram.h"
> -#include "intel_mchbar_regs.h"
> -
> -struct dram_dimm_info {
> -	u16 size;
> -	u8 width, ranks;
> -};
> -
> -struct dram_channel_info {
> -	struct dram_dimm_info dimm_l, dimm_s;
> -	u8 ranks;
> -	bool is_16gb_dimm;
> -};
> -
> -#define DRAM_TYPE_STR(type) [INTEL_DRAM_ ## type] = #type
> -
> -static const char *intel_dram_type_str(enum intel_dram_type type)
> -{
> -	static const char * const str[] = {
> -		DRAM_TYPE_STR(UNKNOWN),
> -		DRAM_TYPE_STR(DDR3),
> -		DRAM_TYPE_STR(DDR4),
> -		DRAM_TYPE_STR(LPDDR3),
> -		DRAM_TYPE_STR(LPDDR4),
> -	};
> -
> -	if (type >= ARRAY_SIZE(str))
> -		type = INTEL_DRAM_UNKNOWN;
> -
> -	return str[type];
> -}
> -
> -#undef DRAM_TYPE_STR
> -
> -static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
> -{
> -	return dimm->ranks * 64 / (dimm->width ?: 1);
> -}
> -
> -/* Returns total Gb for the whole DIMM */
> -static int skl_get_dimm_size(u16 val)
> -{
> -	return (val & SKL_DRAM_SIZE_MASK) * 8;
> -}
> -
> -static int skl_get_dimm_width(u16 val)
> -{
> -	if (skl_get_dimm_size(val) == 0)
> -		return 0;
> -
> -	switch (val & SKL_DRAM_WIDTH_MASK) {
> -	case SKL_DRAM_WIDTH_X8:
> -	case SKL_DRAM_WIDTH_X16:
> -	case SKL_DRAM_WIDTH_X32:
> -		val = (val & SKL_DRAM_WIDTH_MASK) >> SKL_DRAM_WIDTH_SHIFT;
> -		return 8 << val;
> -	default:
> -		MISSING_CASE(val);
> -		return 0;
> -	}
> -}
> -
> -static int skl_get_dimm_ranks(u16 val)
> -{
> -	if (skl_get_dimm_size(val) == 0)
> -		return 0;
> -
> -	val = (val & SKL_DRAM_RANK_MASK) >> SKL_DRAM_RANK_SHIFT;
> -
> -	return val + 1;
> -}
> -
> -/* Returns total Gb for the whole DIMM */
> -static int icl_get_dimm_size(u16 val)
> -{
> -	return (val & ICL_DRAM_SIZE_MASK) * 8 / 2;
> -}
> -
> -static int icl_get_dimm_width(u16 val)
> -{
> -	if (icl_get_dimm_size(val) == 0)
> -		return 0;
> -
> -	switch (val & ICL_DRAM_WIDTH_MASK) {
> -	case ICL_DRAM_WIDTH_X8:
> -	case ICL_DRAM_WIDTH_X16:
> -	case ICL_DRAM_WIDTH_X32:
> -		val = (val & ICL_DRAM_WIDTH_MASK) >> ICL_DRAM_WIDTH_SHIFT;
> -		return 8 << val;
> -	default:
> -		MISSING_CASE(val);
> -		return 0;
> -	}
> -}
> -
> -static int icl_get_dimm_ranks(u16 val)
> -{
> -	if (icl_get_dimm_size(val) == 0)
> -		return 0;
> -
> -	val = (val & ICL_DRAM_RANK_MASK) >> ICL_DRAM_RANK_SHIFT;
> -
> -	return val + 1;
> -}
> -
> -static bool
> -skl_is_16gb_dimm(const struct dram_dimm_info *dimm)
> -{
> -	/* Convert total Gb to Gb per DRAM device */
> -	return dimm->size / (intel_dimm_num_devices(dimm) ?: 1) == 16;
> -}
> -
> -static void
> -skl_dram_get_dimm_info(struct drm_i915_private *i915,
> -		       struct dram_dimm_info *dimm,
> -		       int channel, char dimm_name, u16 val)
> -{
> -	if (GRAPHICS_VER(i915) >= 11) {
> -		dimm->size = icl_get_dimm_size(val);
> -		dimm->width = icl_get_dimm_width(val);
> -		dimm->ranks = icl_get_dimm_ranks(val);
> -	} else {
> -		dimm->size = skl_get_dimm_size(val);
> -		dimm->width = skl_get_dimm_width(val);
> -		dimm->ranks = skl_get_dimm_ranks(val);
> -	}
> -
> -	drm_dbg_kms(&i915->drm,
> -		    "CH%u DIMM %c size: %u Gb, width: X%u, ranks: %u, 16Gb DIMMs: %s\n",
> -		    channel, dimm_name, dimm->size, dimm->width, dimm->ranks,
> -		    str_yes_no(skl_is_16gb_dimm(dimm)));
> -}
> -
> -static int
> -skl_dram_get_channel_info(struct drm_i915_private *i915,
> -			  struct dram_channel_info *ch,
> -			  int channel, u32 val)
> -{
> -	skl_dram_get_dimm_info(i915, &ch->dimm_l,
> -			       channel, 'L', val & 0xffff);
> -	skl_dram_get_dimm_info(i915, &ch->dimm_s,
> -			       channel, 'S', val >> 16);
> -
> -	if (ch->dimm_l.size == 0 && ch->dimm_s.size == 0) {
> -		drm_dbg_kms(&i915->drm, "CH%u not populated\n", channel);
> -		return -EINVAL;
> -	}
> -
> -	if (ch->dimm_l.ranks == 2 || ch->dimm_s.ranks == 2)
> -		ch->ranks = 2;
> -	else if (ch->dimm_l.ranks == 1 && ch->dimm_s.ranks == 1)
> -		ch->ranks = 2;
> -	else
> -		ch->ranks = 1;
> -
> -	ch->is_16gb_dimm = skl_is_16gb_dimm(&ch->dimm_l) ||
> -		skl_is_16gb_dimm(&ch->dimm_s);
> -
> -	drm_dbg_kms(&i915->drm, "CH%u ranks: %u, 16Gb DIMMs: %s\n",
> -		    channel, ch->ranks, str_yes_no(ch->is_16gb_dimm));
> -
> -	return 0;
> -}
> -
> -static bool
> -intel_is_dram_symmetric(const struct dram_channel_info *ch0,
> -			const struct dram_channel_info *ch1)
> -{
> -	return !memcmp(ch0, ch1, sizeof(*ch0)) &&
> -		(ch0->dimm_s.size == 0 ||
> -		 !memcmp(&ch0->dimm_l, &ch0->dimm_s, sizeof(ch0->dimm_l)));
> -}
> -
> -static int
> -skl_dram_get_channels_info(struct drm_i915_private *i915)
> -{
> -	struct dram_info *dram_info = &i915->dram_info;
> -	struct dram_channel_info ch0 = {}, ch1 = {};
> -	u32 val;
> -	int ret;
> -
> -	val = intel_de_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
> -	ret = skl_dram_get_channel_info(i915, &ch0, 0, val);
> -	if (ret == 0)
> -		dram_info->num_channels++;
> -
> -	val = intel_de_read(i915, SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
> -	ret = skl_dram_get_channel_info(i915, &ch1, 1, val);
> -	if (ret == 0)
> -		dram_info->num_channels++;
> -
> -	if (dram_info->num_channels == 0) {
> -		drm_info(&i915->drm, "Number of memory channels is zero\n");
> -		return -EINVAL;
> -	}
> -
> -	if (ch0.ranks == 0 && ch1.ranks == 0) {
> -		drm_info(&i915->drm, "couldn't get memory rank information\n");
> -		return -EINVAL;
> -	}
> -
> -	dram_info->wm_lv_0_adjust_needed = ch0.is_16gb_dimm || ch1.is_16gb_dimm;
> -
> -	dram_info->symmetric_memory = intel_is_dram_symmetric(&ch0, &ch1);
> -
> -	drm_dbg_kms(&i915->drm, "Memory configuration is symmetric? %s\n",
> -		    str_yes_no(dram_info->symmetric_memory));
> -
> -	return 0;
> -}
> -
> -static enum intel_dram_type
> -skl_get_dram_type(struct drm_i915_private *i915)
> -{
> -	u32 val;
> -
> -	val = intel_de_read(i915, SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
> -
> -	switch (val & SKL_DRAM_DDR_TYPE_MASK) {
> -	case SKL_DRAM_DDR_TYPE_DDR3:
> -		return INTEL_DRAM_DDR3;
> -	case SKL_DRAM_DDR_TYPE_DDR4:
> -		return INTEL_DRAM_DDR4;
> -	case SKL_DRAM_DDR_TYPE_LPDDR3:
> -		return INTEL_DRAM_LPDDR3;
> -	case SKL_DRAM_DDR_TYPE_LPDDR4:
> -		return INTEL_DRAM_LPDDR4;
> -	default:
> -		MISSING_CASE(val);
> -		return INTEL_DRAM_UNKNOWN;
> -	}
> -}
> -
> -static int
> -skl_get_dram_info(struct drm_i915_private *i915)
> -{
> -	struct dram_info *dram_info = &i915->dram_info;
> -	int ret;
> -
> -	dram_info->type = skl_get_dram_type(i915);
> -	drm_dbg_kms(&i915->drm, "DRAM type: %s\n",
> -		    intel_dram_type_str(dram_info->type));
> -
> -	ret = skl_dram_get_channels_info(i915);
> -	if (ret)
> -		return ret;
> -
> -	return 0;
> -}
> -
> -/* Returns Gb per DRAM device */
> -static int bxt_get_dimm_size(u32 val)
> -{
> -	switch (val & BXT_DRAM_SIZE_MASK) {
> -	case BXT_DRAM_SIZE_4GBIT:
> -		return 4;
> -	case BXT_DRAM_SIZE_6GBIT:
> -		return 6;
> -	case BXT_DRAM_SIZE_8GBIT:
> -		return 8;
> -	case BXT_DRAM_SIZE_12GBIT:
> -		return 12;
> -	case BXT_DRAM_SIZE_16GBIT:
> -		return 16;
> -	default:
> -		MISSING_CASE(val);
> -		return 0;
> -	}
> -}
> -
> -static int bxt_get_dimm_width(u32 val)
> -{
> -	if (!bxt_get_dimm_size(val))
> -		return 0;
> -
> -	val = (val & BXT_DRAM_WIDTH_MASK) >> BXT_DRAM_WIDTH_SHIFT;
> -
> -	return 8 << val;
> -}
> -
> -static int bxt_get_dimm_ranks(u32 val)
> -{
> -	if (!bxt_get_dimm_size(val))
> -		return 0;
> -
> -	switch (val & BXT_DRAM_RANK_MASK) {
> -	case BXT_DRAM_RANK_SINGLE:
> -		return 1;
> -	case BXT_DRAM_RANK_DUAL:
> -		return 2;
> -	default:
> -		MISSING_CASE(val);
> -		return 0;
> -	}
> -}
> -
> -static enum intel_dram_type bxt_get_dimm_type(u32 val)
> -{
> -	if (!bxt_get_dimm_size(val))
> -		return INTEL_DRAM_UNKNOWN;
> -
> -	switch (val & BXT_DRAM_TYPE_MASK) {
> -	case BXT_DRAM_TYPE_DDR3:
> -		return INTEL_DRAM_DDR3;
> -	case BXT_DRAM_TYPE_LPDDR3:
> -		return INTEL_DRAM_LPDDR3;
> -	case BXT_DRAM_TYPE_DDR4:
> -		return INTEL_DRAM_DDR4;
> -	case BXT_DRAM_TYPE_LPDDR4:
> -		return INTEL_DRAM_LPDDR4;
> -	default:
> -		MISSING_CASE(val);
> -		return INTEL_DRAM_UNKNOWN;
> -	}
> -}
> -
> -static void bxt_get_dimm_info(struct dram_dimm_info *dimm, u32 val)
> -{
> -	dimm->width = bxt_get_dimm_width(val);
> -	dimm->ranks = bxt_get_dimm_ranks(val);
> -
> -	/*
> -	 * Size in register is Gb per DRAM device. Convert to total
> -	 * Gb to match the way we report this for non-LP platforms.
> -	 */
> -	dimm->size = bxt_get_dimm_size(val) * intel_dimm_num_devices(dimm);
> -}
> -
> -static int bxt_get_dram_info(struct drm_i915_private *i915)
> -{
> -	struct dram_info *dram_info = &i915->dram_info;
> -	u32 val;
> -	u8 valid_ranks = 0;
> -	int i;
> -
> -	/*
> -	 * Now read each DUNIT8/9/10/11 to check the rank of each dimms.
> -	 */
> -	for (i = BXT_D_CR_DRP0_DUNIT_START; i <= BXT_D_CR_DRP0_DUNIT_END; i++) {
> -		struct dram_dimm_info dimm;
> -		enum intel_dram_type type;
> -
> -		val = intel_de_read(i915, BXT_D_CR_DRP0_DUNIT(i));
> -		if (val == 0xFFFFFFFF)
> -			continue;
> -
> -		dram_info->num_channels++;
> -
> -		bxt_get_dimm_info(&dimm, val);
> -		type = bxt_get_dimm_type(val);
> -
> -		drm_WARN_ON(&i915->drm, type != INTEL_DRAM_UNKNOWN &&
> -			    dram_info->type != INTEL_DRAM_UNKNOWN &&
> -			    dram_info->type != type);
> -
> -		drm_dbg_kms(&i915->drm,
> -			    "CH%u DIMM size: %u Gb, width: X%u, ranks: %u, type: %s\n",
> -			    i - BXT_D_CR_DRP0_DUNIT_START,
> -			    dimm.size, dimm.width, dimm.ranks,
> -			    intel_dram_type_str(type));
> -
> -		if (valid_ranks == 0)
> -			valid_ranks = dimm.ranks;
> -
> -		if (type != INTEL_DRAM_UNKNOWN)
> -			dram_info->type = type;
> -	}
> -
> -	if (dram_info->type == INTEL_DRAM_UNKNOWN || valid_ranks == 0) {
> -		drm_info(&i915->drm, "couldn't get memory information\n");
> -		return -EINVAL;
> -	}
> -
> -	return 0;
> -}
> -
> -static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
> -{
> -	struct dram_info *dram_info = &dev_priv->dram_info;
> -	u32 val = 0;
> -	int ret;
> -
> -	ret = intel_de_pcode_read(dev_priv, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
> -			     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
> -	if (ret)
> -		return ret;
> -
> -	if (GRAPHICS_VER(dev_priv) >= 12) {
> -		switch (val & 0xf) {
> -		case 0:
> -			dram_info->type = INTEL_DRAM_DDR4;
> -			break;
> -		case 1:
> -			dram_info->type = INTEL_DRAM_DDR5;
> -			break;
> -		case 2:
> -			dram_info->type = INTEL_DRAM_LPDDR5;
> -			break;
> -		case 3:
> -			dram_info->type = INTEL_DRAM_LPDDR4;
> -			break;
> -		case 4:
> -			dram_info->type = INTEL_DRAM_DDR3;
> -			break;
> -		case 5:
> -			dram_info->type = INTEL_DRAM_LPDDR3;
> -			break;
> -		default:
> -			MISSING_CASE(val & 0xf);
> -			return -EINVAL;
> -		}
> -	} else {
> -		switch (val & 0xf) {
> -		case 0:
> -			dram_info->type = INTEL_DRAM_DDR4;
> -			break;
> -		case 1:
> -			dram_info->type = INTEL_DRAM_DDR3;
> -			break;
> -		case 2:
> -			dram_info->type = INTEL_DRAM_LPDDR3;
> -			break;
> -		case 3:
> -			dram_info->type = INTEL_DRAM_LPDDR4;
> -			break;
> -		default:
> -			MISSING_CASE(val & 0xf);
> -			return -EINVAL;
> -		}
> -	}
> -
> -	dram_info->num_channels = (val & 0xf0) >> 4;
> -	dram_info->num_qgv_points = (val & 0xf00) >> 8;
> -	dram_info->num_psf_gv_points = (val & 0x3000) >> 12;
> -
> -	return 0;
> -}
> -
> -static int gen11_get_dram_info(struct drm_i915_private *i915)
> -{
> -	int ret = skl_get_dram_info(i915);
> -
> -	if (ret)
> -		return ret;
> -
> -	return icl_pcode_read_mem_global_info(i915);
> -}
> -
> -static int gen12_get_dram_info(struct drm_i915_private *i915)
> -{
> -	i915->dram_info.wm_lv_0_adjust_needed = false;
> -
> -	return icl_pcode_read_mem_global_info(i915);
> -}
> -
> -void intel_dram_detect(struct drm_i915_private *i915)
> -{
> -	struct dram_info *dram_info = &i915->dram_info;
> -	int ret;
> -
> -	if (GRAPHICS_VER(i915) < 9 || IS_DG2(i915) || !HAS_DISPLAY(i915))
> -		return;
> -
> -	/*
> -	 * Assume level 0 watermark latency adjustment is needed until proven
> -	 * otherwise, this w/a is not needed by bxt/glk.
> -	 */
> -	dram_info->wm_lv_0_adjust_needed = !IS_GEN9_LP(i915);
> -
> -	if (GRAPHICS_VER(i915) >= 12)
> -		ret = gen12_get_dram_info(i915);
> -	else if (GRAPHICS_VER(i915) >= 11)
> -		ret = gen11_get_dram_info(i915);
> -	else if (IS_GEN9_LP(i915))
> -		ret = bxt_get_dram_info(i915);
> -	else
> -		ret = skl_get_dram_info(i915);
> -	if (ret)
> -		return;
> -
> -	drm_dbg_kms(&i915->drm, "DRAM channels: %u\n", dram_info->num_channels);
> -
> -	drm_dbg_kms(&i915->drm, "Watermark level 0 adjustment needed: %s\n",
> -		    str_yes_no(dram_info->wm_lv_0_adjust_needed));
> -}
> diff --git a/drivers/gpu/drm/xe/display/ext/intel_pch.c b/drivers/gpu/drm/xe/display/ext/intel_pch.c
> deleted file mode 100644
> index dc2b15b5c4be..000000000000
> --- a/drivers/gpu/drm/xe/display/ext/intel_pch.c
> +++ /dev/null
> @@ -1,157 +0,0 @@
> -// SPDX-License-Identifier: MIT
> -/*
> - * Copyright 2019 Intel Corporation.
> - */
> -
> -#include "i915_drv.h"
> -#include "i915_utils.h"
> -#include "intel_pch.h"
> -
> -/* Map PCH device id to PCH type, or PCH_NONE if unknown. */
> -static enum intel_pch
> -intel_pch_type(const struct drm_i915_private *dev_priv, unsigned short id)
> -{
> -	switch (id) {
> -	case INTEL_PCH_MCC_DEVICE_ID_TYPE:
> -		drm_dbg_kms(&dev_priv->drm, "Found Mule Creek Canyon PCH\n");
> -		drm_WARN_ON(&dev_priv->drm, !IS_JSL_EHL(dev_priv));
> -		/* MCC is TGP compatible */
> -		return PCH_TGP;
> -	case INTEL_PCH_TGP_DEVICE_ID_TYPE:
> -	case INTEL_PCH_TGP2_DEVICE_ID_TYPE:
> -		drm_dbg_kms(&dev_priv->drm, "Found Tiger Lake LP PCH\n");
> -		drm_WARN_ON(&dev_priv->drm, !IS_TIGERLAKE(dev_priv) &&
> -			    !IS_ROCKETLAKE(dev_priv) &&
> -			    !IS_GEN9_BC(dev_priv));
> -		return PCH_TGP;
> -	case INTEL_PCH_ADP_DEVICE_ID_TYPE:
> -	case INTEL_PCH_ADP2_DEVICE_ID_TYPE:
> -	case INTEL_PCH_ADP3_DEVICE_ID_TYPE:
> -	case INTEL_PCH_ADP4_DEVICE_ID_TYPE:
> -		drm_dbg_kms(&dev_priv->drm, "Found Alder Lake PCH\n");
> -		drm_WARN_ON(&dev_priv->drm, !IS_ALDERLAKE_S(dev_priv) &&
> -			    !IS_ALDERLAKE_P(dev_priv));
> -		return PCH_ADP;
> -	case INTEL_PCH_MTP_DEVICE_ID_TYPE:
> -	case INTEL_PCH_MTP2_DEVICE_ID_TYPE:
> -		drm_dbg_kms(&dev_priv->drm, "Found Meteor Lake PCH\n");
> -		drm_WARN_ON(&dev_priv->drm, !IS_METEORLAKE(dev_priv));
> -		return PCH_MTP;
> -	default:
> -		return PCH_NONE;
> -	}
> -}
> -
> -static bool intel_is_virt_pch(unsigned short id,
> -			      unsigned short svendor, unsigned short sdevice)
> -{
> -	return (id == INTEL_PCH_P2X_DEVICE_ID_TYPE ||
> -		id == INTEL_PCH_P3X_DEVICE_ID_TYPE ||
> -		(id == INTEL_PCH_QEMU_DEVICE_ID_TYPE &&
> -		 svendor == PCI_SUBVENDOR_ID_REDHAT_QUMRANET &&
> -		 sdevice == PCI_SUBDEVICE_ID_QEMU));
> -}
> -
> -static void
> -intel_virt_detect_pch(const struct drm_i915_private *dev_priv,
> -		      unsigned short *pch_id, enum intel_pch *pch_type)
> -{
> -	unsigned short id = 0;
> -
> -	/*
> -	 * In a virtualized passthrough environment we can be in a
> -	 * setup where the ISA bridge is not able to be passed through.
> -	 * In this case, a south bridge can be emulated and we have to
> -	 * make an educated guess as to which PCH is really there.
> -	 */
> -
> -	if (IS_METEORLAKE(dev_priv))
> -		id = INTEL_PCH_MTP_DEVICE_ID_TYPE;
> -	else if (IS_ALDERLAKE_S(dev_priv) || IS_ALDERLAKE_P(dev_priv))
> -		id = INTEL_PCH_ADP_DEVICE_ID_TYPE;
> -	else if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv))
> -		id = INTEL_PCH_TGP_DEVICE_ID_TYPE;
> -	else BUG_ON(1);
> -
> -	if (id)
> -		drm_dbg_kms(&dev_priv->drm, "Assuming PCH ID %04x\n", id);
> -	else
> -		drm_dbg_kms(&dev_priv->drm, "Assuming no PCH\n");
> -
> -	*pch_type = intel_pch_type(dev_priv, id);
> -
> -	/* Sanity check virtual PCH id */
> -	if (drm_WARN_ON(&dev_priv->drm,
> -			id && *pch_type == PCH_NONE))
> -		id = 0;
> -
> -	*pch_id = id;
> -}
> -
> -void intel_detect_pch(struct drm_i915_private *dev_priv)
> -{
> -	struct pci_dev *pch = NULL;
> -	unsigned short id;
> -	enum intel_pch pch_type;
> -
> -	/* DG1 has south engine display on the same PCI device */
> -	if (IS_DG1(dev_priv)) {
> -		dev_priv->pch_type = PCH_DG1;
> -		return;
> -	} else if (IS_DG2(dev_priv)) {
> -		dev_priv->pch_type = PCH_DG2;
> -		return;
> -	}
> -
> -	/*
> -	 * The reason to probe ISA bridge instead of Dev31:Fun0 is to
> -	 * make graphics device passthrough work easy for VMM, that only
> -	 * need to expose ISA bridge to let driver know the real hardware
> -	 * underneath. This is a requirement from virtualization team.
> -	 *
> -	 * In some virtualized environments (e.g. XEN), there is irrelevant
> -	 * ISA bridge in the system. To work reliably, we should scan trhough
> -	 * all the ISA bridge devices and check for the first match, instead
> -	 * of only checking the first one.
> -	 */
> -	while ((pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, pch))) {
> -		if (pch->vendor != PCI_VENDOR_ID_INTEL)
> -			continue;
> -
> -		id = pch->device & INTEL_PCH_DEVICE_ID_MASK;
> -
> -		pch_type = intel_pch_type(dev_priv, id);
> -		if (pch_type != PCH_NONE) {
> -			dev_priv->pch_type = pch_type;
> -			dev_priv->pch_id = id;
> -			break;
> -		} else if (intel_is_virt_pch(id, pch->subsystem_vendor,
> -					     pch->subsystem_device)) {
> -			intel_virt_detect_pch(dev_priv, &id, &pch_type);
> -			dev_priv->pch_type = pch_type;
> -			dev_priv->pch_id = id;
> -			break;
> -		}
> -	}
> -
> -	/*
> -	 * Use PCH_NOP (PCH but no South Display) for PCH platforms without
> -	 * display.
> -	 */
> -	if (pch && !HAS_DISPLAY(dev_priv)) {
> -		drm_dbg_kms(&dev_priv->drm,
> -			    "Display disabled, reverting to NOP PCH\n");
> -		dev_priv->pch_type = PCH_NOP;
> -		dev_priv->pch_id = 0;
> -	} else if (!pch) {
> -		if (i915_run_as_guest() && HAS_DISPLAY(dev_priv)) {
> -			intel_virt_detect_pch(dev_priv, &id, &pch_type);
> -			dev_priv->pch_type = pch_type;
> -			dev_priv->pch_id = id;
> -		} else {
> -			drm_dbg_kms(&dev_priv->drm, "No PCH found.\n");
> -		}
> -	}
> -
> -	pci_dev_put(pch);
> -}
> diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
> index 88f863edc41c..54de3229553b 100644
> --- a/drivers/gpu/drm/xe/xe_device_types.h
> +++ b/drivers/gpu/drm/xe/xe_device_types.h
> @@ -18,7 +18,7 @@
>  
>  #if IS_ENABLED(CONFIG_DRM_XE_DISPLAY)
>  #include "ext/intel_device_info.h"
> -#include "ext/intel_pch.h"
> +#include "soc/intel_pch.h"
>  #include "intel_display_core.h"
>  #endif
>  
> diff --git a/drivers/gpu/drm/xe/xe_display.c b/drivers/gpu/drm/xe/xe_display.c
> index 64e65b29d8de..35dee29da832 100644
> --- a/drivers/gpu/drm/xe/xe_display.c
> +++ b/drivers/gpu/drm/xe/xe_display.c
> @@ -16,8 +16,8 @@
>  #include <drm/xe_drm.h>
>  
>  #include "ext/i915_irq.h"
> -#include "ext/intel_dram.h"
>  #include "ext/intel_pm.h"
> +#include "soc/intel_dram.h"
>  #include "intel_acpi.h"
>  #include "intel_audio.h"
>  #include "intel_bw.h"

-- 
Jani Nikula, Intel Open Source Graphics Center

Re: [Intel-xe] [PATCH 5/6] drm/xe: Build soc files directly

[Maarten Lankhorst]

On 2023-03-31 13:49, Jani Nikula wrote:
> On Fri, 31 Mar 2023, Maarten Lankhorst <maarten.lankhorst@linux.intel.com> wrote:
>> Instead of making a copy that is hard to keep up to date, build the
>> source directly.
>>
>> Add a rule for soc_* files, and use those to build i915.
>>
>> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
>> ---
>>   drivers/gpu/drm/i915/soc/intel_dram.c         |  30 +-
>>   drivers/gpu/drm/xe/Makefile                   |   8 +-
>>   .../gpu/drm/xe/compat-i915-headers/i915_drv.h |   2 +-
>>   .../soc}/intel_dram.h                         |   1 -
>>   .../soc}/intel_pch.h                          |   0
>>   drivers/gpu/drm/xe/display/ext/intel_dram.c   | 495 ------------------
>>   drivers/gpu/drm/xe/display/ext/intel_pch.c    | 157 ------
>>   drivers/gpu/drm/xe/xe_device_types.h          |   2 +-
>>   drivers/gpu/drm/xe/xe_display.c               |   2 +-
>>   9 files changed, 28 insertions(+), 669 deletions(-)
>>   rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_dram.h (80%)
>>   rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_pch.h (100%)
>>   delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_dram.c
>>   delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_pch.c
>>
>> diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
>> index 9649051ed8ed..13fe18a9372f 100644
>> --- a/drivers/gpu/drm/i915/soc/intel_dram.c
>> +++ b/drivers/gpu/drm/i915/soc/intel_dram.c
>> @@ -9,8 +9,12 @@
>>   #include "i915_reg.h"
>>   #include "intel_dram.h"
>>   #include "intel_mchbar_regs.h"
>> -#include "intel_pcode.h"
>> +#ifdef I915
>>   #include "vlv_sideband.h"
>> +#include "display/intel_de.h"
>> +#else
>> +#include "xe_de.h"
>> +#endif
>>   
>>   struct dram_dimm_info {
>>   	u16 size;
>> @@ -43,6 +47,8 @@ static const char *intel_dram_type_str(enum intel_dram_type type)
>>   
>>   #undef DRAM_TYPE_STR
>>   
>> +#ifdef I915
>> +
>>   static void pnv_detect_mem_freq(struct drm_i915_private *dev_priv)
>>   {
>>   	u32 tmp;
>> @@ -191,6 +197,9 @@ static void detect_mem_freq(struct drm_i915_private *i915)
>>   	if (i915->mem_freq)
>>   		drm_dbg(&i915->drm, "DDR speed: %d MHz\n", i915->mem_freq);
>>   }
>> +#else
>> +#define detect_mem_freq(i915) do { } while (0)
>> +#endif
>>   
>>   static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
>>   {
>> @@ -339,14 +348,12 @@ skl_dram_get_channels_info(struct drm_i915_private *i915)
>>   	u32 val;
>>   	int ret;
>>   
>> -	val = intel_uncore_read(&i915->uncore,
>> -				SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
>> +	val = intel_de_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
> This isn't great, because those aren't DE registers. So this won't fly
> upstream.

Could we do this for xe instead?

#define intel_uncore_read(ignore, reg) xe_mmio_read32(i915, reg) ?

Re: [Intel-xe] [PATCH 5/6] drm/xe: Build soc files directly

[Jani Nikula]

On Fri, 31 Mar 2023, Maarten Lankhorst <maarten.lankhorst@linux.intel.com> wrote:
> On 2023-03-31 13:49, Jani Nikula wrote:
>> On Fri, 31 Mar 2023, Maarten Lankhorst <maarten.lankhorst@linux.intel.com> wrote:
>>> Instead of making a copy that is hard to keep up to date, build the
>>> source directly.
>>>
>>> Add a rule for soc_* files, and use those to build i915.
>>>
>>> Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
>>> ---
>>>   drivers/gpu/drm/i915/soc/intel_dram.c         |  30 +-
>>>   drivers/gpu/drm/xe/Makefile                   |   8 +-
>>>   .../gpu/drm/xe/compat-i915-headers/i915_drv.h |   2 +-
>>>   .../soc}/intel_dram.h                         |   1 -
>>>   .../soc}/intel_pch.h                          |   0
>>>   drivers/gpu/drm/xe/display/ext/intel_dram.c   | 495 ------------------
>>>   drivers/gpu/drm/xe/display/ext/intel_pch.c    | 157 ------
>>>   drivers/gpu/drm/xe/xe_device_types.h          |   2 +-
>>>   drivers/gpu/drm/xe/xe_display.c               |   2 +-
>>>   9 files changed, 28 insertions(+), 669 deletions(-)
>>>   rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_dram.h (80%)
>>>   rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_pch.h (100%)
>>>   delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_dram.c
>>>   delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_pch.c
>>>
>>> diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
>>> index 9649051ed8ed..13fe18a9372f 100644
>>> --- a/drivers/gpu/drm/i915/soc/intel_dram.c
>>> +++ b/drivers/gpu/drm/i915/soc/intel_dram.c
>>> @@ -9,8 +9,12 @@
>>>   #include "i915_reg.h"
>>>   #include "intel_dram.h"
>>>   #include "intel_mchbar_regs.h"
>>> -#include "intel_pcode.h"
>>> +#ifdef I915
>>>   #include "vlv_sideband.h"
>>> +#include "display/intel_de.h"
>>> +#else
>>> +#include "xe_de.h"
>>> +#endif
>>>   
>>>   struct dram_dimm_info {
>>>   	u16 size;
>>> @@ -43,6 +47,8 @@ static const char *intel_dram_type_str(enum intel_dram_type type)
>>>   
>>>   #undef DRAM_TYPE_STR
>>>   
>>> +#ifdef I915
>>> +
>>>   static void pnv_detect_mem_freq(struct drm_i915_private *dev_priv)
>>>   {
>>>   	u32 tmp;
>>> @@ -191,6 +197,9 @@ static void detect_mem_freq(struct drm_i915_private *i915)
>>>   	if (i915->mem_freq)
>>>   		drm_dbg(&i915->drm, "DDR speed: %d MHz\n", i915->mem_freq);
>>>   }
>>> +#else
>>> +#define detect_mem_freq(i915) do { } while (0)
>>> +#endif
>>>   
>>>   static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
>>>   {
>>> @@ -339,14 +348,12 @@ skl_dram_get_channels_info(struct drm_i915_private *i915)
>>>   	u32 val;
>>>   	int ret;
>>>   
>>> -	val = intel_uncore_read(&i915->uncore,
>>> -				SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
>>> +	val = intel_de_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
>> This isn't great, because those aren't DE registers. So this won't fly
>> upstream.
>
> Could we do this for xe instead?
>
> #define intel_uncore_read(ignore, reg) xe_mmio_read32(i915, reg) ?

The obvious downside is that it relies on the implict variable i915,
which could also be named something else, and it has taken a long, long
time trying to fix those in i915.

BR,
Jani.


-- 
Jani Nikula, Intel Open Source Graphics Center

[Intel-xe] [PATCH v2 5/6] drm/xe: Build soc files directly

[Maarten Lankhorst]

[-- Attachment #1: Type: text/plain, Size: 28694 bytes --]


On 2023-04-03 10:12, Jani Nikula wrote:
> On Fri, 31 Mar 2023, Maarten Lankhorst<maarten.lankhorst@linux.intel.com>  wrote:
>> On 2023-03-31 13:49, Jani Nikula wrote:
>>> On Fri, 31 Mar 2023, Maarten Lankhorst<maarten.lankhorst@linux.intel.com>  wrote:
>>>> Instead of making a copy that is hard to keep up to date, build the
>>>> source directly.
>>>>
>>>> Add a rule for soc_* files, and use those to build i915.
>>>>
>>>> Signed-off-by: Maarten Lankhorst<maarten.lankhorst@linux.intel.com>
>>>> ---
>>>>    drivers/gpu/drm/i915/soc/intel_dram.c         |  30 +-
>>>>    drivers/gpu/drm/xe/Makefile                   |   8 +-
>>>>    .../gpu/drm/xe/compat-i915-headers/i915_drv.h |   2 +-
>>>>    .../soc}/intel_dram.h                         |   1 -
>>>>    .../soc}/intel_pch.h                          |   0
>>>>    drivers/gpu/drm/xe/display/ext/intel_dram.c   | 495 ------------------
>>>>    drivers/gpu/drm/xe/display/ext/intel_pch.c    | 157 ------
>>>>    drivers/gpu/drm/xe/xe_device_types.h          |   2 +-
>>>>    drivers/gpu/drm/xe/xe_display.c               |   2 +-
>>>>    9 files changed, 28 insertions(+), 669 deletions(-)
>>>>    rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_dram.h (80%)
>>>>    rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_pch.h (100%)
>>>>    delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_dram.c
>>>>    delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_pch.c
>>>>
>>>> diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
>>>> index 9649051ed8ed..13fe18a9372f 100644
>>>> --- a/drivers/gpu/drm/i915/soc/intel_dram.c
>>>> +++ b/drivers/gpu/drm/i915/soc/intel_dram.c
>>>> @@ -9,8 +9,12 @@
>>>>    #include "i915_reg.h"
>>>>    #include "intel_dram.h"
>>>>    #include "intel_mchbar_regs.h"
>>>> -#include "intel_pcode.h"
>>>> +#ifdef I915
>>>>    #include "vlv_sideband.h"
>>>> +#include "display/intel_de.h"
>>>> +#else
>>>> +#include "xe_de.h"
>>>> +#endif
>>>>    
>>>>    struct dram_dimm_info {
>>>>    	u16 size;
>>>> @@ -43,6 +47,8 @@ static const char *intel_dram_type_str(enum intel_dram_type type)
>>>>    
>>>>    #undef DRAM_TYPE_STR
>>>>    
>>>> +#ifdef I915
>>>> +
>>>>    static void pnv_detect_mem_freq(struct drm_i915_private *dev_priv)
>>>>    {
>>>>    	u32 tmp;
>>>> @@ -191,6 +197,9 @@ static void detect_mem_freq(struct drm_i915_private *i915)
>>>>    	if (i915->mem_freq)
>>>>    		drm_dbg(&i915->drm, "DDR speed: %d MHz\n", i915->mem_freq);
>>>>    }
>>>> +#else
>>>> +#define detect_mem_freq(i915) do { } while (0)
>>>> +#endif
>>>>    
>>>>    static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
>>>>    {
>>>> @@ -339,14 +348,12 @@ skl_dram_get_channels_info(struct drm_i915_private *i915)
>>>>    	u32 val;
>>>>    	int ret;
>>>>    
>>>> -	val = intel_uncore_read(&i915->uncore,
>>>> -				SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
>>>> +	val = intel_de_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
>>> This isn't great, because those aren't DE registers. So this won't fly
>>> upstream.
>> Could we do this for xe instead?
>>
>> #define intel_uncore_read(ignore, reg) xe_mmio_read32(i915, reg) ?
> The obvious downside is that it relies on the implict variable i915,
> which could also be named something else, and it has taken a long, long
> time trying to fix those in i915.
>
> BR,
> Jani.

Well, here goes...

Everything can be solved by adding another layer of indirection. O:)

--->8--------

Instead of making a copy that is hard to keep up to date, build the
source directly.

Add a rule for soc_* files, and use those to build i915.

Changes since v1:
- Add a intel_dram_read/intel_dram_pcode_read at the top of the file.

Signed-off-by: Maarten Lankhorst<maarten.lankhorst@linux.intel.com>
---
  drivers/gpu/drm/i915/soc/intel_dram.c         |  51 +-
  drivers/gpu/drm/xe/Makefile                   |   8 +-
  .../gpu/drm/xe/compat-i915-headers/i915_drv.h |   2 +-
  .../soc}/intel_dram.h                         |   1 -
  .../soc}/intel_pch.h                          |   0
  drivers/gpu/drm/xe/display/ext/intel_dram.c   | 495 ------------------
  drivers/gpu/drm/xe/display/ext/intel_pch.c    | 157 ------
  drivers/gpu/drm/xe/xe_device_types.h          |   2 +-
  drivers/gpu/drm/xe/xe_display.c               |   2 +-
  9 files changed, 50 insertions(+), 668 deletions(-)
  rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_dram.h (80%)
  rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_pch.h (100%)
  delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_dram.c
  delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_pch.c

diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
index 9649051ed8ed..5b2979bd8f09 100644
--- a/drivers/gpu/drm/i915/soc/intel_dram.c
+++ b/drivers/gpu/drm/i915/soc/intel_dram.c
@@ -9,8 +9,35 @@
  #include "i915_reg.h"
  #include "intel_dram.h"
  #include "intel_mchbar_regs.h"
+#ifdef I915
  #include "intel_pcode.h"
  #include "vlv_sideband.h"
+#else
+#include "xe_pcode.h"
+#include "xe_mmio.h"
+#include "xe_gt.h"
+#endif
+
+static inline u32 intel_dram_read(struct drm_i915_private *i915,
+				  const i915_reg_t reg)
+{
+#ifdef I915
+	return intel_uncore_read(&i915->uncore, reg);
+#else
+	return xe_mmio_read32(to_gt(i915), reg.reg);
+#endif
+}
+
+static inline u32 intel_dram_pcode_read(struct drm_i915_private *i915,
+					u32 mbox, u32 *val)
+{
+#ifdef I915
+	return snb_pcode_read(&i915->uncore, mbox, val, NULL);
+#else
+	return xe_pcode_read(to_gt(i915), mbox, val, NULL);
+#endif
+}
+
  
  struct dram_dimm_info {
  	u16 size;
@@ -43,6 +70,8 @@ static const char *intel_dram_type_str(enum intel_dram_type type)
  
  #undef DRAM_TYPE_STR
  
+#ifdef I915
+
  static void pnv_detect_mem_freq(struct drm_i915_private *dev_priv)
  {
  	u32 tmp;
@@ -191,6 +220,9 @@ static void detect_mem_freq(struct drm_i915_private *i915)
  	if (i915->mem_freq)
  		drm_dbg(&i915->drm, "DDR speed: %d MHz\n", i915->mem_freq);
  }
+#else
+#define detect_mem_freq(i915) do { } while (0)
+#endif
  
  static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
  {
@@ -339,14 +371,12 @@ skl_dram_get_channels_info(struct drm_i915_private *i915)
  	u32 val;
  	int ret;
  
-	val = intel_uncore_read(&i915->uncore,
-				SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
+	val = intel_dram_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
  	ret = skl_dram_get_channel_info(i915, &ch0, 0, val);
  	if (ret == 0)
  		dram_info->num_channels++;
  
-	val = intel_uncore_read(&i915->uncore,
-				SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
+	val = intel_dram_read(i915, SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
  	ret = skl_dram_get_channel_info(i915, &ch1, 1, val);
  	if (ret == 0)
  		dram_info->num_channels++;
@@ -376,8 +406,7 @@ skl_get_dram_type(struct drm_i915_private *i915)
  {
  	u32 val;
  
-	val = intel_uncore_read(&i915->uncore,
-				SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
+	val = intel_dram_read(i915, SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
  
  	switch (val & SKL_DRAM_DDR_TYPE_MASK) {
  	case SKL_DRAM_DDR_TYPE_DDR3:
@@ -503,7 +532,7 @@ static int bxt_get_dram_info(struct drm_i915_private *i915)
  		struct dram_dimm_info dimm;
  		enum intel_dram_type type;
  
-		val = intel_uncore_read(&i915->uncore, BXT_D_CR_DRP0_DUNIT(i));
+		val = intel_dram_read(i915, BXT_D_CR_DRP0_DUNIT(i));
  		if (val == 0xFFFFFFFF)
  			continue;
  
@@ -543,8 +572,8 @@ static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
  	u32 val = 0;
  	int ret;
  
-	ret = snb_pcode_read(&dev_priv->uncore, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
-			     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
+	ret = intel_dram_pcode_read(dev_priv, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
+				    ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val);
  	if (ret)
  		return ret;
  
@@ -618,7 +647,7 @@ static int gen12_get_dram_info(struct drm_i915_private *i915)
  
  static int xelpdp_get_dram_info(struct drm_i915_private *i915)
  {
-	u32 val = intel_uncore_read(&i915->uncore, MTL_MEM_SS_INFO_GLOBAL);
+	u32 val = intel_dram_read(i915, MTL_MEM_SS_INFO_GLOBAL);
  	struct dram_info *dram_info = &i915->dram_info;
  
  	switch (REG_FIELD_GET(MTL_DDR_TYPE_MASK, val)) {
@@ -687,6 +716,7 @@ void intel_dram_detect(struct drm_i915_private *i915)
  		    str_yes_no(dram_info->wm_lv_0_adjust_needed));
  }
  
+#ifdef I915
  static u32 gen9_edram_size_mb(struct drm_i915_private *i915, u32 cap)
  {
  	static const u8 ways[8] = { 4, 8, 12, 16, 16, 16, 16, 16 };
@@ -722,3 +752,4 @@ void intel_dram_edram_detect(struct drm_i915_private *i915)
  
  	drm_info(&i915->drm, "Found %uMB of eDRAM\n", i915->edram_size_mb);
  }
+#endif
diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
index 669c7b6f552c..3ebef784445e 100644
--- a/drivers/gpu/drm/xe/Makefile
+++ b/drivers/gpu/drm/xe/Makefile
@@ -124,6 +124,10 @@ $(obj)/display/intel_%.o: $(srctree)/drivers/gpu/drm/i915/display/intel_%.c FORC
  	$(call cmd,force_checksrc)
  	$(call if_changed_rule,cc_o_c)
  
+$(obj)/display/soc_%.o: $(srctree)/drivers/gpu/drm/i915/soc/%.c FORCE
+	$(call cmd,force_checksrc)
+	$(call if_changed_rule,cc_o_c)
+
  # Display..
  xe-$(CONFIG_DRM_XE_DISPLAY) += \
  	xe_display.o \
@@ -204,8 +208,8 @@ xe-$(CONFIG_DRM_XE_DISPLAY) += \
  	display/ext/i915_irq.o \
  	display/ext/i9xx_wm.o \
  	display/ext/intel_device_info.o \
-	display/ext/intel_dram.o \
-	display/ext/intel_pch.o \
+	display/soc_intel_dram.o \
+	display/soc_intel_pch.o \
  	display/ext/intel_pm.o
  
  ifeq ($(CONFIG_ACPI),y)
diff --git a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
index 92dfbb8b3ca4..fa8c53e91bce 100644
--- a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
+++ b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
@@ -15,7 +15,7 @@
  #include "xe_pm.h"
  #include "xe_step.h"
  #include "i915_reg_defs.h"
-#include "intel_pch.h"
+#include "soc/intel_pch.h"
  #include "i915_utils.h"
  #include <linux/pm_runtime.h>
  
diff --git a/drivers/gpu/drm/xe/display/ext/intel_dram.h b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
similarity index 80%
rename from drivers/gpu/drm/xe/display/ext/intel_dram.h
rename to drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
index 4ba13c13162c..7ed5b6e1f805 100644
--- a/drivers/gpu/drm/xe/display/ext/intel_dram.h
+++ b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
@@ -8,7 +8,6 @@
  
  struct drm_i915_private;
  
-void intel_dram_edram_detect(struct drm_i915_private *i915);
  void intel_dram_detect(struct drm_i915_private *i915);
  
  #endif /* __INTEL_DRAM_H__ */
diff --git a/drivers/gpu/drm/xe/display/ext/intel_pch.h b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h
similarity index 100%
rename from drivers/gpu/drm/xe/display/ext/intel_pch.h
rename to drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h
diff --git a/drivers/gpu/drm/xe/display/ext/intel_dram.c b/drivers/gpu/drm/xe/display/ext/intel_dram.c
deleted file mode 100644
index a2df0c502c34..000000000000
--- a/drivers/gpu/drm/xe/display/ext/intel_dram.c
+++ /dev/null
@@ -1,495 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright © 2020 Intel Corporation
- */
-
-#include <linux/string_helpers.h>
-
-#include "i915_drv.h"
-#include "i915_reg.h"
-#include "intel_de.h"
-#include "intel_dram.h"
-#include "intel_mchbar_regs.h"
-
-struct dram_dimm_info {
-	u16 size;
-	u8 width, ranks;
-};
-
-struct dram_channel_info {
-	struct dram_dimm_info dimm_l, dimm_s;
-	u8 ranks;
-	bool is_16gb_dimm;
-};
-
-#define DRAM_TYPE_STR(type) [INTEL_DRAM_ ## type] = #type
-
-static const char *intel_dram_type_str(enum intel_dram_type type)
-{
-	static const char * const str[] = {
-		DRAM_TYPE_STR(UNKNOWN),
-		DRAM_TYPE_STR(DDR3),
-		DRAM_TYPE_STR(DDR4),
-		DRAM_TYPE_STR(LPDDR3),
-		DRAM_TYPE_STR(LPDDR4),
-	};
-
-	if (type >= ARRAY_SIZE(str))
-		type = INTEL_DRAM_UNKNOWN;
-
-	return str[type];
-}
-
-#undef DRAM_TYPE_STR
-
-static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
-{
-	return dimm->ranks * 64 / (dimm->width ?: 1);
-}
-
-/* Returns total Gb for the whole DIMM */
-static int skl_get_dimm_size(u16 val)
-{
-	return (val & SKL_DRAM_SIZE_MASK) * 8;
-}
-
-static int skl_get_dimm_width(u16 val)
-{
-	if (skl_get_dimm_size(val) == 0)
-		return 0;
-
-	switch (val & SKL_DRAM_WIDTH_MASK) {
-	case SKL_DRAM_WIDTH_X8:
-	case SKL_DRAM_WIDTH_X16:
-	case SKL_DRAM_WIDTH_X32:
-		val = (val & SKL_DRAM_WIDTH_MASK) >> SKL_DRAM_WIDTH_SHIFT;
-		return 8 << val;
-	default:
-		MISSING_CASE(val);
-		return 0;
-	}
-}
-
-static int skl_get_dimm_ranks(u16 val)
-{
-	if (skl_get_dimm_size(val) == 0)
-		return 0;
-
-	val = (val & SKL_DRAM_RANK_MASK) >> SKL_DRAM_RANK_SHIFT;
-
-	return val + 1;
-}
-
-/* Returns total Gb for the whole DIMM */
-static int icl_get_dimm_size(u16 val)
-{
-	return (val & ICL_DRAM_SIZE_MASK) * 8 / 2;
-}
-
-static int icl_get_dimm_width(u16 val)
-{
-	if (icl_get_dimm_size(val) == 0)
-		return 0;
-
-	switch (val & ICL_DRAM_WIDTH_MASK) {
-	case ICL_DRAM_WIDTH_X8:
-	case ICL_DRAM_WIDTH_X16:
-	case ICL_DRAM_WIDTH_X32:
-		val = (val & ICL_DRAM_WIDTH_MASK) >> ICL_DRAM_WIDTH_SHIFT;
-		return 8 << val;
-	default:
-		MISSING_CASE(val);
-		return 0;
-	}
-}
-
-static int icl_get_dimm_ranks(u16 val)
-{
-	if (icl_get_dimm_size(val) == 0)
-		return 0;
-
-	val = (val & ICL_DRAM_RANK_MASK) >> ICL_DRAM_RANK_SHIFT;
-
-	return val + 1;
-}
-
-static bool
-skl_is_16gb_dimm(const struct dram_dimm_info *dimm)
-{
-	/* Convert total Gb to Gb per DRAM device */
-	return dimm->size / (intel_dimm_num_devices(dimm) ?: 1) == 16;
-}
-
-static void
-skl_dram_get_dimm_info(struct drm_i915_private *i915,
-		       struct dram_dimm_info *dimm,
-		       int channel, char dimm_name, u16 val)
-{
-	if (GRAPHICS_VER(i915) >= 11) {
-		dimm->size = icl_get_dimm_size(val);
-		dimm->width = icl_get_dimm_width(val);
-		dimm->ranks = icl_get_dimm_ranks(val);
-	} else {
-		dimm->size = skl_get_dimm_size(val);
-		dimm->width = skl_get_dimm_width(val);
-		dimm->ranks = skl_get_dimm_ranks(val);
-	}
-
-	drm_dbg_kms(&i915->drm,
-		    "CH%u DIMM %c size: %u Gb, width: X%u, ranks: %u, 16Gb DIMMs: %s\n",
-		    channel, dimm_name, dimm->size, dimm->width, dimm->ranks,
-		    str_yes_no(skl_is_16gb_dimm(dimm)));
-}
-
-static int
-skl_dram_get_channel_info(struct drm_i915_private *i915,
-			  struct dram_channel_info *ch,
-			  int channel, u32 val)
-{
-	skl_dram_get_dimm_info(i915, &ch->dimm_l,
-			       channel, 'L', val & 0xffff);
-	skl_dram_get_dimm_info(i915, &ch->dimm_s,
-			       channel, 'S', val >> 16);
-
-	if (ch->dimm_l.size == 0 && ch->dimm_s.size == 0) {
-		drm_dbg_kms(&i915->drm, "CH%u not populated\n", channel);
-		return -EINVAL;
-	}
-
-	if (ch->dimm_l.ranks == 2 || ch->dimm_s.ranks == 2)
-		ch->ranks = 2;
-	else if (ch->dimm_l.ranks == 1 && ch->dimm_s.ranks == 1)
-		ch->ranks = 2;
-	else
-		ch->ranks = 1;
-
-	ch->is_16gb_dimm = skl_is_16gb_dimm(&ch->dimm_l) ||
-		skl_is_16gb_dimm(&ch->dimm_s);
-
-	drm_dbg_kms(&i915->drm, "CH%u ranks: %u, 16Gb DIMMs: %s\n",
-		    channel, ch->ranks, str_yes_no(ch->is_16gb_dimm));
-
-	return 0;
-}
-
-static bool
-intel_is_dram_symmetric(const struct dram_channel_info *ch0,
-			const struct dram_channel_info *ch1)
-{
-	return !memcmp(ch0, ch1, sizeof(*ch0)) &&
-		(ch0->dimm_s.size == 0 ||
-		 !memcmp(&ch0->dimm_l, &ch0->dimm_s, sizeof(ch0->dimm_l)));
-}
-
-static int
-skl_dram_get_channels_info(struct drm_i915_private *i915)
-{
-	struct dram_info *dram_info = &i915->dram_info;
-	struct dram_channel_info ch0 = {}, ch1 = {};
-	u32 val;
-	int ret;
-
-	val = intel_de_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
-	ret = skl_dram_get_channel_info(i915, &ch0, 0, val);
-	if (ret == 0)
-		dram_info->num_channels++;
-
-	val = intel_de_read(i915, SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
-	ret = skl_dram_get_channel_info(i915, &ch1, 1, val);
-	if (ret == 0)
-		dram_info->num_channels++;
-
-	if (dram_info->num_channels == 0) {
-		drm_info(&i915->drm, "Number of memory channels is zero\n");
-		return -EINVAL;
-	}
-
-	if (ch0.ranks == 0 && ch1.ranks == 0) {
-		drm_info(&i915->drm, "couldn't get memory rank information\n");
-		return -EINVAL;
-	}
-
-	dram_info->wm_lv_0_adjust_needed = ch0.is_16gb_dimm || ch1.is_16gb_dimm;
-
-	dram_info->symmetric_memory = intel_is_dram_symmetric(&ch0, &ch1);
-
-	drm_dbg_kms(&i915->drm, "Memory configuration is symmetric? %s\n",
-		    str_yes_no(dram_info->symmetric_memory));
-
-	return 0;
-}
-
-static enum intel_dram_type
-skl_get_dram_type(struct drm_i915_private *i915)
-{
-	u32 val;
-
-	val = intel_de_read(i915, SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
-
-	switch (val & SKL_DRAM_DDR_TYPE_MASK) {
-	case SKL_DRAM_DDR_TYPE_DDR3:
-		return INTEL_DRAM_DDR3;
-	case SKL_DRAM_DDR_TYPE_DDR4:
-		return INTEL_DRAM_DDR4;
-	case SKL_DRAM_DDR_TYPE_LPDDR3:
-		return INTEL_DRAM_LPDDR3;
-	case SKL_DRAM_DDR_TYPE_LPDDR4:
-		return INTEL_DRAM_LPDDR4;
-	default:
-		MISSING_CASE(val);
-		return INTEL_DRAM_UNKNOWN;
-	}
-}
-
-static int
-skl_get_dram_info(struct drm_i915_private *i915)
-{
-	struct dram_info *dram_info = &i915->dram_info;
-	int ret;
-
-	dram_info->type = skl_get_dram_type(i915);
-	drm_dbg_kms(&i915->drm, "DRAM type: %s\n",
-		    intel_dram_type_str(dram_info->type));
-
-	ret = skl_dram_get_channels_info(i915);
-	if (ret)
-		return ret;
-
-	return 0;
-}
-
-/* Returns Gb per DRAM device */
-static int bxt_get_dimm_size(u32 val)
-{
-	switch (val & BXT_DRAM_SIZE_MASK) {
-	case BXT_DRAM_SIZE_4GBIT:
-		return 4;
-	case BXT_DRAM_SIZE_6GBIT:
-		return 6;
-	case BXT_DRAM_SIZE_8GBIT:
-		return 8;
-	case BXT_DRAM_SIZE_12GBIT:
-		return 12;
-	case BXT_DRAM_SIZE_16GBIT:
-		return 16;
-	default:
-		MISSING_CASE(val);
-		return 0;
-	}
-}
-
-static int bxt_get_dimm_width(u32 val)
-{
-	if (!bxt_get_dimm_size(val))
-		return 0;
-
-	val = (val & BXT_DRAM_WIDTH_MASK) >> BXT_DRAM_WIDTH_SHIFT;
-
-	return 8 << val;
-}
-
-static int bxt_get_dimm_ranks(u32 val)
-{
-	if (!bxt_get_dimm_size(val))
-		return 0;
-
-	switch (val & BXT_DRAM_RANK_MASK) {
-	case BXT_DRAM_RANK_SINGLE:
-		return 1;
-	case BXT_DRAM_RANK_DUAL:
-		return 2;
-	default:
-		MISSING_CASE(val);
-		return 0;
-	}
-}
-
-static enum intel_dram_type bxt_get_dimm_type(u32 val)
-{
-	if (!bxt_get_dimm_size(val))
-		return INTEL_DRAM_UNKNOWN;
-
-	switch (val & BXT_DRAM_TYPE_MASK) {
-	case BXT_DRAM_TYPE_DDR3:
-		return INTEL_DRAM_DDR3;
-	case BXT_DRAM_TYPE_LPDDR3:
-		return INTEL_DRAM_LPDDR3;
-	case BXT_DRAM_TYPE_DDR4:
-		return INTEL_DRAM_DDR4;
-	case BXT_DRAM_TYPE_LPDDR4:
-		return INTEL_DRAM_LPDDR4;
-	default:
-		MISSING_CASE(val);
-		return INTEL_DRAM_UNKNOWN;
-	}
-}
-
-static void bxt_get_dimm_info(struct dram_dimm_info *dimm, u32 val)
-{
-	dimm->width = bxt_get_dimm_width(val);
-	dimm->ranks = bxt_get_dimm_ranks(val);
-
-	/*
-	 * Size in register is Gb per DRAM device. Convert to total
-	 * Gb to match the way we report this for non-LP platforms.
-	 */
-	dimm->size = bxt_get_dimm_size(val) * intel_dimm_num_devices(dimm);
-}
-
-static int bxt_get_dram_info(struct drm_i915_private *i915)
-{
-	struct dram_info *dram_info = &i915->dram_info;
-	u32 val;
-	u8 valid_ranks = 0;
-	int i;
-
-	/*
-	 * Now read each DUNIT8/9/10/11 to check the rank of each dimms.
-	 */
-	for (i = BXT_D_CR_DRP0_DUNIT_START; i <= BXT_D_CR_DRP0_DUNIT_END; i++) {
-		struct dram_dimm_info dimm;
-		enum intel_dram_type type;
-
-		val = intel_de_read(i915, BXT_D_CR_DRP0_DUNIT(i));
-		if (val == 0xFFFFFFFF)
-			continue;
-
-		dram_info->num_channels++;
-
-		bxt_get_dimm_info(&dimm, val);
-		type = bxt_get_dimm_type(val);
-
-		drm_WARN_ON(&i915->drm, type != INTEL_DRAM_UNKNOWN &&
-			    dram_info->type != INTEL_DRAM_UNKNOWN &&
-			    dram_info->type != type);
-
-		drm_dbg_kms(&i915->drm,
-			    "CH%u DIMM size: %u Gb, width: X%u, ranks: %u, type: %s\n",
-			    i - BXT_D_CR_DRP0_DUNIT_START,
-			    dimm.size, dimm.width, dimm.ranks,
-			    intel_dram_type_str(type));
-
-		if (valid_ranks == 0)
-			valid_ranks = dimm.ranks;
-
-		if (type != INTEL_DRAM_UNKNOWN)
-			dram_info->type = type;
-	}
-
-	if (dram_info->type == INTEL_DRAM_UNKNOWN || valid_ranks == 0) {
-		drm_info(&i915->drm, "couldn't get memory information\n");
-		return -EINVAL;
-	}
-
-	return 0;
-}
-
-static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
-{
-	struct dram_info *dram_info = &dev_priv->dram_info;
-	u32 val = 0;
-	int ret;
-
-	ret = intel_de_pcode_read(dev_priv, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
-			     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
-	if (ret)
-		return ret;
-
-	if (GRAPHICS_VER(dev_priv) >= 12) {
-		switch (val & 0xf) {
-		case 0:
-			dram_info->type = INTEL_DRAM_DDR4;
-			break;
-		case 1:
-			dram_info->type = INTEL_DRAM_DDR5;
-			break;
-		case 2:
-			dram_info->type = INTEL_DRAM_LPDDR5;
-			break;
-		case 3:
-			dram_info->type = INTEL_DRAM_LPDDR4;
-			break;
-		case 4:
-			dram_info->type = INTEL_DRAM_DDR3;
-			break;
-		case 5:
-			dram_info->type = INTEL_DRAM_LPDDR3;
-			break;
-		default:
-			MISSING_CASE(val & 0xf);
-			return -EINVAL;
-		}
-	} else {
-		switch (val & 0xf) {
-		case 0:
-			dram_info->type = INTEL_DRAM_DDR4;
-			break;
-		case 1:
-			dram_info->type = INTEL_DRAM_DDR3;
-			break;
-		case 2:
-			dram_info->type = INTEL_DRAM_LPDDR3;
-			break;
-		case 3:
-			dram_info->type = INTEL_DRAM_LPDDR4;
-			break;
-		default:
-			MISSING_CASE(val & 0xf);
-			return -EINVAL;
-		}
-	}
-
-	dram_info->num_channels = (val & 0xf0) >> 4;
-	dram_info->num_qgv_points = (val & 0xf00) >> 8;
-	dram_info->num_psf_gv_points = (val & 0x3000) >> 12;
-
-	return 0;
-}
-
-static int gen11_get_dram_info(struct drm_i915_private *i915)
-{
-	int ret = skl_get_dram_info(i915);
-
-	if (ret)
-		return ret;
-
-	return icl_pcode_read_mem_global_info(i915);
-}
-
-static int gen12_get_dram_info(struct drm_i915_private *i915)
-{
-	i915->dram_info.wm_lv_0_adjust_needed = false;
-
-	return icl_pcode_read_mem_global_info(i915);
-}
-
-void intel_dram_detect(struct drm_i915_private *i915)
-{
-	struct dram_info *dram_info = &i915->dram_info;
-	int ret;
-
-	if (GRAPHICS_VER(i915) < 9 || IS_DG2(i915) || !HAS_DISPLAY(i915))
-		return;
-
-	/*
-	 * Assume level 0 watermark latency adjustment is needed until proven
-	 * otherwise, this w/a is not needed by bxt/glk.
-	 */
-	dram_info->wm_lv_0_adjust_needed = !IS_GEN9_LP(i915);
-
-	if (GRAPHICS_VER(i915) >= 12)
-		ret = gen12_get_dram_info(i915);
-	else if (GRAPHICS_VER(i915) >= 11)
-		ret = gen11_get_dram_info(i915);
-	else if (IS_GEN9_LP(i915))
-		ret = bxt_get_dram_info(i915);
-	else
-		ret = skl_get_dram_info(i915);
-	if (ret)
-		return;
-
-	drm_dbg_kms(&i915->drm, "DRAM channels: %u\n", dram_info->num_channels);
-
-	drm_dbg_kms(&i915->drm, "Watermark level 0 adjustment needed: %s\n",
-		    str_yes_no(dram_info->wm_lv_0_adjust_needed));
-}
diff --git a/drivers/gpu/drm/xe/display/ext/intel_pch.c b/drivers/gpu/drm/xe/display/ext/intel_pch.c
deleted file mode 100644
index dc2b15b5c4be..000000000000
--- a/drivers/gpu/drm/xe/display/ext/intel_pch.c
+++ /dev/null
@@ -1,157 +0,0 @@
-// SPDX-License-Identifier: MIT
-/*
- * Copyright 2019 Intel Corporation.
- */
-
-#include "i915_drv.h"
-#include "i915_utils.h"
-#include "intel_pch.h"
-
-/* Map PCH device id to PCH type, or PCH_NONE if unknown. */
-static enum intel_pch
-intel_pch_type(const struct drm_i915_private *dev_priv, unsigned short id)
-{
-	switch (id) {
-	case INTEL_PCH_MCC_DEVICE_ID_TYPE:
-		drm_dbg_kms(&dev_priv->drm, "Found Mule Creek Canyon PCH\n");
-		drm_WARN_ON(&dev_priv->drm, !IS_JSL_EHL(dev_priv));
-		/* MCC is TGP compatible */
-		return PCH_TGP;
-	case INTEL_PCH_TGP_DEVICE_ID_TYPE:
-	case INTEL_PCH_TGP2_DEVICE_ID_TYPE:
-		drm_dbg_kms(&dev_priv->drm, "Found Tiger Lake LP PCH\n");
-		drm_WARN_ON(&dev_priv->drm, !IS_TIGERLAKE(dev_priv) &&
-			    !IS_ROCKETLAKE(dev_priv) &&
-			    !IS_GEN9_BC(dev_priv));
-		return PCH_TGP;
-	case INTEL_PCH_ADP_DEVICE_ID_TYPE:
-	case INTEL_PCH_ADP2_DEVICE_ID_TYPE:
-	case INTEL_PCH_ADP3_DEVICE_ID_TYPE:
-	case INTEL_PCH_ADP4_DEVICE_ID_TYPE:
-		drm_dbg_kms(&dev_priv->drm, "Found Alder Lake PCH\n");
-		drm_WARN_ON(&dev_priv->drm, !IS_ALDERLAKE_S(dev_priv) &&
-			    !IS_ALDERLAKE_P(dev_priv));
-		return PCH_ADP;
-	case INTEL_PCH_MTP_DEVICE_ID_TYPE:
-	case INTEL_PCH_MTP2_DEVICE_ID_TYPE:
-		drm_dbg_kms(&dev_priv->drm, "Found Meteor Lake PCH\n");
-		drm_WARN_ON(&dev_priv->drm, !IS_METEORLAKE(dev_priv));
-		return PCH_MTP;
-	default:
-		return PCH_NONE;
-	}
-}
-
-static bool intel_is_virt_pch(unsigned short id,
-			      unsigned short svendor, unsigned short sdevice)
-{
-	return (id == INTEL_PCH_P2X_DEVICE_ID_TYPE ||
-		id == INTEL_PCH_P3X_DEVICE_ID_TYPE ||
-		(id == INTEL_PCH_QEMU_DEVICE_ID_TYPE &&
-		 svendor == PCI_SUBVENDOR_ID_REDHAT_QUMRANET &&
-		 sdevice == PCI_SUBDEVICE_ID_QEMU));
-}
-
-static void
-intel_virt_detect_pch(const struct drm_i915_private *dev_priv,
-		      unsigned short *pch_id, enum intel_pch *pch_type)
-{
-	unsigned short id = 0;
-
-	/*
-	 * In a virtualized passthrough environment we can be in a
-	 * setup where the ISA bridge is not able to be passed through.
-	 * In this case, a south bridge can be emulated and we have to
-	 * make an educated guess as to which PCH is really there.
-	 */
-
-	if (IS_METEORLAKE(dev_priv))
-		id = INTEL_PCH_MTP_DEVICE_ID_TYPE;
-	else if (IS_ALDERLAKE_S(dev_priv) || IS_ALDERLAKE_P(dev_priv))
-		id = INTEL_PCH_ADP_DEVICE_ID_TYPE;
-	else if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv))
-		id = INTEL_PCH_TGP_DEVICE_ID_TYPE;
-	else BUG_ON(1);
-
-	if (id)
-		drm_dbg_kms(&dev_priv->drm, "Assuming PCH ID %04x\n", id);
-	else
-		drm_dbg_kms(&dev_priv->drm, "Assuming no PCH\n");
-
-	*pch_type = intel_pch_type(dev_priv, id);
-
-	/* Sanity check virtual PCH id */
-	if (drm_WARN_ON(&dev_priv->drm,
-			id && *pch_type == PCH_NONE))
-		id = 0;
-
-	*pch_id = id;
-}
-
-void intel_detect_pch(struct drm_i915_private *dev_priv)
-{
-	struct pci_dev *pch = NULL;
-	unsigned short id;
-	enum intel_pch pch_type;
-
-	/* DG1 has south engine display on the same PCI device */
-	if (IS_DG1(dev_priv)) {
-		dev_priv->pch_type = PCH_DG1;
-		return;
-	} else if (IS_DG2(dev_priv)) {
-		dev_priv->pch_type = PCH_DG2;
-		return;
-	}
-
-	/*
-	 * The reason to probe ISA bridge instead of Dev31:Fun0 is to
-	 * make graphics device passthrough work easy for VMM, that only
-	 * need to expose ISA bridge to let driver know the real hardware
-	 * underneath. This is a requirement from virtualization team.
-	 *
-	 * In some virtualized environments (e.g. XEN), there is irrelevant
-	 * ISA bridge in the system. To work reliably, we should scan trhough
-	 * all the ISA bridge devices and check for the first match, instead
-	 * of only checking the first one.
-	 */
-	while ((pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, pch))) {
-		if (pch->vendor != PCI_VENDOR_ID_INTEL)
-			continue;
-
-		id = pch->device & INTEL_PCH_DEVICE_ID_MASK;
-
-		pch_type = intel_pch_type(dev_priv, id);
-		if (pch_type != PCH_NONE) {
-			dev_priv->pch_type = pch_type;
-			dev_priv->pch_id = id;
-			break;
-		} else if (intel_is_virt_pch(id, pch->subsystem_vendor,
-					     pch->subsystem_device)) {
-			intel_virt_detect_pch(dev_priv, &id, &pch_type);
-			dev_priv->pch_type = pch_type;
-			dev_priv->pch_id = id;
-			break;
-		}
-	}
-
-	/*
-	 * Use PCH_NOP (PCH but no South Display) for PCH platforms without
-	 * display.
-	 */
-	if (pch && !HAS_DISPLAY(dev_priv)) {
-		drm_dbg_kms(&dev_priv->drm,
-			    "Display disabled, reverting to NOP PCH\n");
-		dev_priv->pch_type = PCH_NOP;
-		dev_priv->pch_id = 0;
-	} else if (!pch) {
-		if (i915_run_as_guest() && HAS_DISPLAY(dev_priv)) {
-			intel_virt_detect_pch(dev_priv, &id, &pch_type);
-			dev_priv->pch_type = pch_type;
-			dev_priv->pch_id = id;
-		} else {
-			drm_dbg_kms(&dev_priv->drm, "No PCH found.\n");
-		}
-	}
-
-	pci_dev_put(pch);
-}
diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index 88f863edc41c..54de3229553b 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -18,7 +18,7 @@
  
  #if IS_ENABLED(CONFIG_DRM_XE_DISPLAY)
  #include "ext/intel_device_info.h"
-#include "ext/intel_pch.h"
+#include "soc/intel_pch.h"
  #include "intel_display_core.h"
  #endif
  
diff --git a/drivers/gpu/drm/xe/xe_display.c b/drivers/gpu/drm/xe/xe_display.c
index 64e65b29d8de..35dee29da832 100644
--- a/drivers/gpu/drm/xe/xe_display.c
+++ b/drivers/gpu/drm/xe/xe_display.c
@@ -16,8 +16,8 @@
  #include <drm/xe_drm.h>
  
  #include "ext/i915_irq.h"
-#include "ext/intel_dram.h"
  #include "ext/intel_pm.h"
+#include "soc/intel_dram.h"
  #include "intel_acpi.h"
  #include "intel_audio.h"
  #include "intel_bw.h"
-- 
2.34.1


[-- Attachment #2: Type: text/html, Size: 30373 bytes --]

Re: [Intel-xe] [PATCH v2 5/6] drm/xe: Build soc files directly

[Jani Nikula]

On Mon, 03 Apr 2023, Maarten Lankhorst <maarten.lankhorst@linux.intel.com> wrote:
> On 2023-04-03 10:12, Jani Nikula wrote:
>> On Fri, 31 Mar 2023, Maarten Lankhorst<maarten.lankhorst@linux.intel.com>  wrote:
>>> On 2023-03-31 13:49, Jani Nikula wrote:
>>>> On Fri, 31 Mar 2023, Maarten Lankhorst<maarten.lankhorst@linux.intel.com>  wrote:
>>>>> Instead of making a copy that is hard to keep up to date, build the
>>>>> source directly.
>>>>>
>>>>> Add a rule for soc_* files, and use those to build i915.
>>>>>
>>>>> Signed-off-by: Maarten Lankhorst<maarten.lankhorst@linux.intel.com>
>>>>> ---
>>>>>    drivers/gpu/drm/i915/soc/intel_dram.c         |  30 +-
>>>>>    drivers/gpu/drm/xe/Makefile                   |   8 +-
>>>>>    .../gpu/drm/xe/compat-i915-headers/i915_drv.h |   2 +-
>>>>>    .../soc}/intel_dram.h                         |   1 -
>>>>>    .../soc}/intel_pch.h                          |   0
>>>>>    drivers/gpu/drm/xe/display/ext/intel_dram.c   | 495 ------------------
>>>>>    drivers/gpu/drm/xe/display/ext/intel_pch.c    | 157 ------
>>>>>    drivers/gpu/drm/xe/xe_device_types.h          |   2 +-
>>>>>    drivers/gpu/drm/xe/xe_display.c               |   2 +-
>>>>>    9 files changed, 28 insertions(+), 669 deletions(-)
>>>>>    rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_dram.h (80%)
>>>>>    rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_pch.h (100%)
>>>>>    delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_dram.c
>>>>>    delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_pch.c
>>>>>
>>>>> diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
>>>>> index 9649051ed8ed..13fe18a9372f 100644
>>>>> --- a/drivers/gpu/drm/i915/soc/intel_dram.c
>>>>> +++ b/drivers/gpu/drm/i915/soc/intel_dram.c
>>>>> @@ -9,8 +9,12 @@
>>>>>    #include "i915_reg.h"
>>>>>    #include "intel_dram.h"
>>>>>    #include "intel_mchbar_regs.h"
>>>>> -#include "intel_pcode.h"
>>>>> +#ifdef I915
>>>>>    #include "vlv_sideband.h"
>>>>> +#include "display/intel_de.h"
>>>>> +#else
>>>>> +#include "xe_de.h"
>>>>> +#endif
>>>>>    
>>>>>    struct dram_dimm_info {
>>>>>    	u16 size;
>>>>> @@ -43,6 +47,8 @@ static const char *intel_dram_type_str(enum intel_dram_type type)
>>>>>    
>>>>>    #undef DRAM_TYPE_STR
>>>>>    
>>>>> +#ifdef I915
>>>>> +
>>>>>    static void pnv_detect_mem_freq(struct drm_i915_private *dev_priv)
>>>>>    {
>>>>>    	u32 tmp;
>>>>> @@ -191,6 +197,9 @@ static void detect_mem_freq(struct drm_i915_private *i915)
>>>>>    	if (i915->mem_freq)
>>>>>    		drm_dbg(&i915->drm, "DDR speed: %d MHz\n", i915->mem_freq);
>>>>>    }
>>>>> +#else
>>>>> +#define detect_mem_freq(i915) do { } while (0)
>>>>> +#endif
>>>>>    
>>>>>    static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
>>>>>    {
>>>>> @@ -339,14 +348,12 @@ skl_dram_get_channels_info(struct drm_i915_private *i915)
>>>>>    	u32 val;
>>>>>    	int ret;
>>>>>    
>>>>> -	val = intel_uncore_read(&i915->uncore,
>>>>> -				SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
>>>>> +	val = intel_de_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
>>>> This isn't great, because those aren't DE registers. So this won't fly
>>>> upstream.
>>> Could we do this for xe instead?
>>>
>>> #define intel_uncore_read(ignore, reg) xe_mmio_read32(i915, reg) ?
>> The obvious downside is that it relies on the implict variable i915,
>> which could also be named something else, and it has taken a long, long
>> time trying to fix those in i915.
>>
>> BR,
>> Jani.
>
> Well, here goes...
>
> Everything can be solved by adding another layer of indirection. O:)

I'm toying with adding the abstraction layer at intel_uncore level.

xe_device_types.h:

struct xe_device {
	struct fake_uncore {
		spinlock_t lock;
	} uncore;
}

compat-i915-headers/intel_uncore.h:

static inline struct xe_gt *__fake_uncore_to_gt(struct fake_uncore *uncore)
{
	struct xe_device *xe = container_of(uncore, struct xe_device, uncore);

	return to_gt(xe);	
}

static inline u32 intel_uncore_read(struct fake_uncore *uncore, i915_reg_t reg)
{
	return xe_mmio_read32(__fake_uncore_to_gt(uncore), reg.reg);
}

etc.

It intentionally doesn't work for

	struct intel_uncore *uncore = &i915->uncore;

But that doesn't happen in display code. (Could also name it struct
intel_uncore, but the different naming would err on the safe side.)

The end result, we could ditch xe_de.h, and all the intel_uncore stuff,
including intel_de.h, would work.


BR,
Jani.


>
> --->8--------
>
> Instead of making a copy that is hard to keep up to date, build the
> source directly.
>
> Add a rule for soc_* files, and use those to build i915.
>
> Changes since v1:
> - Add a intel_dram_read/intel_dram_pcode_read at the top of the file.
>
> Signed-off-by: Maarten Lankhorst<maarten.lankhorst@linux.intel.com>
> ---
>   drivers/gpu/drm/i915/soc/intel_dram.c         |  51 +-
>   drivers/gpu/drm/xe/Makefile                   |   8 +-
>   .../gpu/drm/xe/compat-i915-headers/i915_drv.h |   2 +-
>   .../soc}/intel_dram.h                         |   1 -
>   .../soc}/intel_pch.h                          |   0
>   drivers/gpu/drm/xe/display/ext/intel_dram.c   | 495 ------------------
>   drivers/gpu/drm/xe/display/ext/intel_pch.c    | 157 ------
>   drivers/gpu/drm/xe/xe_device_types.h          |   2 +-
>   drivers/gpu/drm/xe/xe_display.c               |   2 +-
>   9 files changed, 50 insertions(+), 668 deletions(-)
>   rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_dram.h (80%)
>   rename drivers/gpu/drm/xe/{display/ext => compat-i915-headers/soc}/intel_pch.h (100%)
>   delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_dram.c
>   delete mode 100644 drivers/gpu/drm/xe/display/ext/intel_pch.c
>
> diff --git a/drivers/gpu/drm/i915/soc/intel_dram.c b/drivers/gpu/drm/i915/soc/intel_dram.c
> index 9649051ed8ed..5b2979bd8f09 100644
> --- a/drivers/gpu/drm/i915/soc/intel_dram.c
> +++ b/drivers/gpu/drm/i915/soc/intel_dram.c
> @@ -9,8 +9,35 @@
>   #include "i915_reg.h"
>   #include "intel_dram.h"
>   #include "intel_mchbar_regs.h"
> +#ifdef I915
>   #include "intel_pcode.h"
>   #include "vlv_sideband.h"
> +#else
> +#include "xe_pcode.h"
> +#include "xe_mmio.h"
> +#include "xe_gt.h"
> +#endif
> +
> +static inline u32 intel_dram_read(struct drm_i915_private *i915,
> +				  const i915_reg_t reg)
> +{
> +#ifdef I915
> +	return intel_uncore_read(&i915->uncore, reg);
> +#else
> +	return xe_mmio_read32(to_gt(i915), reg.reg);
> +#endif
> +}
> +
> +static inline u32 intel_dram_pcode_read(struct drm_i915_private *i915,
> +					u32 mbox, u32 *val)
> +{
> +#ifdef I915
> +	return snb_pcode_read(&i915->uncore, mbox, val, NULL);
> +#else
> +	return xe_pcode_read(to_gt(i915), mbox, val, NULL);
> +#endif
> +}
> +
>   
>   struct dram_dimm_info {
>   	u16 size;
> @@ -43,6 +70,8 @@ static const char *intel_dram_type_str(enum intel_dram_type type)
>   
>   #undef DRAM_TYPE_STR
>   
> +#ifdef I915
> +
>   static void pnv_detect_mem_freq(struct drm_i915_private *dev_priv)
>   {
>   	u32 tmp;
> @@ -191,6 +220,9 @@ static void detect_mem_freq(struct drm_i915_private *i915)
>   	if (i915->mem_freq)
>   		drm_dbg(&i915->drm, "DDR speed: %d MHz\n", i915->mem_freq);
>   }
> +#else
> +#define detect_mem_freq(i915) do { } while (0)
> +#endif
>   
>   static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
>   {
> @@ -339,14 +371,12 @@ skl_dram_get_channels_info(struct drm_i915_private *i915)
>   	u32 val;
>   	int ret;
>   
> -	val = intel_uncore_read(&i915->uncore,
> -				SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
> +	val = intel_dram_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
>   	ret = skl_dram_get_channel_info(i915, &ch0, 0, val);
>   	if (ret == 0)
>   		dram_info->num_channels++;
>   
> -	val = intel_uncore_read(&i915->uncore,
> -				SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
> +	val = intel_dram_read(i915, SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
>   	ret = skl_dram_get_channel_info(i915, &ch1, 1, val);
>   	if (ret == 0)
>   		dram_info->num_channels++;
> @@ -376,8 +406,7 @@ skl_get_dram_type(struct drm_i915_private *i915)
>   {
>   	u32 val;
>   
> -	val = intel_uncore_read(&i915->uncore,
> -				SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
> +	val = intel_dram_read(i915, SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
>   
>   	switch (val & SKL_DRAM_DDR_TYPE_MASK) {
>   	case SKL_DRAM_DDR_TYPE_DDR3:
> @@ -503,7 +532,7 @@ static int bxt_get_dram_info(struct drm_i915_private *i915)
>   		struct dram_dimm_info dimm;
>   		enum intel_dram_type type;
>   
> -		val = intel_uncore_read(&i915->uncore, BXT_D_CR_DRP0_DUNIT(i));
> +		val = intel_dram_read(i915, BXT_D_CR_DRP0_DUNIT(i));
>   		if (val == 0xFFFFFFFF)
>   			continue;
>   
> @@ -543,8 +572,8 @@ static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
>   	u32 val = 0;
>   	int ret;
>   
> -	ret = snb_pcode_read(&dev_priv->uncore, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
> -			     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
> +	ret = intel_dram_pcode_read(dev_priv, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
> +				    ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val);
>   	if (ret)
>   		return ret;
>   
> @@ -618,7 +647,7 @@ static int gen12_get_dram_info(struct drm_i915_private *i915)
>   
>   static int xelpdp_get_dram_info(struct drm_i915_private *i915)
>   {
> -	u32 val = intel_uncore_read(&i915->uncore, MTL_MEM_SS_INFO_GLOBAL);
> +	u32 val = intel_dram_read(i915, MTL_MEM_SS_INFO_GLOBAL);
>   	struct dram_info *dram_info = &i915->dram_info;
>   
>   	switch (REG_FIELD_GET(MTL_DDR_TYPE_MASK, val)) {
> @@ -687,6 +716,7 @@ void intel_dram_detect(struct drm_i915_private *i915)
>   		    str_yes_no(dram_info->wm_lv_0_adjust_needed));
>   }
>   
> +#ifdef I915
>   static u32 gen9_edram_size_mb(struct drm_i915_private *i915, u32 cap)
>   {
>   	static const u8 ways[8] = { 4, 8, 12, 16, 16, 16, 16, 16 };
> @@ -722,3 +752,4 @@ void intel_dram_edram_detect(struct drm_i915_private *i915)
>   
>   	drm_info(&i915->drm, "Found %uMB of eDRAM\n", i915->edram_size_mb);
>   }
> +#endif
> diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
> index 669c7b6f552c..3ebef784445e 100644
> --- a/drivers/gpu/drm/xe/Makefile
> +++ b/drivers/gpu/drm/xe/Makefile
> @@ -124,6 +124,10 @@ $(obj)/display/intel_%.o: $(srctree)/drivers/gpu/drm/i915/display/intel_%.c FORC
>   	$(call cmd,force_checksrc)
>   	$(call if_changed_rule,cc_o_c)
>   
> +$(obj)/display/soc_%.o: $(srctree)/drivers/gpu/drm/i915/soc/%.c FORCE
> +	$(call cmd,force_checksrc)
> +	$(call if_changed_rule,cc_o_c)
> +
>   # Display..
>   xe-$(CONFIG_DRM_XE_DISPLAY) += \
>   	xe_display.o \
> @@ -204,8 +208,8 @@ xe-$(CONFIG_DRM_XE_DISPLAY) += \
>   	display/ext/i915_irq.o \
>   	display/ext/i9xx_wm.o \
>   	display/ext/intel_device_info.o \
> -	display/ext/intel_dram.o \
> -	display/ext/intel_pch.o \
> +	display/soc_intel_dram.o \
> +	display/soc_intel_pch.o \
>   	display/ext/intel_pm.o
>   
>   ifeq ($(CONFIG_ACPI),y)
> diff --git a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> index 92dfbb8b3ca4..fa8c53e91bce 100644
> --- a/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> +++ b/drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h
> @@ -15,7 +15,7 @@
>   #include "xe_pm.h"
>   #include "xe_step.h"
>   #include "i915_reg_defs.h"
> -#include "intel_pch.h"
> +#include "soc/intel_pch.h"
>   #include "i915_utils.h"
>   #include <linux/pm_runtime.h>
>   
> diff --git a/drivers/gpu/drm/xe/display/ext/intel_dram.h b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
> similarity index 80%
> rename from drivers/gpu/drm/xe/display/ext/intel_dram.h
> rename to drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
> index 4ba13c13162c..7ed5b6e1f805 100644
> --- a/drivers/gpu/drm/xe/display/ext/intel_dram.h
> +++ b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_dram.h
> @@ -8,7 +8,6 @@
>   
>   struct drm_i915_private;
>   
> -void intel_dram_edram_detect(struct drm_i915_private *i915);
>   void intel_dram_detect(struct drm_i915_private *i915);
>   
>   #endif /* __INTEL_DRAM_H__ */
> diff --git a/drivers/gpu/drm/xe/display/ext/intel_pch.h b/drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h
> similarity index 100%
> rename from drivers/gpu/drm/xe/display/ext/intel_pch.h
> rename to drivers/gpu/drm/xe/compat-i915-headers/soc/intel_pch.h
> diff --git a/drivers/gpu/drm/xe/display/ext/intel_dram.c b/drivers/gpu/drm/xe/display/ext/intel_dram.c
> deleted file mode 100644
> index a2df0c502c34..000000000000
> --- a/drivers/gpu/drm/xe/display/ext/intel_dram.c
> +++ /dev/null
> @@ -1,495 +0,0 @@
> -// SPDX-License-Identifier: MIT
> -/*
> - * Copyright © 2020 Intel Corporation
> - */
> -
> -#include <linux/string_helpers.h>
> -
> -#include "i915_drv.h"
> -#include "i915_reg.h"
> -#include "intel_de.h"
> -#include "intel_dram.h"
> -#include "intel_mchbar_regs.h"
> -
> -struct dram_dimm_info {
> -	u16 size;
> -	u8 width, ranks;
> -};
> -
> -struct dram_channel_info {
> -	struct dram_dimm_info dimm_l, dimm_s;
> -	u8 ranks;
> -	bool is_16gb_dimm;
> -};
> -
> -#define DRAM_TYPE_STR(type) [INTEL_DRAM_ ## type] = #type
> -
> -static const char *intel_dram_type_str(enum intel_dram_type type)
> -{
> -	static const char * const str[] = {
> -		DRAM_TYPE_STR(UNKNOWN),
> -		DRAM_TYPE_STR(DDR3),
> -		DRAM_TYPE_STR(DDR4),
> -		DRAM_TYPE_STR(LPDDR3),
> -		DRAM_TYPE_STR(LPDDR4),
> -	};
> -
> -	if (type >= ARRAY_SIZE(str))
> -		type = INTEL_DRAM_UNKNOWN;
> -
> -	return str[type];
> -}
> -
> -#undef DRAM_TYPE_STR
> -
> -static int intel_dimm_num_devices(const struct dram_dimm_info *dimm)
> -{
> -	return dimm->ranks * 64 / (dimm->width ?: 1);
> -}
> -
> -/* Returns total Gb for the whole DIMM */
> -static int skl_get_dimm_size(u16 val)
> -{
> -	return (val & SKL_DRAM_SIZE_MASK) * 8;
> -}
> -
> -static int skl_get_dimm_width(u16 val)
> -{
> -	if (skl_get_dimm_size(val) == 0)
> -		return 0;
> -
> -	switch (val & SKL_DRAM_WIDTH_MASK) {
> -	case SKL_DRAM_WIDTH_X8:
> -	case SKL_DRAM_WIDTH_X16:
> -	case SKL_DRAM_WIDTH_X32:
> -		val = (val & SKL_DRAM_WIDTH_MASK) >> SKL_DRAM_WIDTH_SHIFT;
> -		return 8 << val;
> -	default:
> -		MISSING_CASE(val);
> -		return 0;
> -	}
> -}
> -
> -static int skl_get_dimm_ranks(u16 val)
> -{
> -	if (skl_get_dimm_size(val) == 0)
> -		return 0;
> -
> -	val = (val & SKL_DRAM_RANK_MASK) >> SKL_DRAM_RANK_SHIFT;
> -
> -	return val + 1;
> -}
> -
> -/* Returns total Gb for the whole DIMM */
> -static int icl_get_dimm_size(u16 val)
> -{
> -	return (val & ICL_DRAM_SIZE_MASK) * 8 / 2;
> -}
> -
> -static int icl_get_dimm_width(u16 val)
> -{
> -	if (icl_get_dimm_size(val) == 0)
> -		return 0;
> -
> -	switch (val & ICL_DRAM_WIDTH_MASK) {
> -	case ICL_DRAM_WIDTH_X8:
> -	case ICL_DRAM_WIDTH_X16:
> -	case ICL_DRAM_WIDTH_X32:
> -		val = (val & ICL_DRAM_WIDTH_MASK) >> ICL_DRAM_WIDTH_SHIFT;
> -		return 8 << val;
> -	default:
> -		MISSING_CASE(val);
> -		return 0;
> -	}
> -}
> -
> -static int icl_get_dimm_ranks(u16 val)
> -{
> -	if (icl_get_dimm_size(val) == 0)
> -		return 0;
> -
> -	val = (val & ICL_DRAM_RANK_MASK) >> ICL_DRAM_RANK_SHIFT;
> -
> -	return val + 1;
> -}
> -
> -static bool
> -skl_is_16gb_dimm(const struct dram_dimm_info *dimm)
> -{
> -	/* Convert total Gb to Gb per DRAM device */
> -	return dimm->size / (intel_dimm_num_devices(dimm) ?: 1) == 16;
> -}
> -
> -static void
> -skl_dram_get_dimm_info(struct drm_i915_private *i915,
> -		       struct dram_dimm_info *dimm,
> -		       int channel, char dimm_name, u16 val)
> -{
> -	if (GRAPHICS_VER(i915) >= 11) {
> -		dimm->size = icl_get_dimm_size(val);
> -		dimm->width = icl_get_dimm_width(val);
> -		dimm->ranks = icl_get_dimm_ranks(val);
> -	} else {
> -		dimm->size = skl_get_dimm_size(val);
> -		dimm->width = skl_get_dimm_width(val);
> -		dimm->ranks = skl_get_dimm_ranks(val);
> -	}
> -
> -	drm_dbg_kms(&i915->drm,
> -		    "CH%u DIMM %c size: %u Gb, width: X%u, ranks: %u, 16Gb DIMMs: %s\n",
> -		    channel, dimm_name, dimm->size, dimm->width, dimm->ranks,
> -		    str_yes_no(skl_is_16gb_dimm(dimm)));
> -}
> -
> -static int
> -skl_dram_get_channel_info(struct drm_i915_private *i915,
> -			  struct dram_channel_info *ch,
> -			  int channel, u32 val)
> -{
> -	skl_dram_get_dimm_info(i915, &ch->dimm_l,
> -			       channel, 'L', val & 0xffff);
> -	skl_dram_get_dimm_info(i915, &ch->dimm_s,
> -			       channel, 'S', val >> 16);
> -
> -	if (ch->dimm_l.size == 0 && ch->dimm_s.size == 0) {
> -		drm_dbg_kms(&i915->drm, "CH%u not populated\n", channel);
> -		return -EINVAL;
> -	}
> -
> -	if (ch->dimm_l.ranks == 2 || ch->dimm_s.ranks == 2)
> -		ch->ranks = 2;
> -	else if (ch->dimm_l.ranks == 1 && ch->dimm_s.ranks == 1)
> -		ch->ranks = 2;
> -	else
> -		ch->ranks = 1;
> -
> -	ch->is_16gb_dimm = skl_is_16gb_dimm(&ch->dimm_l) ||
> -		skl_is_16gb_dimm(&ch->dimm_s);
> -
> -	drm_dbg_kms(&i915->drm, "CH%u ranks: %u, 16Gb DIMMs: %s\n",
> -		    channel, ch->ranks, str_yes_no(ch->is_16gb_dimm));
> -
> -	return 0;
> -}
> -
> -static bool
> -intel_is_dram_symmetric(const struct dram_channel_info *ch0,
> -			const struct dram_channel_info *ch1)
> -{
> -	return !memcmp(ch0, ch1, sizeof(*ch0)) &&
> -		(ch0->dimm_s.size == 0 ||
> -		 !memcmp(&ch0->dimm_l, &ch0->dimm_s, sizeof(ch0->dimm_l)));
> -}
> -
> -static int
> -skl_dram_get_channels_info(struct drm_i915_private *i915)
> -{
> -	struct dram_info *dram_info = &i915->dram_info;
> -	struct dram_channel_info ch0 = {}, ch1 = {};
> -	u32 val;
> -	int ret;
> -
> -	val = intel_de_read(i915, SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
> -	ret = skl_dram_get_channel_info(i915, &ch0, 0, val);
> -	if (ret == 0)
> -		dram_info->num_channels++;
> -
> -	val = intel_de_read(i915, SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
> -	ret = skl_dram_get_channel_info(i915, &ch1, 1, val);
> -	if (ret == 0)
> -		dram_info->num_channels++;
> -
> -	if (dram_info->num_channels == 0) {
> -		drm_info(&i915->drm, "Number of memory channels is zero\n");
> -		return -EINVAL;
> -	}
> -
> -	if (ch0.ranks == 0 && ch1.ranks == 0) {
> -		drm_info(&i915->drm, "couldn't get memory rank information\n");
> -		return -EINVAL;
> -	}
> -
> -	dram_info->wm_lv_0_adjust_needed = ch0.is_16gb_dimm || ch1.is_16gb_dimm;
> -
> -	dram_info->symmetric_memory = intel_is_dram_symmetric(&ch0, &ch1);
> -
> -	drm_dbg_kms(&i915->drm, "Memory configuration is symmetric? %s\n",
> -		    str_yes_no(dram_info->symmetric_memory));
> -
> -	return 0;
> -}
> -
> -static enum intel_dram_type
> -skl_get_dram_type(struct drm_i915_private *i915)
> -{
> -	u32 val;
> -
> -	val = intel_de_read(i915, SKL_MAD_INTER_CHANNEL_0_0_0_MCHBAR_MCMAIN);
> -
> -	switch (val & SKL_DRAM_DDR_TYPE_MASK) {
> -	case SKL_DRAM_DDR_TYPE_DDR3:
> -		return INTEL_DRAM_DDR3;
> -	case SKL_DRAM_DDR_TYPE_DDR4:
> -		return INTEL_DRAM_DDR4;
> -	case SKL_DRAM_DDR_TYPE_LPDDR3:
> -		return INTEL_DRAM_LPDDR3;
> -	case SKL_DRAM_DDR_TYPE_LPDDR4:
> -		return INTEL_DRAM_LPDDR4;
> -	default:
> -		MISSING_CASE(val);
> -		return INTEL_DRAM_UNKNOWN;
> -	}
> -}
> -
> -static int
> -skl_get_dram_info(struct drm_i915_private *i915)
> -{
> -	struct dram_info *dram_info = &i915->dram_info;
> -	int ret;
> -
> -	dram_info->type = skl_get_dram_type(i915);
> -	drm_dbg_kms(&i915->drm, "DRAM type: %s\n",
> -		    intel_dram_type_str(dram_info->type));
> -
> -	ret = skl_dram_get_channels_info(i915);
> -	if (ret)
> -		return ret;
> -
> -	return 0;
> -}
> -
> -/* Returns Gb per DRAM device */
> -static int bxt_get_dimm_size(u32 val)
> -{
> -	switch (val & BXT_DRAM_SIZE_MASK) {
> -	case BXT_DRAM_SIZE_4GBIT:
> -		return 4;
> -	case BXT_DRAM_SIZE_6GBIT:
> -		return 6;
> -	case BXT_DRAM_SIZE_8GBIT:
> -		return 8;
> -	case BXT_DRAM_SIZE_12GBIT:
> -		return 12;
> -	case BXT_DRAM_SIZE_16GBIT:
> -		return 16;
> -	default:
> -		MISSING_CASE(val);
> -		return 0;
> -	}
> -}
> -
> -static int bxt_get_dimm_width(u32 val)
> -{
> -	if (!bxt_get_dimm_size(val))
> -		return 0;
> -
> -	val = (val & BXT_DRAM_WIDTH_MASK) >> BXT_DRAM_WIDTH_SHIFT;
> -
> -	return 8 << val;
> -}
> -
> -static int bxt_get_dimm_ranks(u32 val)
> -{
> -	if (!bxt_get_dimm_size(val))
> -		return 0;
> -
> -	switch (val & BXT_DRAM_RANK_MASK) {
> -	case BXT_DRAM_RANK_SINGLE:
> -		return 1;
> -	case BXT_DRAM_RANK_DUAL:
> -		return 2;
> -	default:
> -		MISSING_CASE(val);
> -		return 0;
> -	}
> -}
> -
> -static enum intel_dram_type bxt_get_dimm_type(u32 val)
> -{
> -	if (!bxt_get_dimm_size(val))
> -		return INTEL_DRAM_UNKNOWN;
> -
> -	switch (val & BXT_DRAM_TYPE_MASK) {
> -	case BXT_DRAM_TYPE_DDR3:
> -		return INTEL_DRAM_DDR3;
> -	case BXT_DRAM_TYPE_LPDDR3:
> -		return INTEL_DRAM_LPDDR3;
> -	case BXT_DRAM_TYPE_DDR4:
> -		return INTEL_DRAM_DDR4;
> -	case BXT_DRAM_TYPE_LPDDR4:
> -		return INTEL_DRAM_LPDDR4;
> -	default:
> -		MISSING_CASE(val);
> -		return INTEL_DRAM_UNKNOWN;
> -	}
> -}
> -
> -static void bxt_get_dimm_info(struct dram_dimm_info *dimm, u32 val)
> -{
> -	dimm->width = bxt_get_dimm_width(val);
> -	dimm->ranks = bxt_get_dimm_ranks(val);
> -
> -	/*
> -	 * Size in register is Gb per DRAM device. Convert to total
> -	 * Gb to match the way we report this for non-LP platforms.
> -	 */
> -	dimm->size = bxt_get_dimm_size(val) * intel_dimm_num_devices(dimm);
> -}
> -
> -static int bxt_get_dram_info(struct drm_i915_private *i915)
> -{
> -	struct dram_info *dram_info = &i915->dram_info;
> -	u32 val;
> -	u8 valid_ranks = 0;
> -	int i;
> -
> -	/*
> -	 * Now read each DUNIT8/9/10/11 to check the rank of each dimms.
> -	 */
> -	for (i = BXT_D_CR_DRP0_DUNIT_START; i <= BXT_D_CR_DRP0_DUNIT_END; i++) {
> -		struct dram_dimm_info dimm;
> -		enum intel_dram_type type;
> -
> -		val = intel_de_read(i915, BXT_D_CR_DRP0_DUNIT(i));
> -		if (val == 0xFFFFFFFF)
> -			continue;
> -
> -		dram_info->num_channels++;
> -
> -		bxt_get_dimm_info(&dimm, val);
> -		type = bxt_get_dimm_type(val);
> -
> -		drm_WARN_ON(&i915->drm, type != INTEL_DRAM_UNKNOWN &&
> -			    dram_info->type != INTEL_DRAM_UNKNOWN &&
> -			    dram_info->type != type);
> -
> -		drm_dbg_kms(&i915->drm,
> -			    "CH%u DIMM size: %u Gb, width: X%u, ranks: %u, type: %s\n",
> -			    i - BXT_D_CR_DRP0_DUNIT_START,
> -			    dimm.size, dimm.width, dimm.ranks,
> -			    intel_dram_type_str(type));
> -
> -		if (valid_ranks == 0)
> -			valid_ranks = dimm.ranks;
> -
> -		if (type != INTEL_DRAM_UNKNOWN)
> -			dram_info->type = type;
> -	}
> -
> -	if (dram_info->type == INTEL_DRAM_UNKNOWN || valid_ranks == 0) {
> -		drm_info(&i915->drm, "couldn't get memory information\n");
> -		return -EINVAL;
> -	}
> -
> -	return 0;
> -}
> -
> -static int icl_pcode_read_mem_global_info(struct drm_i915_private *dev_priv)
> -{
> -	struct dram_info *dram_info = &dev_priv->dram_info;
> -	u32 val = 0;
> -	int ret;
> -
> -	ret = intel_de_pcode_read(dev_priv, ICL_PCODE_MEM_SUBSYSYSTEM_INFO |
> -			     ICL_PCODE_MEM_SS_READ_GLOBAL_INFO, &val, NULL);
> -	if (ret)
> -		return ret;
> -
> -	if (GRAPHICS_VER(dev_priv) >= 12) {
> -		switch (val & 0xf) {
> -		case 0:
> -			dram_info->type = INTEL_DRAM_DDR4;
> -			break;
> -		case 1:
> -			dram_info->type = INTEL_DRAM_DDR5;
> -			break;
> -		case 2:
> -			dram_info->type = INTEL_DRAM_LPDDR5;
> -			break;
> -		case 3:
> -			dram_info->type = INTEL_DRAM_LPDDR4;
> -			break;
> -		case 4:
> -			dram_info->type = INTEL_DRAM_DDR3;
> -			break;
> -		case 5:
> -			dram_info->type = INTEL_DRAM_LPDDR3;
> -			break;
> -		default:
> -			MISSING_CASE(val & 0xf);
> -			return -EINVAL;
> -		}
> -	} else {
> -		switch (val & 0xf) {
> -		case 0:
> -			dram_info->type = INTEL_DRAM_DDR4;
> -			break;
> -		case 1:
> -			dram_info->type = INTEL_DRAM_DDR3;
> -			break;
> -		case 2:
> -			dram_info->type = INTEL_DRAM_LPDDR3;
> -			break;
> -		case 3:
> -			dram_info->type = INTEL_DRAM_LPDDR4;
> -			break;
> -		default:
> -			MISSING_CASE(val & 0xf);
> -			return -EINVAL;
> -		}
> -	}
> -
> -	dram_info->num_channels = (val & 0xf0) >> 4;
> -	dram_info->num_qgv_points = (val & 0xf00) >> 8;
> -	dram_info->num_psf_gv_points = (val & 0x3000) >> 12;
> -
> -	return 0;
> -}
> -
> -static int gen11_get_dram_info(struct drm_i915_private *i915)
> -{
> -	int ret = skl_get_dram_info(i915);
> -
> -	if (ret)
> -		return ret;
> -
> -	return icl_pcode_read_mem_global_info(i915);
> -}
> -
> -static int gen12_get_dram_info(struct drm_i915_private *i915)
> -{
> -	i915->dram_info.wm_lv_0_adjust_needed = false;
> -
> -	return icl_pcode_read_mem_global_info(i915);
> -}
> -
> -void intel_dram_detect(struct drm_i915_private *i915)
> -{
> -	struct dram_info *dram_info = &i915->dram_info;
> -	int ret;
> -
> -	if (GRAPHICS_VER(i915) < 9 || IS_DG2(i915) || !HAS_DISPLAY(i915))
> -		return;
> -
> -	/*
> -	 * Assume level 0 watermark latency adjustment is needed until proven
> -	 * otherwise, this w/a is not needed by bxt/glk.
> -	 */
> -	dram_info->wm_lv_0_adjust_needed = !IS_GEN9_LP(i915);
> -
> -	if (GRAPHICS_VER(i915) >= 12)
> -		ret = gen12_get_dram_info(i915);
> -	else if (GRAPHICS_VER(i915) >= 11)
> -		ret = gen11_get_dram_info(i915);
> -	else if (IS_GEN9_LP(i915))
> -		ret = bxt_get_dram_info(i915);
> -	else
> -		ret = skl_get_dram_info(i915);
> -	if (ret)
> -		return;
> -
> -	drm_dbg_kms(&i915->drm, "DRAM channels: %u\n", dram_info->num_channels);
> -
> -	drm_dbg_kms(&i915->drm, "Watermark level 0 adjustment needed: %s\n",
> -		    str_yes_no(dram_info->wm_lv_0_adjust_needed));
> -}
> diff --git a/drivers/gpu/drm/xe/display/ext/intel_pch.c b/drivers/gpu/drm/xe/display/ext/intel_pch.c
> deleted file mode 100644
> index dc2b15b5c4be..000000000000
> --- a/drivers/gpu/drm/xe/display/ext/intel_pch.c
> +++ /dev/null
> @@ -1,157 +0,0 @@
> -// SPDX-License-Identifier: MIT
> -/*
> - * Copyright 2019 Intel Corporation.
> - */
> -
> -#include "i915_drv.h"
> -#include "i915_utils.h"
> -#include "intel_pch.h"
> -
> -/* Map PCH device id to PCH type, or PCH_NONE if unknown. */
> -static enum intel_pch
> -intel_pch_type(const struct drm_i915_private *dev_priv, unsigned short id)
> -{
> -	switch (id) {
> -	case INTEL_PCH_MCC_DEVICE_ID_TYPE:
> -		drm_dbg_kms(&dev_priv->drm, "Found Mule Creek Canyon PCH\n");
> -		drm_WARN_ON(&dev_priv->drm, !IS_JSL_EHL(dev_priv));
> -		/* MCC is TGP compatible */
> -		return PCH_TGP;
> -	case INTEL_PCH_TGP_DEVICE_ID_TYPE:
> -	case INTEL_PCH_TGP2_DEVICE_ID_TYPE:
> -		drm_dbg_kms(&dev_priv->drm, "Found Tiger Lake LP PCH\n");
> -		drm_WARN_ON(&dev_priv->drm, !IS_TIGERLAKE(dev_priv) &&
> -			    !IS_ROCKETLAKE(dev_priv) &&
> -			    !IS_GEN9_BC(dev_priv));
> -		return PCH_TGP;
> -	case INTEL_PCH_ADP_DEVICE_ID_TYPE:
> -	case INTEL_PCH_ADP2_DEVICE_ID_TYPE:
> -	case INTEL_PCH_ADP3_DEVICE_ID_TYPE:
> -	case INTEL_PCH_ADP4_DEVICE_ID_TYPE:
> -		drm_dbg_kms(&dev_priv->drm, "Found Alder Lake PCH\n");
> -		drm_WARN_ON(&dev_priv->drm, !IS_ALDERLAKE_S(dev_priv) &&
> -			    !IS_ALDERLAKE_P(dev_priv));
> -		return PCH_ADP;
> -	case INTEL_PCH_MTP_DEVICE_ID_TYPE:
> -	case INTEL_PCH_MTP2_DEVICE_ID_TYPE:
> -		drm_dbg_kms(&dev_priv->drm, "Found Meteor Lake PCH\n");
> -		drm_WARN_ON(&dev_priv->drm, !IS_METEORLAKE(dev_priv));
> -		return PCH_MTP;
> -	default:
> -		return PCH_NONE;
> -	}
> -}
> -
> -static bool intel_is_virt_pch(unsigned short id,
> -			      unsigned short svendor, unsigned short sdevice)
> -{
> -	return (id == INTEL_PCH_P2X_DEVICE_ID_TYPE ||
> -		id == INTEL_PCH_P3X_DEVICE_ID_TYPE ||
> -		(id == INTEL_PCH_QEMU_DEVICE_ID_TYPE &&
> -		 svendor == PCI_SUBVENDOR_ID_REDHAT_QUMRANET &&
> -		 sdevice == PCI_SUBDEVICE_ID_QEMU));
> -}
> -
> -static void
> -intel_virt_detect_pch(const struct drm_i915_private *dev_priv,
> -		      unsigned short *pch_id, enum intel_pch *pch_type)
> -{
> -	unsigned short id = 0;
> -
> -	/*
> -	 * In a virtualized passthrough environment we can be in a
> -	 * setup where the ISA bridge is not able to be passed through.
> -	 * In this case, a south bridge can be emulated and we have to
> -	 * make an educated guess as to which PCH is really there.
> -	 */
> -
> -	if (IS_METEORLAKE(dev_priv))
> -		id = INTEL_PCH_MTP_DEVICE_ID_TYPE;
> -	else if (IS_ALDERLAKE_S(dev_priv) || IS_ALDERLAKE_P(dev_priv))
> -		id = INTEL_PCH_ADP_DEVICE_ID_TYPE;
> -	else if (IS_TIGERLAKE(dev_priv) || IS_ROCKETLAKE(dev_priv))
> -		id = INTEL_PCH_TGP_DEVICE_ID_TYPE;
> -	else BUG_ON(1);
> -
> -	if (id)
> -		drm_dbg_kms(&dev_priv->drm, "Assuming PCH ID %04x\n", id);
> -	else
> -		drm_dbg_kms(&dev_priv->drm, "Assuming no PCH\n");
> -
> -	*pch_type = intel_pch_type(dev_priv, id);
> -
> -	/* Sanity check virtual PCH id */
> -	if (drm_WARN_ON(&dev_priv->drm,
> -			id && *pch_type == PCH_NONE))
> -		id = 0;
> -
> -	*pch_id = id;
> -}
> -
> -void intel_detect_pch(struct drm_i915_private *dev_priv)
> -{
> -	struct pci_dev *pch = NULL;
> -	unsigned short id;
> -	enum intel_pch pch_type;
> -
> -	/* DG1 has south engine display on the same PCI device */
> -	if (IS_DG1(dev_priv)) {
> -		dev_priv->pch_type = PCH_DG1;
> -		return;
> -	} else if (IS_DG2(dev_priv)) {
> -		dev_priv->pch_type = PCH_DG2;
> -		return;
> -	}
> -
> -	/*
> -	 * The reason to probe ISA bridge instead of Dev31:Fun0 is to
> -	 * make graphics device passthrough work easy for VMM, that only
> -	 * need to expose ISA bridge to let driver know the real hardware
> -	 * underneath. This is a requirement from virtualization team.
> -	 *
> -	 * In some virtualized environments (e.g. XEN), there is irrelevant
> -	 * ISA bridge in the system. To work reliably, we should scan trhough
> -	 * all the ISA bridge devices and check for the first match, instead
> -	 * of only checking the first one.
> -	 */
> -	while ((pch = pci_get_class(PCI_CLASS_BRIDGE_ISA << 8, pch))) {
> -		if (pch->vendor != PCI_VENDOR_ID_INTEL)
> -			continue;
> -
> -		id = pch->device & INTEL_PCH_DEVICE_ID_MASK;
> -
> -		pch_type = intel_pch_type(dev_priv, id);
> -		if (pch_type != PCH_NONE) {
> -			dev_priv->pch_type = pch_type;
> -			dev_priv->pch_id = id;
> -			break;
> -		} else if (intel_is_virt_pch(id, pch->subsystem_vendor,
> -					     pch->subsystem_device)) {
> -			intel_virt_detect_pch(dev_priv, &id, &pch_type);
> -			dev_priv->pch_type = pch_type;
> -			dev_priv->pch_id = id;
> -			break;
> -		}
> -	}
> -
> -	/*
> -	 * Use PCH_NOP (PCH but no South Display) for PCH platforms without
> -	 * display.
> -	 */
> -	if (pch && !HAS_DISPLAY(dev_priv)) {
> -		drm_dbg_kms(&dev_priv->drm,
> -			    "Display disabled, reverting to NOP PCH\n");
> -		dev_priv->pch_type = PCH_NOP;
> -		dev_priv->pch_id = 0;
> -	} else if (!pch) {
> -		if (i915_run_as_guest() && HAS_DISPLAY(dev_priv)) {
> -			intel_virt_detect_pch(dev_priv, &id, &pch_type);
> -			dev_priv->pch_type = pch_type;
> -			dev_priv->pch_id = id;
> -		} else {
> -			drm_dbg_kms(&dev_priv->drm, "No PCH found.\n");
> -		}
> -	}
> -
> -	pci_dev_put(pch);
> -}
> diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
> index 88f863edc41c..54de3229553b 100644
> --- a/drivers/gpu/drm/xe/xe_device_types.h
> +++ b/drivers/gpu/drm/xe/xe_device_types.h
> @@ -18,7 +18,7 @@
>   
>   #if IS_ENABLED(CONFIG_DRM_XE_DISPLAY)
>   #include "ext/intel_device_info.h"
> -#include "ext/intel_pch.h"
> +#include "soc/intel_pch.h"
>   #include "intel_display_core.h"
>   #endif
>   
> diff --git a/drivers/gpu/drm/xe/xe_display.c b/drivers/gpu/drm/xe/xe_display.c
> index 64e65b29d8de..35dee29da832 100644
> --- a/drivers/gpu/drm/xe/xe_display.c
> +++ b/drivers/gpu/drm/xe/xe_display.c
> @@ -16,8 +16,8 @@
>   #include <drm/xe_drm.h>
>   
>   #include "ext/i915_irq.h"
> -#include "ext/intel_dram.h"
>   #include "ext/intel_pm.h"
> +#include "soc/intel_dram.h"
>   #include "intel_acpi.h"
>   #include "intel_audio.h"
>   #include "intel_bw.h"

-- 
Jani Nikula, Intel Open Source Graphics Center

[Intel-xe] [PATCH 6/6] drm/xe: Fixup __intel_de_wait_for_register macro.

[Maarten Lankhorst]

Create a dummy when output reg is null, this is required for the changes
to MTL display support.

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
---
 drivers/gpu/drm/xe/display/xe_de.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/gpu/drm/xe/display/xe_de.h b/drivers/gpu/drm/xe/display/xe_de.h
index 9f92fdb4159a..ce42e2938a06 100644
--- a/drivers/gpu/drm/xe/display/xe_de.h
+++ b/drivers/gpu/drm/xe/display/xe_de.h
@@ -70,6 +70,10 @@ __intel_de_wait_for_register(struct drm_i915_private *i915, i915_reg_t reg,
 			     unsigned int fast_timeout_us,
 			     unsigned int slow_timeout_ms, u32 *out_value)
 {
+	u32 dummy;
+	if (!out_value)
+		out_value = &dummy;
+
 	return wait_for_atomic(((*out_value = xe_mmio_read32(to_gt(i915), reg.reg)) & mask) == value,
 			slow_timeout_ms);
 }
-- 
2.34.1

Re: [Intel-xe] [PATCH 6/6] drm/xe: Fixup __intel_de_wait_for_register macro.

[Lucas De Marchi]

On Fri, Mar 31, 2023 at 12:24:19PM +0200, Maarten Lankhorst wrote:
>Create a dummy when output reg is null, this is required for the changes
>to MTL display support.

drm/xe/display here too. Or if you prefer to handle this as a fixup
commit to be squashed in next rebase, commit this is --fixup=<commit>

>
>Signed-off-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
>---
> drivers/gpu/drm/xe/display/xe_de.h | 4 ++++
> 1 file changed, 4 insertions(+)
>
>diff --git a/drivers/gpu/drm/xe/display/xe_de.h b/drivers/gpu/drm/xe/display/xe_de.h
>index 9f92fdb4159a..ce42e2938a06 100644
>--- a/drivers/gpu/drm/xe/display/xe_de.h
>+++ b/drivers/gpu/drm/xe/display/xe_de.h
>@@ -70,6 +70,10 @@ __intel_de_wait_for_register(struct drm_i915_private *i915, i915_reg_t reg,
> 			     unsigned int fast_timeout_us,
> 			     unsigned int slow_timeout_ms, u32 *out_value)
> {
>+	u32 dummy;

missing new line.


More commonly these kind of situations are handled the other way around
though:

	u32 x;

	ret = wait_for_atomic(((x = ...

	if (out_value)
		*out_value = x;

	return ret;

I don't mind either way though, up to you.

Note that in xe we got rid of wait_for() and friends though.
Do we want to drop this one too?

Lucas De Marchi

>+	if (!out_value)
>+		out_value = &dummy;
>+
> 	return wait_for_atomic(((*out_value = xe_mmio_read32(to_gt(i915), reg.reg)) & mask) == value,
> 			slow_timeout_ms);
> }
>-- 
>2.34.1
>

[Intel-xe] ✓ CI.Patch_applied: success for drm/xe: Meteorlake fixes.

[Patchwork]

== Series Details ==

Series: drm/xe: Meteorlake fixes.
URL   : https://patchwork.freedesktop.org/series/115930/
State : success

== Summary ==

=== Applying kernel patches on branch 'drm-xe-next' with base: ===
commit 4b3608ff06e78b835a2e50eec94cd041c95e4326
Author:     Matt Roper <matthew.d.roper@intel.com>
AuthorDate: Wed Mar 29 10:33:34 2023 -0700
Commit:     Matt Roper <matthew.d.roper@intel.com>
CommitDate: Thu Mar 30 12:20:22 2023 -0700

    drm/xe: Don't emit extra MI_BATCH_BUFFER_END in WA batchbuffer
    
    The MI_BATCH_BUFFER_END is already added automatically by
    __xe_bb_create_job(); including it in the construction of the workaround
    batchbuffer results in an unnecessary duplicate.
    
    Link: https://lore.kernel.org/r/20230329173334.4015124-4-matthew.d.roper@intel.com
    Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
    Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
=== git am output follows ===
Applying: drm/xe: Fix meteorlake stolen memory
Applying: drm/xe: Use full_gt batchbuffer allocation for media tiles.
Applying: drm/i915: Fix comparison in intel_dram.
Applying: drm/xe: Fix XE_LPDP and meteorlake display info.
Applying: drm/xe: Build soc files directly
Applying: drm/xe: Fixup __intel_de_wait_for_register macro.

[Intel-xe] ✗ CI.KUnit: failure for drm/xe: Meteorlake fixes.

[Patchwork]

== Series Details ==

Series: drm/xe: Meteorlake fixes.
URL   : https://patchwork.freedesktop.org/series/115930/
State : failure

== Summary ==

+ trap cleanup EXIT
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/xe/.kunitconfig
[10:26:41] Configuring KUnit Kernel ...
Generating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig
[10:26:46] Building KUnit Kernel ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig
Building with:
$ make ARCH=um O=.kunit --jobs=48
ERROR:root:In file included from /usr/include/stdlib.h:1013,
                 from ../arch/um/os-Linux/irq.c:8:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
In file included from /usr/include/stdlib.h:1013,
                 from ../arch/um/os-Linux/helper.c:6:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
In file included from /usr/include/stdlib.h:1013,
                 from ../arch/um/drivers/chan_user.c:6:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
In file included from /usr/include/stdlib.h:1013,
                 from ../arch/um/drivers/fd.c:7:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
In file included from /usr/include/stdlib.h:1013,
                 from ../arch/x86/um/os-Linux/registers.c:8:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
In file included from /usr/include/stdlib.h:1013,
                 from ../arch/um/os-Linux/execvp.c:24:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
In file included from /usr/include/stdlib.h:1013,
                 from ../arch/um/os-Linux/main.c:8:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
In file included from /usr/include/stdlib.h:1013,
                 from ../arch/um/os-Linux/mem.c:8:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
In file included from /usr/include/stdlib.h:1013,
                 from ../arch/um/os-Linux/file.c:8:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
make[3]: *** [../scripts/Makefile.build:252: arch/um/os-Linux/irq.o] Error 1
make[3]: *** Waiting for unfinished jobs....
In file included from /usr/include/stdlib.h:1013,
                 from ../arch/x86/um/os-Linux/task_size.c:3:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
make[4]: *** [../scripts/Makefile.build:252: arch/x86/um/os-Linux/registers.o] Error 1
make[4]: *** Waiting for unfinished jobs....
make[3]: *** [../scripts/Makefile.build:252: arch/um/os-Linux/execvp.o] Error 1
make[3]: *** [../scripts/Makefile.build:252: arch/um/drivers/fd.o] Error 1
make[3]: *** Waiting for unfinished jobs....
make[4]: *** [../scripts/Makefile.build:252: arch/x86/um/os-Linux/task_size.o] Error 1
make[3]: *** [../scripts/Makefile.build:494: arch/x86/um/os-Linux] Error 2
make[3]: *** Waiting for unfinished jobs....
make[3]: *** [../scripts/Makefile.build:252: arch/um/drivers/chan_user.o] Error 1
make[3]: *** [../scripts/Makefile.build:252: arch/um/os-Linux/helper.o] Error 1
make[3]: *** [../scripts/Makefile.build:252: arch/um/os-Linux/main.o] Error 1
make[3]: *** [../scripts/Makefile.build:252: arch/um/os-Linux/mem.o] Error 1
make[3]: *** [../scripts/Makefile.build:252: arch/um/os-Linux/file.o] Error 1
In file included from /usr/include/stdlib.h:1013,
                 from ../arch/um/os-Linux/skas/process.c:7:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
make[4]: *** [../scripts/Makefile.build:252: arch/um/os-Linux/skas/process.o] Error 1
make[3]: *** [../scripts/Makefile.build:494: arch/um/os-Linux/skas] Error 2
make[2]: *** [../scripts/Makefile.build:494: arch/um/os-Linux] Error 2
make[2]: *** Waiting for unfinished jobs....
make[2]: *** [../scripts/Makefile.build:494: arch/x86/um] Error 2
make[2]: *** [../scripts/Makefile.build:494: arch/um/drivers] Error 2
In file included from /usr/include/stdlib.h:1013,
                 from arch/um/kernel/config.c:7:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h: In function ‘atof’:
/usr/include/x86_64-linux-gnu/bits/stdlib-float.h:26:1: error: SSE register return with SSE disabled
   26 | {
      | ^
make[3]: *** [../scripts/Makefile.build:252: arch/um/kernel/config.o] Error 1
make[3]: *** Waiting for unfinished jobs....
make[2]: *** [../scripts/Makefile.build:494: arch/um/kernel] Error 2
make[1]: *** [/kernel/Makefile:2025: .] Error 2
make: *** [Makefile:226: __sub-make] Error 2

+ cleanup
++ stat -c %u:%g /kernel
+ chown -R 1003:1003 /kernel

[Intel-xe] ✗ CI.Patch_applied: failure for drm/xe: Meteorlake fixes. (rev2)

[Patchwork]

== Series Details ==

Series: drm/xe: Meteorlake fixes. (rev2)
URL   : https://patchwork.freedesktop.org/series/115930/
State : failure

== Summary ==

=== Applying kernel patches on branch 'drm-xe-next' with base: ===
commit cbd8b47f28266aad8f9786be9a1899e3f0e0f2e0
Author:     Matt Roper <matthew.d.roper@intel.com>
AuthorDate: Mon Mar 27 10:34:33 2023 -0700
Commit:     Matt Roper <matthew.d.roper@intel.com>
CommitDate: Fri Mar 31 17:06:44 2023 -0700

    fixup! drm/xe/pat: Handle unicast vs MCR PAT registers
    
    Somehow I dropped the error message when sending this; add it back.
    
    To be squashed next time the Xe driver is rebased.
    
    Fixes: 11946af00717 ("drm/xe/pat: Handle unicast vs MCR PAT registers")
    Reviewed-by: Nirmoy Das <nirmoy.das@intel.com>
    Link: https://lore.kernel.org/r/20230327173433.2957645-1-matthew.d.roper@intel.com
    Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
=== git am output follows ===
error: patch failed: drivers/gpu/drm/i915/soc/intel_dram.c:9
error: drivers/gpu/drm/i915/soc/intel_dram.c: patch does not apply
error: patch failed: drivers/gpu/drm/xe/Makefile:124
error: drivers/gpu/drm/xe/Makefile: patch does not apply
error: patch failed: drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h:15
error: drivers/gpu/drm/xe/compat-i915-headers/i915_drv.h: patch does not apply
error: patch failed: drivers/gpu/drm/xe/display/ext/intel_dram.h:8
error: drivers/gpu/drm/xe/display/ext/intel_dram.h: patch does not apply
error: patch failed: drivers/gpu/drm/xe/xe_device_types.h:18
error: drivers/gpu/drm/xe/xe_device_types.h: patch does not apply
error: patch failed: drivers/gpu/drm/xe/xe_display.c:16
error: drivers/gpu/drm/xe/xe_display.c: patch does not apply
hint: Use 'git am --show-current-patch' to see the failed patch
Applying: drm/xe: Fix meteorlake stolen memory
Applying: drm/xe: Use full_gt batchbuffer allocation for media tiles.
Applying: drm/i915: Fix comparison in intel_dram.
Applying: drm/xe: Fix XE_LPDP and meteorlake display info.
Applying: drm/xe: Build soc files directly
Patch failed at 0005 drm/xe: Build soc files directly
When you have resolved this problem, run "git am --continue".
If you prefer to skip this patch, run "git am --skip" instead.
To restore the original branch and stop patching, run "git am --abort".