[PATCH 0/7] Include <asm/cpu_device_id.h> without #ifdef CONFIG_X86

[PATCH 0/7] Include <asm/cpu_device_id.h> without #ifdef CONFIG_X86

[Pavel Popov]

<asm/cpu_device_id.h> exists only on x86, so generic code that uses it
wraps the include in #ifdef CONFIG_X86. The P2PDMA series [1] adds a
generic version of the header, so the include can be unconditional.

This series drops the #ifdef around the include in the remaining
users, and in ccp, which does not use the header, the include itself.
Some of them also include <asm/intel-family.h>, which has no
generic version. That include is removed: on x86 it is already pulled
in by <asm/cpu_device_id.h>, and on other architectures the INTEL_*
model defines are not used. The #ifdef stays around code that needs
x86 for other reasons: CPU match tables using the INTEL_* model
defines and boot_cpu_data accesses.

Patches 1-6 are independent of each other and can go through the
respective trees once [1] has landed. Patch 7 does not depend on [1].

Based on mainline (7.4-rc2+) plus [1]. Built with x86_64 and arm64
allmodconfig, each patch separately.

[1] https://lore.kernel.org/r/20261111234439.4098-1-pavel.e.popov@intel.com

Pavel Popov (7):
  ACPI: mipi-disco-img: Include <asm/cpu_device_id.h> unconditionally
  drm/i915: Include <asm/cpu_device_id.h> unconditionally
  drm/amdgpu: Include <asm/cpu_device_id.h> unconditionally
  cpufreq: ondemand: Include <asm/cpu_device_id.h> unconditionally
  platform/x86: Include <asm/cpu_device_id.h> unconditionally in soc.h
  tpm: Include <asm/cpu_device_id.h> unconditionally
  crypto: ccp - Drop unused <asm/cpu_device_id.h> include

 drivers/acpi/mipi-disco-img.c              | 5 ++---
 drivers/char/tpm/tpm.h                     | 2 --
 drivers/cpufreq/cpufreq_ondemand.h         | 4 ++--
 drivers/crypto/ccp/ccp-dev.c               | 3 ---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 3 ---
 drivers/gpu/drm/i915/intel_cpu_info.c      | 5 ++---
 include/linux/platform_data/x86/soc.h      | 5 ++---
 7 files changed, 8 insertions(+), 19 deletions(-)


base-commit: 067c5c24cef23d913b67eaaf71ae67e7169a6620
prerequisite-patch-id: 2a8952c8b8b6b019585a8bc449db628f458658cb
prerequisite-patch-id: 6b1715c7e4f230fedfcadf38303ce5966922caff
prerequisite-patch-id: 88872ad297c6b8ec80cb4ec4057ab281e10c148d
-- 
2.54.0

[PATCH 1/7] ACPI: mipi-disco-img: Include <asm/cpu_device_id.h> unconditionally

[Pavel Popov]

<asm/cpu_device_id.h> now has a generic version, so the #ifdef
CONFIG_X86 around the include is no longer needed.

<asm/intel-family.h> is included by <asm/cpu_device_id.h>. The CPU
match table still needs CONFIG_X86 for the INTEL_* model defines.

Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>
---
Depends on <asm-generic/cpu_device_id.h> from [1]; please apply after
it is in mainline.

[1] https://lore.kernel.org/r/20261111234439.4098-1-pavel.e.popov@intel.com

 drivers/acpi/mipi-disco-img.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/acpi/mipi-disco-img.c b/drivers/acpi/mipi-disco-img.c
index ea4bffe3c02..3d841d2bdfd 100644
--- a/drivers/acpi/mipi-disco-img.c
+++ b/drivers/acpi/mipi-disco-img.c
@@ -30,6 +30,8 @@
 
 #include <media/v4l2-fwnode.h>
 
+#include <asm/cpu_device_id.h>
+
 #include "internal.h"
 
 static LIST_HEAD(acpi_mipi_crs_csi2_list);
@@ -724,9 +726,6 @@ void acpi_mipi_crs_csi2_cleanup(void)
 }
 
 #ifdef CONFIG_X86
-#include <asm/cpu_device_id.h>
-#include <asm/intel-family.h>
-
 /* CPU matches for Dell generations with broken ACPI MIPI DISCO info */
 static const struct x86_cpu_id dell_broken_mipi_disco_cpu_gens[] = {
 	X86_MATCH_VFM(INTEL_TIGERLAKE, NULL),
-- 
2.54.0

[PATCH 2/7] drm/i915: Include <asm/cpu_device_id.h> unconditionally

[Pavel Popov]

<asm/cpu_device_id.h> now has a generic version, so the #ifdef
CONFIG_X86 around the include is no longer needed.

<asm/intel-family.h> is included by <asm/cpu_device_id.h>. The CPU
match table still needs CONFIG_X86 for the INTEL_* model defines.

Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>
---
Depends on <asm-generic/cpu_device_id.h> from [1]; please apply after
it is in mainline.

[1] https://lore.kernel.org/r/20261111234439.4098-1-pavel.e.popov@intel.com

 drivers/gpu/drm/i915/intel_cpu_info.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/i915/intel_cpu_info.c b/drivers/gpu/drm/i915/intel_cpu_info.c
index e52d0ac713a..a6b46cc841a 100644
--- a/drivers/gpu/drm/i915/intel_cpu_info.c
+++ b/drivers/gpu/drm/i915/intel_cpu_info.c
@@ -6,12 +6,11 @@
  * intel_device_info.h by having a separate file.
  */
 
+#include <asm/cpu_device_id.h>
+
 #include "intel_cpu_info.h"
 
 #ifdef CONFIG_X86
-#include <asm/cpu_device_id.h>
-#include <asm/intel-family.h>
-
 static const struct x86_cpu_id g8_cpu_ids[] = {
 	X86_MATCH_VFM(INTEL_ALDERLAKE,		NULL),
 	X86_MATCH_VFM(INTEL_ALDERLAKE_L,	NULL),
-- 
2.54.0

Re: [PATCH 2/7] drm/i915: Include <asm/cpu_device_id.h> unconditionally

[Raag Jadav]

Cc: Andy

On Wed, Nov 11, 2026 at 04:48:15PM -0700, Pavel Popov wrote:
> <asm/cpu_device_id.h> now has a generic version, so the #ifdef
> CONFIG_X86 around the include is no longer needed.
> 
> <asm/intel-family.h> is included by <asm/cpu_device_id.h>. The CPU
> match table still needs CONFIG_X86 for the INTEL_* model defines.
> 
> Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>

Reviewed-by: Raag Jadav <raag.jadav@intel.com>

> ---
> Depends on <asm-generic/cpu_device_id.h> from [1]; please apply after
> it is in mainline.
> 
> [1] https://lore.kernel.org/r/20261111234439.4098-1-pavel.e.popov@intel.com
> 
>  drivers/gpu/drm/i915/intel_cpu_info.c | 5 ++---
>  1 file changed, 2 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/intel_cpu_info.c b/drivers/gpu/drm/i915/intel_cpu_info.c
> index e52d0ac713a..a6b46cc841a 100644
> --- a/drivers/gpu/drm/i915/intel_cpu_info.c
> +++ b/drivers/gpu/drm/i915/intel_cpu_info.c
> @@ -6,12 +6,11 @@
>   * intel_device_info.h by having a separate file.
>   */
>  
> +#include <asm/cpu_device_id.h>
> +
>  #include "intel_cpu_info.h"
>  
>  #ifdef CONFIG_X86
> -#include <asm/cpu_device_id.h>
> -#include <asm/intel-family.h>
> -
>  static const struct x86_cpu_id g8_cpu_ids[] = {
>  	X86_MATCH_VFM(INTEL_ALDERLAKE,		NULL),
>  	X86_MATCH_VFM(INTEL_ALDERLAKE_L,	NULL),
> -- 
> 2.54.0
> 

[PATCH 3/7] drm/amdgpu: Include <asm/cpu_device_id.h> unconditionally

[Pavel Popov]

<asm/cpu_device_id.h> now has a generic version, so the #ifdef
CONFIG_X86 around the include is no longer needed.

<asm/intel-family.h> is included by <asm/cpu_device_id.h>.

Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>
---
Depends on <asm-generic/cpu_device_id.h> from [1]; please apply after
it is in mainline.

[1] https://lore.kernel.org/r/20261111234439.4098-1-pavel.e.popov@intel.com

 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 19a566187df..fc4e7273630 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -88,10 +88,7 @@
 
 #include <drm/drm_drv.h>
 
-#if IS_ENABLED(CONFIG_X86)
-#include <asm/intel-family.h>
 #include <asm/cpu_device_id.h>
-#endif
 
 MODULE_FIRMWARE("amdgpu/vega10_gpu_info.bin");
 MODULE_FIRMWARE("amdgpu/vega12_gpu_info.bin");
-- 
2.54.0

[PATCH 4/7] cpufreq: ondemand: Include <asm/cpu_device_id.h> unconditionally

[Pavel Popov]

<asm/cpu_device_id.h> now has a generic version, so the #ifdef
CONFIG_X86 around the include is no longer needed.

od_should_io_be_busy() still needs CONFIG_X86 for boot_cpu_data.

Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>
---
Depends on <asm-generic/cpu_device_id.h> from [1]; please apply after
it is in mainline.

[1] https://lore.kernel.org/r/20261111234439.4098-1-pavel.e.popov@intel.com

 drivers/cpufreq/cpufreq_ondemand.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_ondemand.h b/drivers/cpufreq/cpufreq_ondemand.h
index 2ca8f1aaf2e..0e33865f9d3 100644
--- a/drivers/cpufreq/cpufreq_ondemand.h
+++ b/drivers/cpufreq/cpufreq_ondemand.h
@@ -6,6 +6,8 @@
  * Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
  */
 
+#include <asm/cpu_device_id.h>
+
 #include "cpufreq_governor.h"
 
 struct od_policy_dbs_info {
@@ -26,8 +28,6 @@ struct od_dbs_tuners {
 };
 
 #ifdef CONFIG_X86
-#include <asm/cpu_device_id.h>
-
 /*
  * Not all CPUs want IO time to be accounted as busy; this depends on
  * how efficient idling at a higher frequency/voltage is.
-- 
2.54.0

Re: [PATCH 4/7] cpufreq: ondemand: Include <asm/cpu_device_id.h> unconditionally

[Zhongqiu Han]

On 11/12/2026 7:48 AM, Pavel Popov wrote:
> <asm/cpu_device_id.h> now has a generic version, so the #ifdef
> CONFIG_X86 around the include is no longer needed.
> 
> od_should_io_be_busy() still needs CONFIG_X86 for boot_cpu_data.
> 
> Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>
> ---
> Depends on <asm-generic/cpu_device_id.h> from [1]; please apply after
> it is in mainline.


Yeah, this patch strictly depends on [1] (series patch 1/3) being in
mainline first; otherwise non-x86 builds break.

Reviewed-by: Zhongqiu Han <zhongqiu.han@oss.qualcomm.com>

> 
> [1] https://lore.kernel.org/r/20261111234439.4098-1-pavel.e.popov@intel.com
> 
>   drivers/cpufreq/cpufreq_ondemand.h | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/cpufreq/cpufreq_ondemand.h b/drivers/cpufreq/cpufreq_ondemand.h
> index 2ca8f1aaf2e..0e33865f9d3 100644
> --- a/drivers/cpufreq/cpufreq_ondemand.h
> +++ b/drivers/cpufreq/cpufreq_ondemand.h
> @@ -6,6 +6,8 @@
>    * Author: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
>    */
>   
> +#include <asm/cpu_device_id.h>
> +
>   #include "cpufreq_governor.h"
>   
>   struct od_policy_dbs_info {
> @@ -26,8 +28,6 @@ struct od_dbs_tuners {
>   };
>   
>   #ifdef CONFIG_X86
> -#include <asm/cpu_device_id.h>
> -
>   /*
>    * Not all CPUs want IO time to be accounted as busy; this depends on
>    * how efficient idling at a higher frequency/voltage is.


-- 
Thx and BRs,
Zhongqiu Han

[PATCH 5/7] platform/x86: Include <asm/cpu_device_id.h> unconditionally in soc.h

[Pavel Popov]

<asm/cpu_device_id.h> now has a generic version, so the #ifdef
CONFIG_X86 around the include is no longer needed. The include of
<linux/device-id/x86_cpu.h>, which is generic, moves out of the
#ifdef as well.

The soc_intel_is_*() helpers still need CONFIG_X86 for the INTEL_*
model defines.

Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>
---
Depends on <asm-generic/cpu_device_id.h> from [1]; please apply after
it is in mainline.

[1] https://lore.kernel.org/r/20261111234439.4098-1-pavel.e.popov@intel.com

 include/linux/platform_data/x86/soc.h | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/include/linux/platform_data/x86/soc.h b/include/linux/platform_data/x86/soc.h
index a6a6b313dfa..ee696e576a0 100644
--- a/include/linux/platform_data/x86/soc.h
+++ b/include/linux/platform_data/x86/soc.h
@@ -9,13 +9,12 @@
 #define __PLATFORM_DATA_X86_SOC_H
 
 #include <linux/types.h>
-
-#if IS_ENABLED(CONFIG_X86)
-
 #include <linux/device-id/x86_cpu.h>
 
 #include <asm/cpu_device_id.h>
 
+#if IS_ENABLED(CONFIG_X86)
+
 #define SOC_INTEL_IS_CPU(soc, type)				\
 static inline bool soc_intel_is_##soc(void)			\
 {								\
-- 
2.54.0

[PATCH 6/7] tpm: Include <asm/cpu_device_id.h> unconditionally

[Pavel Popov]

<asm/cpu_device_id.h> now has a generic version, so the #ifdef
CONFIG_X86 around the include is no longer needed.

Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>
---
Depends on <asm-generic/cpu_device_id.h> from [1]; please apply after
it is in mainline.

[1] https://lore.kernel.org/r/20261111234439.4098-1-pavel.e.popov@intel.com

 drivers/char/tpm/tpm.h | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/char/tpm/tpm.h b/drivers/char/tpm/tpm.h
index fa554c5ad80..ad1534aa00c 100644
--- a/drivers/char/tpm/tpm.h
+++ b/drivers/char/tpm/tpm.h
@@ -28,9 +28,7 @@
 #include <linux/tpm.h>
 #include <linux/tpm_eventlog.h>
 
-#ifdef CONFIG_X86
 #include <asm/cpu_device_id.h>
-#endif
 
 #define TPM_MINOR		224	/* officially assigned */
 #define TPM_NUM_DEVICES		65536
-- 
2.54.0

Re: [PATCH 6/7] tpm: Include <asm/cpu_device_id.h> unconditionally

[Jarkko Sakkinen]

On Wed, Nov 11, 2026 at 04:48:19PM -0700, Pavel Popov wrote:
> <asm/cpu_device_id.h> now has a generic version, so the #ifdef
> CONFIG_X86 around the include is no longer needed.
> 
> Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>

Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>

Are these expected to be picked separately by subsystems?

Br, Jarkko

RE: [PATCH 6/7] tpm: Include <asm/cpu_device_id.h> unconditionally

[Popov, Pavel E]

> Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
>
> Are these expected to be picked separately by subsystems?

Yes, each through its own subsystem tree, once [1] is in mainline.
The tpm patch does not build on non-x86 before that.

[1] https://lore.kernel.org/r/20261111234439.4098-1-pavel.e.popov@intel.com

Thanks,
Pavel

> -----Original Message-----
> From: Jarkko Sakkinen <jarkko@kernel.org>
> Sent: Wednesday, November 11, 2026 5:40 PM
> To: Popov, Pavel E <pavel.e.popov@intel.com>
> Cc: Rafael J. Wysocki <rafael@kernel.org>; Viresh Kumar
> <viresh.kumar@linaro.org>; Jani Nikula <jani.nikula@linux.intel.com>; Joonas
> Lahtinen <joonas.lahtinen@linux.intel.com>; Vivi, Rodrigo
> <rodrigo.vivi@intel.com>; Tvrtko Ursulin <tursulin@ursulin.net>; Alex Deucher
> <alexander.deucher@amd.com>; Christian König
> <christian.koenig@amd.com>; Hans de Goede <hansg@kernel.org>; Ilpo
> Järvinen <ilpo.jarvinen@linux.intel.com>; Peter Huewe
> <peterhuewe@gmx.de>; Tom Lendacky <thomas.lendacky@amd.com>;
> Herbert Xu <herbert@gondor.apana.org.au>; Leon Romanovsky
> <leon@kernel.org>; Bjorn Helgaas <bhelgaas@google.com>; Logan Gunthorpe
> <logang@deltatee.com>; Arnd Bergmann <arnd@arndb.de>; Len Brown
> <lenb@kernel.org>; Sakari Ailus <sakari.ailus@linux.intel.com>; Jason
> Gunthorpe <jgg@ziepe.ca>; David Airlie <airlied@gmail.com>; Simona Vetter
> <simona@ffwll.ch>; David S . Miller <davem@davemloft.net>;
> x86@kernel.org; linux-acpi@vger.kernel.org; linux-pm@vger.kernel.org; intel-
> gfx@lists.freedesktop.org; amd-gfx@lists.freedesktop.org; dri-
> devel@lists.freedesktop.org; platform-driver-x86@vger.kernel.org; linux-
> integrity@vger.kernel.org; linux-crypto@vger.kernel.org; linux-
> arch@vger.kernel.org; linux-pci@vger.kernel.org; linux-kernel@vger.kernel.org
> Subject: Re: [PATCH 6/7] tpm: Include <asm/cpu_device_id.h> unconditionally
> 
> On Wed, Nov 11, 2026 at 04:48:19PM -0700, Pavel Popov wrote:
> > <asm/cpu_device_id.h> now has a generic version, so the #ifdef
> > CONFIG_X86 around the include is no longer needed.
> >
> > Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>
> 
> Reviewed-by: Jarkko Sakkinen <jarkko@kernel.org>
> 
> Are these expected to be picked separately by subsystems?
> 
> Br, Jarkko

[PATCH 7/7] crypto: ccp - Drop unused <asm/cpu_device_id.h> include

[Pavel Popov]

Nothing in ccp-dev.c uses <asm/cpu_device_id.h>; the CPU match table
it was included for is long gone. Drop the include and the #ifdef
CONFIG_X86 around it.

Signed-off-by: Pavel Popov <pavel.e.popov@intel.com>
---
 drivers/crypto/ccp/ccp-dev.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/drivers/crypto/ccp/ccp-dev.c b/drivers/crypto/ccp/ccp-dev.c
index 41d9bdac95d..017e7a6551f 100644
--- a/drivers/crypto/ccp/ccp-dev.c
+++ b/drivers/crypto/ccp/ccp-dev.c
@@ -21,9 +21,6 @@
 #include <linux/hw_random.h>
 #include <linux/cpu.h>
 #include <linux/atomic.h>
-#ifdef CONFIG_X86
-#include <asm/cpu_device_id.h>
-#endif
 #include <linux/ccp.h>
 
 #include "ccp-dev.h"
-- 
2.54.0