ICL changes the registers and addresses to 64 bits.
I also briefly looked at implementing an u64 version of the PCI config read functions, but I concluded this wouldn't be trivial, so it's not worth doing it for a single user that can't have any racing problems while reading the register in two separate operations.
v2: - Scrub the development (non-public) changelog (Joonas). - Remove the i915.ko bits so this can be easily backported in order to properly avoid stolen memory even on machines without i915.ko (Joonas). - CC stable for the reasons above.
Issue: VIZ-9250 CC: stable@vger.kernel.org Cc: Ingo Molnar mingo@kernel.org Cc: H. Peter Anvin hpa@zytor.com Cc: x86@kernel.org Cc: Daniele Ceraolo Spurio daniele.ceraolospurio@intel.com Cc: Joonas Lahtinen joonas.lahtinen@linux.intel.com Signed-off-by: Paulo Zanoni paulo.r.zanoni@intel.com --- arch/x86/kernel/early-quirks.c | 18 ++++++++++++++++++ include/drm/i915_drm.h | 4 +++- 2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index bae0d32e327b..72c2cf961d44 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/arch/x86/kernel/early-quirks.c @@ -340,6 +340,18 @@ static resource_size_t __init gen3_stolen_base(int num, int slot, int func, return bsm & INTEL_BSM_MASK; }
+static resource_size_t __init gen11_stolen_base(int num, int slot, int func, + resource_size_t stolen_size) +{ + u64 bsm; + + bsm = read_pci_config(num, slot, func, INTEL_GEN11_BSM_DW0); + bsm &= INTEL_BSM_MASK; + bsm |= (u64)read_pci_config(num, slot, func, INTEL_GEN11_BSM_DW1) << 32; + + return bsm; +} + static resource_size_t __init i830_stolen_size(int num, int slot, int func) { u16 gmch_ctrl; @@ -500,6 +512,11 @@ static const struct intel_early_ops chv_early_ops __initconst = { .stolen_size = chv_stolen_size, };
+static const struct intel_early_ops gen11_early_ops __initconst = { + .stolen_base = gen11_stolen_base, + .stolen_size = gen9_stolen_size, +}; + static const struct pci_device_id intel_early_ids[] __initconst = { INTEL_I830_IDS(&i830_early_ops), INTEL_I845G_IDS(&i845_early_ops), @@ -531,6 +548,7 @@ static const struct pci_device_id intel_early_ids[] __initconst = { INTEL_CFL_IDS(&gen9_early_ops), INTEL_GLK_IDS(&gen9_early_ops), INTEL_CNL_IDS(&gen9_early_ops), + INTEL_ICL_11_IDS(&gen11_early_ops), };
struct resource intel_graphics_stolen_res __ro_after_init = DEFINE_RES_MEM(0, 0); diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h index c9e5a6621b95..c44703f471b3 100644 --- a/include/drm/i915_drm.h +++ b/include/drm/i915_drm.h @@ -95,7 +95,9 @@ extern struct resource intel_graphics_stolen_res; #define I845_TSEG_SIZE_512K (2 << 1) #define I845_TSEG_SIZE_1M (3 << 1)
-#define INTEL_BSM 0x5c +#define INTEL_BSM 0x5c +#define INTEL_GEN11_BSM_DW0 0xc0 +#define INTEL_GEN11_BSM_DW1 0xc4 #define INTEL_BSM_MASK (-(1u << 20))
#endif /* _I915_DRM_H_ */
Ingo, do you prefer to merge through our tree with your ack?
Quoting Paulo Zanoni (2018-05-04 23:32:51)
ICL changes the registers and addresses to 64 bits.
I also briefly looked at implementing an u64 version of the PCI config read functions, but I concluded this wouldn't be trivial, so it's not worth doing it for a single user that can't have any racing problems while reading the register in two separate operations.
v2:
- Scrub the development (non-public) changelog (Joonas).
- Remove the i915.ko bits so this can be easily backported in order to properly avoid stolen memory even on machines without i915.ko (Joonas).
- CC stable for the reasons above.
Issue: VIZ-9250
Fixes: 412310019a20 ("drm/i915/icl: Add initial Icelake definitions.")
CC: stable@vger.kernel.org
This should not be needed, it was introduced in v4.17-rc1 only.
Reviewed-by: Joonas Lahtinen joonas.lahtinen@linux.intel.com
Regards, Joonas
Cc: Ingo Molnar mingo@kernel.org Cc: H. Peter Anvin hpa@zytor.com Cc: x86@kernel.org Cc: Daniele Ceraolo Spurio daniele.ceraolospurio@intel.com Cc: Joonas Lahtinen joonas.lahtinen@linux.intel.com Signed-off-by: Paulo Zanoni paulo.r.zanoni@intel.com
arch/x86/kernel/early-quirks.c | 18 ++++++++++++++++++ include/drm/i915_drm.h | 4 +++- 2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index bae0d32e327b..72c2cf961d44 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/arch/x86/kernel/early-quirks.c @@ -340,6 +340,18 @@ static resource_size_t __init gen3_stolen_base(int num, int slot, int func, return bsm & INTEL_BSM_MASK; } +static resource_size_t __init gen11_stolen_base(int num, int slot, int func,
resource_size_t stolen_size)
+{
u64 bsm;
bsm = read_pci_config(num, slot, func, INTEL_GEN11_BSM_DW0);
bsm &= INTEL_BSM_MASK;
bsm |= (u64)read_pci_config(num, slot, func, INTEL_GEN11_BSM_DW1) << 32;
return bsm;
+}
static resource_size_t __init i830_stolen_size(int num, int slot, int func) { u16 gmch_ctrl; @@ -500,6 +512,11 @@ static const struct intel_early_ops chv_early_ops __initconst = { .stolen_size = chv_stolen_size, }; +static const struct intel_early_ops gen11_early_ops __initconst = {
.stolen_base = gen11_stolen_base,
.stolen_size = gen9_stolen_size,
+};
static const struct pci_device_id intel_early_ids[] __initconst = { INTEL_I830_IDS(&i830_early_ops), INTEL_I845G_IDS(&i845_early_ops), @@ -531,6 +548,7 @@ static const struct pci_device_id intel_early_ids[] __initconst = { INTEL_CFL_IDS(&gen9_early_ops), INTEL_GLK_IDS(&gen9_early_ops), INTEL_CNL_IDS(&gen9_early_ops),
INTEL_ICL_11_IDS(&gen11_early_ops),
}; struct resource intel_graphics_stolen_res __ro_after_init = DEFINE_RES_MEM(0, 0); diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h index c9e5a6621b95..c44703f471b3 100644 --- a/include/drm/i915_drm.h +++ b/include/drm/i915_drm.h @@ -95,7 +95,9 @@ extern struct resource intel_graphics_stolen_res; #define I845_TSEG_SIZE_512K (2 << 1) #define I845_TSEG_SIZE_1M (3 << 1) -#define INTEL_BSM 0x5c +#define INTEL_BSM 0x5c +#define INTEL_GEN11_BSM_DW0 0xc0 +#define INTEL_GEN11_BSM_DW1 0xc4 #define INTEL_BSM_MASK (-(1u << 20))
#endif /* _I915_DRM_H_ */
2.14.3
Em Seg, 2018-05-07 às 11:46 +0300, Joonas Lahtinen escreveu:
Ingo, do you prefer to merge through our tree with your ack?
Ping?
Quoting Paulo Zanoni (2018-05-04 23:32:51)
ICL changes the registers and addresses to 64 bits.
I also briefly looked at implementing an u64 version of the PCI config read functions, but I concluded this wouldn't be trivial, so it's not worth doing it for a single user that can't have any racing problems while reading the register in two separate operations.
v2:
- Scrub the development (non-public) changelog (Joonas).
- Remove the i915.ko bits so this can be easily backported in
order to properly avoid stolen memory even on machines without i915.ko (Joonas).
- CC stable for the reasons above.
Issue: VIZ-9250
Fixes: 412310019a20 ("drm/i915/icl: Add initial Icelake definitions.")
CC: stable@vger.kernel.org
This should not be needed, it was introduced in v4.17-rc1 only.
Reviewed-by: Joonas Lahtinen joonas.lahtinen@linux.intel.com
Regards, Joonas
Cc: Ingo Molnar mingo@kernel.org Cc: H. Peter Anvin hpa@zytor.com Cc: x86@kernel.org Cc: Daniele Ceraolo Spurio daniele.ceraolospurio@intel.com Cc: Joonas Lahtinen joonas.lahtinen@linux.intel.com Signed-off-by: Paulo Zanoni paulo.r.zanoni@intel.com
arch/x86/kernel/early-quirks.c | 18 ++++++++++++++++++ include/drm/i915_drm.h | 4 +++- 2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index bae0d32e327b..72c2cf961d44 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/arch/x86/kernel/early-quirks.c @@ -340,6 +340,18 @@ static resource_size_t __init gen3_stolen_base(int num, int slot, int func, return bsm & INTEL_BSM_MASK; } +static resource_size_t __init gen11_stolen_base(int num, int slot, int func,
resource_size_t
stolen_size) +{
u64 bsm;
bsm = read_pci_config(num, slot, func,
INTEL_GEN11_BSM_DW0);
bsm &= INTEL_BSM_MASK;
bsm |= (u64)read_pci_config(num, slot, func,
INTEL_GEN11_BSM_DW1) << 32;
return bsm;
+}
static resource_size_t __init i830_stolen_size(int num, int slot, int func) { u16 gmch_ctrl; @@ -500,6 +512,11 @@ static const struct intel_early_ops chv_early_ops __initconst = { .stolen_size = chv_stolen_size, }; +static const struct intel_early_ops gen11_early_ops __initconst = {
.stolen_base = gen11_stolen_base,
.stolen_size = gen9_stolen_size,
+};
static const struct pci_device_id intel_early_ids[] __initconst = { INTEL_I830_IDS(&i830_early_ops), INTEL_I845G_IDS(&i845_early_ops), @@ -531,6 +548,7 @@ static const struct pci_device_id intel_early_ids[] __initconst = { INTEL_CFL_IDS(&gen9_early_ops), INTEL_GLK_IDS(&gen9_early_ops), INTEL_CNL_IDS(&gen9_early_ops),
INTEL_ICL_11_IDS(&gen11_early_ops),
}; struct resource intel_graphics_stolen_res __ro_after_init = DEFINE_RES_MEM(0, 0); diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h index c9e5a6621b95..c44703f471b3 100644 --- a/include/drm/i915_drm.h +++ b/include/drm/i915_drm.h @@ -95,7 +95,9 @@ extern struct resource intel_graphics_stolen_res; #define I845_TSEG_SIZE_512K (2 << 1) #define I845_TSEG_SIZE_1M (3 << 1) -#define INTEL_BSM 0x5c +#define INTEL_BSM 0x5c +#define INTEL_GEN11_BSM_DW0 0xc0 +#define INTEL_GEN11_BSM_DW1 0xc4 #define INTEL_BSM_MASK (-(1u << 20))
#endif /* _I915_DRM_H_ */
2.14.3
On Fri, Jun 01, 2018 at 02:44:51PM -0700, Paulo Zanoni wrote:
Em Seg, 2018-05-07 às 11:46 +0300, Joonas Lahtinen escreveu:
Ingo, do you prefer to merge through our tree with your ack?
Ping?
Quoting Paulo Zanoni (2018-05-04 23:32:51)
ICL changes the registers and addresses to 64 bits.
I also briefly looked at implementing an u64 version of the PCI config read functions, but I concluded this wouldn't be trivial, so it's not worth doing it for a single user that can't have any racing problems while reading the register in two separate operations.
v2:
- Scrub the development (non-public) changelog (Joonas).
- Remove the i915.ko bits so this can be easily backported in
order to properly avoid stolen memory even on machines without i915.ko (Joonas).
- CC stable for the reasons above.
Issue: VIZ-9250
Fixes: 412310019a20 ("drm/i915/icl: Add initial Icelake definitions.")
CC: stable@vger.kernel.org
This should not be needed, it was introduced in v4.17-rc1 only.
Reviewed-by: Joonas Lahtinen joonas.lahtinen@linux.intel.com
Regards, Joonas
Cc: Thomas Gleixner tglx@linutronix.de
Cc: Ingo Molnar mingo@kernel.org Cc: H. Peter Anvin hpa@zytor.com Cc: x86@kernel.org
guys, could we push this through drm-intel? ack? nack? comments?
Cc: Daniele Ceraolo Spurio daniele.ceraolospurio@intel.com Cc: Joonas Lahtinen joonas.lahtinen@linux.intel.com Signed-off-by: Paulo Zanoni paulo.r.zanoni@intel.com
arch/x86/kernel/early-quirks.c | 18 ++++++++++++++++++ include/drm/i915_drm.h | 4 +++- 2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index bae0d32e327b..72c2cf961d44 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/arch/x86/kernel/early-quirks.c @@ -340,6 +340,18 @@ static resource_size_t __init gen3_stolen_base(int num, int slot, int func, return bsm & INTEL_BSM_MASK; } +static resource_size_t __init gen11_stolen_base(int num, int slot, int func,
resource_size_t
stolen_size) +{
u64 bsm;
bsm = read_pci_config(num, slot, func,
INTEL_GEN11_BSM_DW0);
bsm &= INTEL_BSM_MASK;
bsm |= (u64)read_pci_config(num, slot, func,
INTEL_GEN11_BSM_DW1) << 32;
return bsm;
+}
static resource_size_t __init i830_stolen_size(int num, int slot, int func) { u16 gmch_ctrl; @@ -500,6 +512,11 @@ static const struct intel_early_ops chv_early_ops __initconst = { .stolen_size = chv_stolen_size, }; +static const struct intel_early_ops gen11_early_ops __initconst = {
.stolen_base = gen11_stolen_base,
.stolen_size = gen9_stolen_size,
+};
static const struct pci_device_id intel_early_ids[] __initconst = { INTEL_I830_IDS(&i830_early_ops), INTEL_I845G_IDS(&i845_early_ops), @@ -531,6 +548,7 @@ static const struct pci_device_id intel_early_ids[] __initconst = { INTEL_CFL_IDS(&gen9_early_ops), INTEL_GLK_IDS(&gen9_early_ops), INTEL_CNL_IDS(&gen9_early_ops),
INTEL_ICL_11_IDS(&gen11_early_ops),
}; struct resource intel_graphics_stolen_res __ro_after_init = DEFINE_RES_MEM(0, 0); diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h index c9e5a6621b95..c44703f471b3 100644 --- a/include/drm/i915_drm.h +++ b/include/drm/i915_drm.h @@ -95,7 +95,9 @@ extern struct resource intel_graphics_stolen_res; #define I845_TSEG_SIZE_512K (2 << 1) #define I845_TSEG_SIZE_1M (3 << 1) -#define INTEL_BSM 0x5c +#define INTEL_BSM 0x5c +#define INTEL_GEN11_BSM_DW0 0xc0 +#define INTEL_GEN11_BSM_DW1 0xc4 #define INTEL_BSM_MASK (-(1u << 20))
#endif /* _I915_DRM_H_ */
2.14.3
Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx
On Mon, Jun 18, 2018 at 10:47:21AM -0700, Rodrigo Vivi wrote:
On Fri, Jun 01, 2018 at 02:44:51PM -0700, Paulo Zanoni wrote:
Em Seg, 2018-05-07 às 11:46 +0300, Joonas Lahtinen escreveu:
Ingo, do you prefer to merge through our tree with your ack?
Ping?
Quoting Paulo Zanoni (2018-05-04 23:32:51)
ICL changes the registers and addresses to 64 bits.
I also briefly looked at implementing an u64 version of the PCI config read functions, but I concluded this wouldn't be trivial, so it's not worth doing it for a single user that can't have any racing problems while reading the register in two separate operations.
v2:
- Scrub the development (non-public) changelog (Joonas).
- Remove the i915.ko bits so this can be easily backported in
order to properly avoid stolen memory even on machines without i915.ko (Joonas).
- CC stable for the reasons above.
Issue: VIZ-9250
Fixes: 412310019a20 ("drm/i915/icl: Add initial Icelake definitions.")
CC: stable@vger.kernel.org
This should not be needed, it was introduced in v4.17-rc1 only.
Reviewed-by: Joonas Lahtinen joonas.lahtinen@linux.intel.com
Regards, Joonas
Cc: Thomas Gleixner tglx@linutronix.de
Cc: Ingo Molnar mingo@kernel.org Cc: H. Peter Anvin hpa@zytor.com Cc: x86@kernel.org
guys, could we push this through drm-intel? ack? nack? comments?
Is there any concern with this patch? any ack for push this through drm-intel? or any nack with explanations please?
I'd like to push this for 4.19 because this is one of the patches that blocks us on using drm-tip on ICL.
Thanks in advance, Rodrigo.
Cc: Daniele Ceraolo Spurio daniele.ceraolospurio@intel.com Cc: Joonas Lahtinen joonas.lahtinen@linux.intel.com Signed-off-by: Paulo Zanoni paulo.r.zanoni@intel.com
arch/x86/kernel/early-quirks.c | 18 ++++++++++++++++++ include/drm/i915_drm.h | 4 +++- 2 files changed, 21 insertions(+), 1 deletion(-)
diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index bae0d32e327b..72c2cf961d44 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/arch/x86/kernel/early-quirks.c @@ -340,6 +340,18 @@ static resource_size_t __init gen3_stolen_base(int num, int slot, int func, return bsm & INTEL_BSM_MASK; } +static resource_size_t __init gen11_stolen_base(int num, int slot, int func,
resource_size_t
stolen_size) +{
u64 bsm;
bsm = read_pci_config(num, slot, func,
INTEL_GEN11_BSM_DW0);
bsm &= INTEL_BSM_MASK;
bsm |= (u64)read_pci_config(num, slot, func,
INTEL_GEN11_BSM_DW1) << 32;
return bsm;
+}
static resource_size_t __init i830_stolen_size(int num, int slot, int func) { u16 gmch_ctrl; @@ -500,6 +512,11 @@ static const struct intel_early_ops chv_early_ops __initconst = { .stolen_size = chv_stolen_size, }; +static const struct intel_early_ops gen11_early_ops __initconst = {
.stolen_base = gen11_stolen_base,
.stolen_size = gen9_stolen_size,
+};
static const struct pci_device_id intel_early_ids[] __initconst = { INTEL_I830_IDS(&i830_early_ops), INTEL_I845G_IDS(&i845_early_ops), @@ -531,6 +548,7 @@ static const struct pci_device_id intel_early_ids[] __initconst = { INTEL_CFL_IDS(&gen9_early_ops), INTEL_GLK_IDS(&gen9_early_ops), INTEL_CNL_IDS(&gen9_early_ops),
INTEL_ICL_11_IDS(&gen11_early_ops),
}; struct resource intel_graphics_stolen_res __ro_after_init = DEFINE_RES_MEM(0, 0); diff --git a/include/drm/i915_drm.h b/include/drm/i915_drm.h index c9e5a6621b95..c44703f471b3 100644 --- a/include/drm/i915_drm.h +++ b/include/drm/i915_drm.h @@ -95,7 +95,9 @@ extern struct resource intel_graphics_stolen_res; #define I845_TSEG_SIZE_512K (2 << 1) #define I845_TSEG_SIZE_1M (3 << 1) -#define INTEL_BSM 0x5c +#define INTEL_BSM 0x5c +#define INTEL_GEN11_BSM_DW0 0xc0 +#define INTEL_GEN11_BSM_DW1 0xc4 #define INTEL_BSM_MASK (-(1u << 20))
#endif /* _I915_DRM_H_ */
2.14.3
Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx
* Rodrigo Vivi rodrigo.vivi@intel.com wrote:
Cc: Thomas Gleixner tglx@linutronix.de
Cc: Ingo Molnar mingo@kernel.org Cc: H. Peter Anvin hpa@zytor.com Cc: x86@kernel.org
guys, could we push this through drm-intel? ack? nack? comments?
Is there any concern with this patch? any ack for push this through drm-intel? or any nack with explanations please?
I'd like to push this for 4.19 because this is one of the patches that blocks us on using drm-tip on ICL.
Sorry - no objections, and I suppose it will be all tested properly before going upstream:
Acked-by: Ingo Molnar mingo@kernel.org
Thanks,
Ingo
On Wed, Jul 04, 2018 at 07:50:42AM +0200, Ingo Molnar wrote:
- Rodrigo Vivi rodrigo.vivi@intel.com wrote:
Cc: Thomas Gleixner tglx@linutronix.de
Cc: Ingo Molnar mingo@kernel.org Cc: H. Peter Anvin hpa@zytor.com Cc: x86@kernel.org
guys, could we push this through drm-intel? ack? nack? comments?
Is there any concern with this patch? any ack for push this through drm-intel? or any nack with explanations please?
I'd like to push this for 4.19 because this is one of the patches that blocks us on using drm-tip on ICL.
Sorry - no objections, and I suppose it will be all tested properly before going upstream:
np. yeap, we have that running on ICLs here with internal branch for a while already.
Acked-by: Ingo Molnar mingo@kernel.org
Thanks, Rodrigo.
Thanks,
Ingo
On Wed, Jul 04, 2018 at 07:50:42AM +0200, Ingo Molnar wrote:
- Rodrigo Vivi rodrigo.vivi@intel.com wrote:
Cc: Thomas Gleixner tglx@linutronix.de
Cc: Ingo Molnar mingo@kernel.org Cc: H. Peter Anvin hpa@zytor.com Cc: x86@kernel.org
guys, could we push this through drm-intel? ack? nack? comments?
Is there any concern with this patch? any ack for push this through drm-intel? or any nack with explanations please?
I'd like to push this for 4.19 because this is one of the patches that blocks us on using drm-tip on ICL.
Sorry - no objections, and I suppose it will be all tested properly before going upstream:
Acked-by: Ingo Molnar mingo@kernel.org
series pushed to drm-intel-next-queued... Probably landing on 4.19 still
Thanks for patches, reviews and acks, specially x86 one.
Thanks,
Ingo _______________________________________________ Intel-gfx mailing list Intel-gfx@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/intel-gfx
linux-stable-mirror@lists.linaro.org