GICv3: Emulate GICD_IGRPMODR as RAZ / WI
PATCH v3 1/13 GICv3: Emulate GICD_IGRPMODR as RAZ / WI
Refer GIC v3 specification (Arm IHI 0069H ID020922), IGRPMODR is emulated as RAZ / WI for the guests as "GICD_CTLR.ARE_S==0" is true.
Xen is currently supported to run in non-secure mode, so guests will run in non-secure mode only.
Also, if Xen was supposed to run in secure mode with guests, the programming of the interrupts (ie whether it belongs to secure/non secure and group 0/1) will be done by Xen only. The guests will not be allowed to change this.
---
Observed the issue while running Zephyr on R52.
Also, found that KVM has similar behaviour.
Changes from:-
v1 - Moved the definitions of GICD_IGRPMODR, GICD_IGRPMODRN to gic_v3 specific header.
v2 - Updated the commit message.
xen/arch/arm/include/asm/gic_v3_defs.h | 2 ++ xen/arch/arm/vgic-v3.c | 4 ++++ 2 files changed, 6 insertions(+)
xen/arch/arm/include/asm/gic_v3_defs.h +2 -0
@@ -30,6 +30,8 @@#define GICD_CLRSPI_NSR (0x048)#define GICD_SETSPI_SR (0x050)#define GICD_CLRSPI_SR (0x058) +#define GICD_IGRPMODR (0xD00) +#define GICD_IGRPMODRN (0xD7C)#define GICD_IROUTER (0x6000)#define GICD_IROUTER32 (0x6100)#define GICD_IROUTER1019 (0x7FD8)
xen/arch/arm/vgic-v3.c +4 -0
@@ -685,6 +685,7 @@ static int __vgic_v3_distr_common_mmio_read(const char *name, struct vcpu *v,switch ( reg ) {case VRANGE32 (GICD_IGROUPR ,GICD_IGROUPRN ): +case VRANGE32 (GICD_IGRPMODR ,GICD_IGRPMODRN ):/* We do not implement security extensions for guests, read zero */ if ( dabt.size !=DABT_WORD ) goto bad_width; goto read_as_zero;
@@ -781,6 +782,7 @@ static int __vgic_v3_distr_common_mmio_write(const char *name, struct vcpu *v,switch ( reg ) {case VRANGE32 (GICD_IGROUPR ,GICD_IGROUPRN ): +case VRANGE32 (GICD_IGRPMODR ,GICD_IGRPMODRN ):/* We do not implement security extensions for guests, write ignore */ goto write_ignore_32;
@@ -1192,6 +1194,7 @@ static int vgic_v3_distr_mmio_read(struct vcpu *v, mmio_info_t *info,case VRANGE32 (GICD_ICACTIVER ,GICD_ICACTIVERN ):case VRANGE32 (GICD_IPRIORITYR ,GICD_IPRIORITYRN ):case VRANGE32 (GICD_ICFGR ,GICD_ICFGRN ): +case VRANGE32 (GICD_IGRPMODR ,GICD_IGRPMODRN ):/* * Above all register are common with GICR and GICD * Manage in common
@@ -1379,6 +1382,7 @@ static int vgic_v3_distr_mmio_write(struct vcpu *v, mmio_info_t *info,case VRANGE32 (GICD_ICACTIVER ,GICD_ICACTIVERN ):case VRANGE32 (GICD_IPRIORITYR ,GICD_IPRIORITYRN ):case VRANGE32 (GICD_ICFGR ,GICD_ICFGRN ): +case VRANGE32 (GICD_IGRPMODR ,GICD_IGRPMODRN ):/* Above registers are common with GICR and GICD * Manage in common */ return __vgic_v3_distr_common_mmio_write("vGICD", v, info,
Signature · 1 line
2.17.1
Hi Ayan,
Quoted 10 lines · Ayan Kumar Halder
Refer GIC v3 specification (Arm IHI 0069H ID020922), IGRPMODR is emulated as RAZ / WI for the guests as "GICD_CTLR.ARE_S==0" is true.
Xen is currently supported to run in non-secure mode, so guests will run in non-secure mode only.Also, if Xen was supposed to run in secure mode with guests, the programming of the interrupts (ie whether it belongs to secure/non secure and group 0/1) will be done by Xen only. The guests will not be allowed to change this.
Thanks a lot this description is a lot better.
Quoted 2 lines
Signed-off-by Ayan Kumar Halder <[email protected]>
I do not think this is something for 4.17 so it will need to be on hold until staging reopens unless someone thinks otherwise ?
Cheers
Bertrand
Quoted 67 lines
---
Observed the issue while running Zephyr on R52.
Also, found that KVM has similar behaviour.Changes from:-
v1 - Moved the definitions ofGICD_IGRPMODR,GICD_IGRPMODRNtogic_v3specific header.v2 - Updated the commit message.
xen/arch/arm/include/asm/gic_v3_defs.h| 2 ++xen/arch/arm/vgic-v3.c| 4 ++++2 files changed, 6 insertions(+)xen/arch/arm/include/asm/gic_v3_defs.h +2
-0@@ -30,6 +30,8 @@#define GICD_CLRSPI_NSR (0x048)#define GICD_SETSPI_SR (0x050)#define GICD_CLRSPI_SR (0x058) +#define GICD_IGRPMODR (0xD00) +#define GICD_IGRPMODRN (0xD7C)#define GICD_IROUTER (0x6000)#define GICD_IROUTER32 (0x6100)#define GICD_IROUTER1019 (0x7FD8)xen/arch/arm/vgic-v3.c +4
-0@@ -685,6 +685,7 @@ static int __vgic_v3_distr_common_mmio_read(const char *name, struct vcpu *v,switch ( reg ) {case VRANGE32 (GICD_IGROUPR ,GICD_IGROUPRN ): +case VRANGE32 (GICD_IGRPMODR ,GICD_IGRPMODRN ):/* We do not implement security extensions for guests, read zero */ if ( dabt.size !=DABT_WORD ) goto bad_width; goto read_as_zero;@@ -781,6 +782,7 @@ static int __vgic_v3_distr_common_mmio_write(const char *name, struct vcpu *v,switch ( reg ) {case VRANGE32 (GICD_IGROUPR ,GICD_IGROUPRN ): +case VRANGE32 (GICD_IGRPMODR ,GICD_IGRPMODRN ):/* We do not implement security extensions for guests, write ignore */ goto write_ignore_32;@@ -1192,6 +1194,7 @@ static int vgic_v3_distr_mmio_read(struct vcpu *v, mmio_info_t *info,case VRANGE32 (GICD_ICACTIVER ,GICD_ICACTIVERN ):case VRANGE32 (GICD_IPRIORITYR ,GICD_IPRIORITYRN ):case VRANGE32 (GICD_ICFGR ,GICD_ICFGRN ): +case VRANGE32 (GICD_IGRPMODR ,GICD_IGRPMODRN ):/* * Above all register are common with GICR and GICD * Manage in common @@ -1379,6 +1382,7 @@ static int vgic_v3_distr_mmio_write(struct vcpu *v, mmio_info_t *info,case VRANGE32 (GICD_ICACTIVER ,GICD_ICACTIVERN ):case VRANGE32 (GICD_IPRIORITYR ,GICD_IPRIORITYRN ):case VRANGE32 (GICD_ICFGR ,GICD_ICFGRN ): +case VRANGE32 (GICD_IGRPMODR ,GICD_IGRPMODRN ):/* Above registers are common with GICR and GICD * Manage in common */ return __vgic_v3_distr_common_mmio_write("vGICD", v, info,Signature · 1 line
2.17.1
Hi,
Quoted 11 lines · Bertrand Marquis
Quoted 10 lines · Ayan Kumar Halder
Refer GIC v3 specification (Arm IHI 0069H ID020922), IGRPMODR is emulated as RAZ / WI for the guests as "GICD_CTLR.ARE_S==0" is true.
Xen is currently supported to run in non-secure mode, so guests will run in non-secure mode only.Also, if Xen was supposed to run in secure mode with guests, the programming of the interrupts (ie whether it belongs to secure/non secure and group 0/1) will be done by Xen only. The guests will not be allowed to change this.
Quoted 7 lines
Thanks a lot this description is a lot better.
Quoted 2 lines
Signed-off-by Ayan Kumar Halder <[email protected]>Reviewed-by Bertrand Marquis <[email protected]>I do not think this is something for 4.17 so it will need to be on hold until staging reopens unless someone thinks otherwise ?
I have pushed the patch to my branch for-next/4.18. I will merge it once the tree is re-opened.
Cheers,
Signature · 1 line
Julien Grall