* [PATCH 0/8] x86/platform/UV: UV Update PatchSet 2
@ 2017-01-13 15:21 Mike Travis
2017-01-13 15:21 ` [PATCH 1/8] x86/platform/UV: Fix panic with missing UVsystab support Mike Travis
` (7 more replies)
0 siblings, 8 replies; 15+ messages in thread
From: Mike Travis @ 2017-01-13 15:21 UTC (permalink / raw)
To: Ingo Molnar, Thomas Gleixner, H. Peter Anvin; +Cc: x86, linux-kernel
Second UV Update PatchSet containing:
* Fix the panic where KEXEC'd kernel does not have access to
EFI runtime mappings.
* Fix the panic with 2 socket configs.
* Recognition and initialization of UV Hubless systems.
* Addition of UV Hubless support for RMC system NMI command.
* Addition of a basic CPU health check using NMI.
* Verify new NMI action is valid.
* Default NMI action standardized on "dump" (stack/regs).
* Move check for "is UV system" from native_smp_prepare_cpus() to
uv_system_init() since it needs to now check for hubless as well.
--
^ permalink raw reply [flat|nested] 15+ messages in thread* [PATCH 1/8] x86/platform/UV: Fix panic with missing UVsystab support 2017-01-13 15:21 [PATCH 0/8] x86/platform/UV: UV Update PatchSet 2 Mike Travis @ 2017-01-13 15:21 ` Mike Travis 2017-01-14 9:57 ` [tip:x86/platform] " tip-bot for Mike Travis 2017-01-13 15:21 ` [PATCH 2/8] x86/platform/UV: Fix 2 socket config problem Mike Travis ` (6 subsequent siblings) 7 siblings, 1 reply; 15+ messages in thread From: Mike Travis @ 2017-01-13 15:21 UTC (permalink / raw) To: Ingo Molnar, Thomas Gleixner, H. Peter Anvin Cc: x86, linux-kernel, Dimitri Sivanich [-- Attachment #1: uv4_update_get_uvsystab --] [-- Type: text/plain, Size: 2035 bytes --] Fix the panic where KEXEC'd kernel does not have access to EFI runtime mappings. This may cause the extended UVsystab to not be available. The solution is to revert to non-UV mode and continue with limited capabilities. Signed-off-by: Mike Travis <travis@sgi.com> Acked-by: Dimitri Sivanich <sivanich@hpe.com> Reviewed-by: Russ Anderson <rja@hpe.com> Reviewed-by: Alex Thorlton <athorlton@sgi.com> --- arch/x86/kernel/apic/x2apic_uv_x.c | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-) --- linux.orig/arch/x86/kernel/apic/x2apic_uv_x.c +++ linux/arch/x86/kernel/apic/x2apic_uv_x.c @@ -1172,19 +1172,25 @@ static void __init decode_gam_rng_tbl(un index, _min_socket, _max_socket, _min_pnode, _max_pnode); } -static void __init decode_uv_systab(void) +static int __init decode_uv_systab(void) { struct uv_systab *st; int i; + if (uv_hub_info->hub_revision < UV4_HUB_REVISION_BASE) + return 0; /* No extended UVsystab required */ + st = uv_systab; - if ((!st || st->revision < UV_SYSTAB_VERSION_UV4) && !is_uv4_hub()) - return; - if (st->revision != UV_SYSTAB_VERSION_UV4_LATEST) { - pr_crit( + if ((!st) || (st->revision < UV_SYSTAB_VERSION_UV4_LATEST)) { + int rev = st ? st->revision : 0; + + pr_err( "UV: BIOS UVsystab version(%x) mismatch, expecting(%x)\n", - st->revision, UV_SYSTAB_VERSION_UV4_LATEST); - BUG(); + rev, UV_SYSTAB_VERSION_UV4_LATEST); + pr_err( + "UV: Cannot support UV operations, switching to generic PC\n"); + uv_system_type = UV_NONE; + return -EINVAL; } for (i = 0; st->entry[i].type != UV_SYSTAB_TYPE_UNUSED; i++) { @@ -1205,6 +1211,7 @@ static void __init decode_uv_systab(void break; } } + return 0; } /* @@ -1373,7 +1380,8 @@ void __init uv_system_init(void) map_low_mmrs(); uv_bios_init(); /* get uv_systab for decoding */ - decode_uv_systab(); + if (decode_uv_systab() < 0) + return; /* UVsystab problem, abort UV init */ build_socket_tables(); build_uv_gr_table(); uv_init_hub_info(&hub_info); -- ^ permalink raw reply [flat|nested] 15+ messages in thread
* [tip:x86/platform] x86/platform/UV: Fix panic with missing UVsystab support 2017-01-13 15:21 ` [PATCH 1/8] x86/platform/UV: Fix panic with missing UVsystab support Mike Travis @ 2017-01-14 9:57 ` tip-bot for Mike Travis 0 siblings, 0 replies; 15+ messages in thread From: tip-bot for Mike Travis @ 2017-01-14 9:57 UTC (permalink / raw) To: linux-tip-commits Cc: athorlton, mingo, torvalds, travis, rja, sivanich, peterz, hpa, linux-kernel, tglx Commit-ID: eee5715efd8c268724b14c956de6af5d4931f470 Gitweb: http://git.kernel.org/tip/eee5715efd8c268724b14c956de6af5d4931f470 Author: Mike Travis <travis@sgi.com> AuthorDate: Fri, 13 Jan 2017 09:21:11 -0600 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Sat, 14 Jan 2017 09:26:35 +0100 x86/platform/UV: Fix panic with missing UVsystab support Fix the panic where KEXEC'd kernel does not have access to EFI runtime mappings. This may cause the extended UVsystab to not be available. The solution is to revert to non-UV mode and continue with limited capabilities. Signed-off-by: Mike Travis <travis@sgi.com> Reviewed-by: Russ Anderson <rja@hpe.com> Reviewed-by: Alex Thorlton <athorlton@sgi.com> Acked-by: Dimitri Sivanich <sivanich@hpe.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Link: http://lkml.kernel.org/r/20170113152111.118886202@asylum.americas.sgi.com Signed-off-by: Ingo Molnar <mingo@kernel.org> --- arch/x86/kernel/apic/x2apic_uv_x.c | 24 ++++++++++++++++-------- 1 file changed, 16 insertions(+), 8 deletions(-) diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index 35690a1..4393078 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -1172,19 +1172,25 @@ static void __init decode_gam_rng_tbl(unsigned long ptr) index, _min_socket, _max_socket, _min_pnode, _max_pnode); } -static void __init decode_uv_systab(void) +static int __init decode_uv_systab(void) { struct uv_systab *st; int i; + if (uv_hub_info->hub_revision < UV4_HUB_REVISION_BASE) + return 0; /* No extended UVsystab required */ + st = uv_systab; - if ((!st || st->revision < UV_SYSTAB_VERSION_UV4) && !is_uv4_hub()) - return; - if (st->revision != UV_SYSTAB_VERSION_UV4_LATEST) { - pr_crit( + if ((!st) || (st->revision < UV_SYSTAB_VERSION_UV4_LATEST)) { + int rev = st ? st->revision : 0; + + pr_err( "UV: BIOS UVsystab version(%x) mismatch, expecting(%x)\n", - st->revision, UV_SYSTAB_VERSION_UV4_LATEST); - BUG(); + rev, UV_SYSTAB_VERSION_UV4_LATEST); + pr_err( + "UV: Cannot support UV operations, switching to generic PC\n"); + uv_system_type = UV_NONE; + return -EINVAL; } for (i = 0; st->entry[i].type != UV_SYSTAB_TYPE_UNUSED; i++) { @@ -1205,6 +1211,7 @@ static void __init decode_uv_systab(void) break; } } + return 0; } /* @@ -1373,7 +1380,8 @@ void __init uv_system_init(void) map_low_mmrs(); uv_bios_init(); /* get uv_systab for decoding */ - decode_uv_systab(); + if (decode_uv_systab() < 0) + return; /* UVsystab problem, abort UV init */ build_socket_tables(); build_uv_gr_table(); uv_init_hub_info(&hub_info); ^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH 2/8] x86/platform/UV: Fix 2 socket config problem 2017-01-13 15:21 [PATCH 0/8] x86/platform/UV: UV Update PatchSet 2 Mike Travis 2017-01-13 15:21 ` [PATCH 1/8] x86/platform/UV: Fix panic with missing UVsystab support Mike Travis @ 2017-01-13 15:21 ` Mike Travis 2017-01-14 9:58 ` [tip:x86/platform] " tip-bot for Mike Travis 2017-01-13 15:21 ` [PATCH 3/8] x86/platform/UV: Add Support for UV4 Hubless systems Mike Travis ` (5 subsequent siblings) 7 siblings, 1 reply; 15+ messages in thread From: Mike Travis @ 2017-01-13 15:21 UTC (permalink / raw) To: Ingo Molnar, Thomas Gleixner, H. Peter Anvin Cc: x86, linux-kernel, Dimitri Sivanich [-- Attachment #1: uv4_update_gnode_extra --] [-- Type: text/plain, Size: 1445 bytes --] A UV4 chassis with only 2 sockets configured can unexpectedly target the wrong UV hub. Fix the problem by limiting the minimum size of a partition to 4 sockets even if only 2 are configured. Signed-off-by: Mike Travis <travis@sgi.com> Acked-by: Dimitri Sivanich <sivanich@hpe.com> Reviewed-by: Russ Anderson <rja@hpe.com> --- arch/x86/kernel/apic/x2apic_uv_x.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) --- linux.orig/arch/x86/kernel/apic/x2apic_uv_x.c +++ linux/arch/x86/kernel/apic/x2apic_uv_x.c @@ -56,6 +56,7 @@ static struct { unsigned int socketid_shift; /* aka pnode_shift for UV1/2/3 */ unsigned int pnode_mask; unsigned int gpa_shift; + unsigned int gnode_shift; } uv_cpuid; int uv_min_hub_revision_id; @@ -133,6 +134,7 @@ static int __init early_get_pnodeid(void break; case UV4_HUB_PART_NUMBER: uv_min_hub_revision_id += UV4_HUB_REVISION_BASE - 1; + uv_cpuid.gnode_shift = 2; /* min partition is 4 sockets */ break; } @@ -1074,8 +1076,10 @@ void __init uv_init_hub_info(struct uv_h (1UL << uv_cpuid.gpa_shift) - 1; node_id.v = uv_read_local_mmr(UVH_NODE_ID); + uv_cpuid.gnode_shift = max_t(unsigned int, + uv_cpuid.gnode_shift, mn.n_val); hub_info->gnode_extra = - (node_id.s.node_id & ~((1 << mn.n_val) - 1)) >> 1; + (node_id.s.node_id & ~((1 << uv_cpuid.gnode_shift) - 1)) >> 1; hub_info->gnode_upper = ((unsigned long)hub_info->gnode_extra << mn.m_val); -- ^ permalink raw reply [flat|nested] 15+ messages in thread
* [tip:x86/platform] x86/platform/UV: Fix 2 socket config problem 2017-01-13 15:21 ` [PATCH 2/8] x86/platform/UV: Fix 2 socket config problem Mike Travis @ 2017-01-14 9:58 ` tip-bot for Mike Travis 0 siblings, 0 replies; 15+ messages in thread From: tip-bot for Mike Travis @ 2017-01-14 9:58 UTC (permalink / raw) To: linux-tip-commits Cc: linux-kernel, hpa, peterz, torvalds, rja, tglx, sivanich, mingo, travis Commit-ID: 81a71176740624ef3b1bff50c51e7b4aa187353d Gitweb: http://git.kernel.org/tip/81a71176740624ef3b1bff50c51e7b4aa187353d Author: Mike Travis <travis@sgi.com> AuthorDate: Fri, 13 Jan 2017 09:21:12 -0600 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Sat, 14 Jan 2017 09:26:35 +0100 x86/platform/UV: Fix 2 socket config problem A UV4 chassis with only 2 sockets configured can unexpectedly target the wrong UV hub. Fix the problem by limiting the minimum size of a partition to 4 sockets even if only 2 are configured. Signed-off-by: Mike Travis <travis@sgi.com> Reviewed-by: Russ Anderson <rja@hpe.com> Acked-by: Dimitri Sivanich <sivanich@hpe.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Link: http://lkml.kernel.org/r/20170113152111.313888353@asylum.americas.sgi.com Signed-off-by: Ingo Molnar <mingo@kernel.org> --- arch/x86/kernel/apic/x2apic_uv_x.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index 4393078..97ea712 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -56,6 +56,7 @@ static struct { unsigned int socketid_shift; /* aka pnode_shift for UV1/2/3 */ unsigned int pnode_mask; unsigned int gpa_shift; + unsigned int gnode_shift; } uv_cpuid; int uv_min_hub_revision_id; @@ -133,6 +134,7 @@ static int __init early_get_pnodeid(void) break; case UV4_HUB_PART_NUMBER: uv_min_hub_revision_id += UV4_HUB_REVISION_BASE - 1; + uv_cpuid.gnode_shift = 2; /* min partition is 4 sockets */ break; } @@ -1074,8 +1076,10 @@ void __init uv_init_hub_info(struct uv_hub_info_s *hub_info) (1UL << uv_cpuid.gpa_shift) - 1; node_id.v = uv_read_local_mmr(UVH_NODE_ID); + uv_cpuid.gnode_shift = max_t(unsigned int, + uv_cpuid.gnode_shift, mn.n_val); hub_info->gnode_extra = - (node_id.s.node_id & ~((1 << mn.n_val) - 1)) >> 1; + (node_id.s.node_id & ~((1 << uv_cpuid.gnode_shift) - 1)) >> 1; hub_info->gnode_upper = ((unsigned long)hub_info->gnode_extra << mn.m_val); ^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH 3/8] x86/platform/UV: Add Support for UV4 Hubless systems 2017-01-13 15:21 [PATCH 0/8] x86/platform/UV: UV Update PatchSet 2 Mike Travis 2017-01-13 15:21 ` [PATCH 1/8] x86/platform/UV: Fix panic with missing UVsystab support Mike Travis 2017-01-13 15:21 ` [PATCH 2/8] x86/platform/UV: Fix 2 socket config problem Mike Travis @ 2017-01-13 15:21 ` Mike Travis 2017-01-13 15:21 ` [PATCH 4/8] x86/platform/UV: Add Support for UV4 Hubless NMIs Mike Travis ` (4 subsequent siblings) 7 siblings, 0 replies; 15+ messages in thread From: Mike Travis @ 2017-01-13 15:21 UTC (permalink / raw) To: Ingo Molnar, Thomas Gleixner, H. Peter Anvin Cc: x86, linux-kernel, Dimitri Sivanich [-- Attachment #1: uv4_add-hubless-support --] [-- Type: text/plain, Size: 2955 bytes --] Add recognition and support for UV4 hubless systems. Signed-off-by: Mike Travis <travis@sgi.com> Acked-by: Dimitri Sivanich <sivanich@hpe.com> Reviewed-by: Russ Anderson <rja@hpe.com> --- arch/x86/include/asm/uv/uv.h | 2 ++ arch/x86/kernel/apic/x2apic_uv_x.c | 33 +++++++++++++++++++++++++++++++-- 2 files changed, 33 insertions(+), 2 deletions(-) --- linux.orig/arch/x86/include/asm/uv/uv.h +++ linux/arch/x86/include/asm/uv/uv.h @@ -10,6 +10,7 @@ struct mm_struct; extern enum uv_system_type get_uv_system_type(void); extern int is_uv_system(void); +extern int is_uv_hubless(void); extern void uv_cpu_init(void); extern void uv_nmi_init(void); extern void uv_system_init(void); @@ -23,6 +24,7 @@ extern const struct cpumask *uv_flush_tl static inline enum uv_system_type get_uv_system_type(void) { return UV_NONE; } static inline int is_uv_system(void) { return 0; } +static inline int is_uv_hubless(void) { return 0; } static inline void uv_cpu_init(void) { } static inline void uv_system_init(void) { } static inline const struct cpumask * --- linux.orig/arch/x86/kernel/apic/x2apic_uv_x.c +++ linux/arch/x86/kernel/apic/x2apic_uv_x.c @@ -44,6 +44,7 @@ DEFINE_PER_CPU(int, x2apic_extra_bits); #define PR_DEVEL(fmt, args...) pr_devel("%s: " fmt, __func__, args) static enum uv_system_type uv_system_type; +static bool uv_hubless_system; static u64 gru_start_paddr, gru_end_paddr; static u64 gru_dist_base, gru_first_node_paddr = -1LL, gru_last_node_paddr; static u64 gru_dist_lmask, gru_dist_umask; @@ -222,8 +223,14 @@ static int __init uv_acpi_madt_oem_check int pnodeid; int uv_apic; - if (strncmp(oem_id, "SGI", 3) != 0) + if (strncmp(oem_id, "SGI", 3) != 0) { + if (strncmp(oem_id, "NSGI", 4) == 0) { + uv_hubless_system = true; + pr_info("UV: OEM IDs %s/%s, HUBLESS\n", + oem_id, oem_table_id); + } return 0; + } if (numa_off) { pr_err("UV: NUMA is off, disabling UV support\n"); @@ -296,6 +303,12 @@ int is_uv_system(void) } EXPORT_SYMBOL_GPL(is_uv_system); +int is_uv_hubless(void) +{ + return uv_hubless_system; +} +EXPORT_SYMBOL_GPL(is_uv_hubless); + void **__uv_hub_info_list; EXPORT_SYMBOL_GPL(__uv_hub_info_list); @@ -1365,7 +1378,7 @@ static void __init build_socket_tables(v } } -void __init uv_system_init(void) +static void __init uv_system_init_hub(void) { struct uv_hub_info_s hub_info = {0}; int bytes, cpu, nodeid; @@ -1502,4 +1515,20 @@ void __init uv_system_init(void) reboot_type = BOOT_ACPI; } +/* + * There is a small amount of UV specific code needed to initialize a + * UV system that does not have a "UV HUB" (referred to as "hubless"). + */ +void __init uv_system_init(void) +{ + if (likely(!is_uv_system() && !is_uv_hubless())) + return; + + /* Far more common is a standard UV system (with a HUB) */ + if (is_uv_system()) + uv_system_init_hub(); + + /* Initialize UV hubless system here */ +} + apic_driver(apic_x2apic_uv_x); -- ^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH 4/8] x86/platform/UV: Add Support for UV4 Hubless NMIs 2017-01-13 15:21 [PATCH 0/8] x86/platform/UV: UV Update PatchSet 2 Mike Travis ` (2 preceding siblings ...) 2017-01-13 15:21 ` [PATCH 3/8] x86/platform/UV: Add Support for UV4 Hubless systems Mike Travis @ 2017-01-13 15:21 ` Mike Travis 2017-01-14 7:52 ` Ingo Molnar 2017-01-13 15:21 ` [PATCH 5/8] x86/platform/UV: Add basic CPU NMI health check Mike Travis ` (3 subsequent siblings) 7 siblings, 1 reply; 15+ messages in thread From: Mike Travis @ 2017-01-13 15:21 UTC (permalink / raw) To: Ingo Molnar, Thomas Gleixner, H. Peter Anvin Cc: x86, linux-kernel, Dimitri Sivanich [-- Attachment #1: uv4_add-hubless-nmi --] [-- Type: text/plain, Size: 10088 bytes --] Merge new UV Hubless NMI support into existing UV NMI handler. Signed-off-by: Mike Travis <travis@sgi.com> Acked-by: Dimitri Sivanich <sivanich@hpe.com> Reviewed-by: Russ Anderson <rja@hpe.com> --- arch/x86/include/asm/uv/uv_hub.h | 3 arch/x86/kernel/apic/x2apic_uv_x.c | 2 arch/x86/platform/uv/uv_nmi.c | 185 ++++++++++++++++++++++++++++++++++--- 3 files changed, 177 insertions(+), 13 deletions(-) --- linux.orig/arch/x86/include/asm/uv/uv_hub.h +++ linux/arch/x86/include/asm/uv/uv_hub.h @@ -772,6 +772,7 @@ static inline int uv_num_possible_blades /* Per Hub NMI support */ extern void uv_nmi_setup(void); +extern void uv_nmi_setup_hubless(void); /* BMC sets a bit this MMR non-zero before sending an NMI */ #define UVH_NMI_MMR UVH_SCRATCH5 @@ -799,6 +800,8 @@ struct uv_hub_nmi_s { atomic_t read_mmr_count; /* count of MMR reads */ atomic_t nmi_count; /* count of true UV NMIs */ unsigned long nmi_value; /* last value read from NMI MMR */ + bool hub_present; /* false means UV hubless system */ + bool pch_owner; /* indicates this hub owns PCH */ }; struct uv_cpu_nmi_s { --- linux.orig/arch/x86/kernel/apic/x2apic_uv_x.c +++ linux/arch/x86/kernel/apic/x2apic_uv_x.c @@ -1529,6 +1529,8 @@ void __init uv_system_init(void) uv_system_init_hub(); /* Initialize UV hubless system here */ + else + uv_nmi_setup_hubless(); } apic_driver(apic_x2apic_uv_x); --- linux.orig/arch/x86/platform/uv/uv_nmi.c +++ linux/arch/x86/platform/uv/uv_nmi.c @@ -67,6 +67,19 @@ static struct uv_hub_nmi_s **uv_hub_nmi_ DEFINE_PER_CPU(struct uv_cpu_nmi_s, uv_cpu_nmi); EXPORT_PER_CPU_SYMBOL_GPL(uv_cpu_nmi); +/* UV hubless values */ +static int *pch_base; +#define NMI_CONTROL_PORT 0x70 +#define NMI_DUMMY_PORT 0x71 +#define GPI_NMI_STS_GPP_D_0 0x164 +#define GPI_NMI_ENA_GPP_D_0 0x174 +#define STS_GPP_D_0_MASK 0x1 +#define PAD_CFG_DW0_GPP_D_0 0x4c0 +#define GPIROUTNMI (1ul << 17) +#define PCH_PCR_GPIO_1_BASE 0xfdae0000ul +#define PCH_PCR_GPIO_ADDRESS(offset) \ + (int *)((unsigned long long)pch_base | offset) + static unsigned long nmi_mmr; static unsigned long nmi_mmr_clear; static unsigned long nmi_mmr_pending; @@ -144,6 +157,19 @@ module_param_named(wait_count, uv_nmi_wa static int uv_nmi_retry_count = 500; module_param_named(retry_count, uv_nmi_retry_count, int, 0644); +static bool uv_pch_intr_enable = true; +static bool uv_pch_intr_now_enabled; +module_param_named(pch_intr_enable, uv_pch_intr_enable, bool, 0644); + +static int uv_nmi_debug; +module_param_named(debug, uv_nmi_debug, int, 0644); + +#define nmi_debug(fmt, ...) \ + do { \ + if (uv_nmi_debug) \ + pr_info(fmt, ##__VA_ARGS__); \ + } while (0) + /* * Valid NMI Actions: * "dump" - dump process stack for each cpu @@ -192,6 +218,78 @@ static inline void uv_local_mmr_clear_nm } /* + * UV hubless NMI handler functions + */ +static inline void uv_reassert_nmi(void) +{ + /* (from arch/x86/include/asm/mach_traps.h) */ + outb(0x8f, NMI_CONTROL_PORT); + inb(NMI_DUMMY_PORT); /* dummy read */ + outb(0x0f, NMI_CONTROL_PORT); + inb(NMI_DUMMY_PORT); /* dummy read */ +} + +static void uv_init_hubless_pch_io(int offset, int mask, int data) +{ + int *addr = PCH_PCR_GPIO_ADDRESS(offset); + int readd = readl(addr); + + if (mask) { /* OR in new data */ + int writed = (readd & ~mask) | data; + + nmi_debug("UV:PCH: %p = %x & %x | %x (%x)\n", + addr, readd, ~mask, data, writed); + writel(writed, addr); + + } else if (readd & data) { /* clear status bit */ + nmi_debug("UV:PCH: %p = %x\n", addr, data); + writel(data, addr); + } + + (void)readl(addr); /* flush write data */ +} + +static void uv_nmi_setup_hubless_intr(void) +{ + uv_pch_intr_now_enabled = uv_pch_intr_enable; + + uv_init_hubless_pch_io( + PAD_CFG_DW0_GPP_D_0, GPIROUTNMI, + uv_pch_intr_now_enabled ? GPIROUTNMI : 0); + + nmi_debug("UV:NMI: GPP_D_0 interrupt %s\n", + uv_pch_intr_now_enabled ? "enabled" : "disabled"); +} + +static int uv_nmi_test_hubless(struct uv_hub_nmi_s *hub_nmi) +{ + int *pstat = PCH_PCR_GPIO_ADDRESS(GPI_NMI_STS_GPP_D_0); + int status = *pstat; + + hub_nmi->nmi_value = status; + atomic_inc(&hub_nmi->read_mmr_count); + + if (!(status & STS_GPP_D_0_MASK)) /* NOT UV external NMI */ + return 0; + + *pstat = STS_GPP_D_0_MASK; /* IS UV NMI, clear GPP_D_0 status */ + (void)*pstat; /* flush write */ + + return 1; +} + +static int uv_test_nmi(struct uv_hub_nmi_s *hub_nmi) +{ + if (hub_nmi->hub_present) + return uv_nmi_test_mmr(hub_nmi); + + if (hub_nmi->pch_owner) /* Only PCH owner can check status */ + return uv_nmi_test_hubless(hub_nmi); + + return -1; +} + +/* * If first cpu in on this hub, set hub_nmi "in_nmi" and "owner" values and * return true. If first cpu in on the system, set global "in_nmi" flag. */ @@ -214,6 +312,7 @@ static int uv_check_nmi(struct uv_hub_nm { int cpu = smp_processor_id(); int nmi = 0; + int nmi_detected = 0; local64_inc(&uv_nmi_count); this_cpu_inc(uv_cpu_nmi.queries); @@ -224,20 +323,26 @@ static int uv_check_nmi(struct uv_hub_nm break; if (raw_spin_trylock(&hub_nmi->nmi_lock)) { + nmi_detected = uv_test_nmi(hub_nmi); - /* check hub MMR NMI flag */ - if (uv_nmi_test_mmr(hub_nmi)) { + /* check flag for UV external NMI */ + if (nmi_detected > 0) { uv_set_in_nmi(cpu, hub_nmi); nmi = 1; break; } - /* MMR NMI flag is clear */ + /* A non-PCH node in a hubless system waits for NMI */ + else if (nmi_detected < 0) + goto slave_wait; + + /* MMR/PCH NMI flag is clear */ raw_spin_unlock(&hub_nmi->nmi_lock); } else { + /* wait a moment for the hub nmi locker to set flag */ - cpu_relax(); +slave_wait: cpu_relax(); udelay(uv_nmi_slave_delay); /* re-check hub in_nmi flag */ @@ -246,13 +351,20 @@ static int uv_check_nmi(struct uv_hub_nm break; } - /* check if this BMC missed setting the MMR NMI flag */ + /* + * check if this BMC missed setting the MMR NMI flag (or) + * UV hubless system where only PCH owner can check flag + */ if (!nmi) { nmi = atomic_read(&uv_in_nmi); if (nmi) uv_set_in_nmi(cpu, hub_nmi); } + /* if we're holding the hub lock, release it now */ + if (nmi_detected < 0) + raw_spin_unlock(&hub_nmi->nmi_lock); + } while (0); if (!nmi) @@ -269,7 +381,10 @@ static inline void uv_clear_nmi(int cpu) if (cpu == atomic_read(&hub_nmi->cpu_owner)) { atomic_set(&hub_nmi->cpu_owner, -1); atomic_set(&hub_nmi->in_nmi, 0); - uv_local_mmr_clear_nmi(); + if (hub_nmi->hub_present) + uv_local_mmr_clear_nmi(); + else + uv_reassert_nmi(); raw_spin_unlock(&hub_nmi->nmi_lock); } } @@ -302,6 +417,7 @@ static int uv_nmi_wait_cpus(int first) { int i, j, k, n = num_online_cpus(); int last_k = 0, waiting = 0; + int cpu = smp_processor_id(); if (first) { cpumask_copy(uv_nmi_cpu_mask, cpu_online_mask); @@ -310,6 +426,12 @@ static int uv_nmi_wait_cpus(int first) k = n - cpumask_weight(uv_nmi_cpu_mask); } + /* PCH NMI causes only one cpu to respond */ + if (first && uv_pch_intr_now_enabled) { + cpumask_clear_cpu(cpu, uv_nmi_cpu_mask); + return n - k - 1; + } + udelay(uv_nmi_initial_delay); for (i = 0; i < uv_nmi_retry_count; i++) { int loop_delay = uv_nmi_loop_delay; @@ -358,7 +480,7 @@ static void uv_nmi_wait(int master) break; /* if not all made it in, send IPI NMI to them */ - pr_alert("UV: Sending NMI IPI to %d non-responding CPUs: %*pbl\n", + pr_alert("UV: Sending NMI IPI to %d CPUs: %*pbl\n", cpumask_weight(uv_nmi_cpu_mask), cpumask_pr_args(uv_nmi_cpu_mask)); @@ -634,6 +756,13 @@ int uv_handle_nmi(unsigned int reason, s else if (uv_nmi_action_is("kdb") || uv_nmi_action_is("kgdb")) uv_call_kgdb_kdb(cpu, regs, master); + /* Unknown NMI action */ + else { + if (master) + pr_alert("UV: unknown NMI action: %s\n", uv_nmi_action); + uv_nmi_sync_exit(master); + } + /* Clear per_cpu "in nmi" flag */ this_cpu_write(uv_cpu_nmi.state, UV_NMI_STATE_OUT); @@ -648,6 +777,7 @@ int uv_handle_nmi(unsigned int reason, s atomic_set(&uv_nmi_cpu, -1); atomic_set(&uv_in_nmi, 0); atomic_set(&uv_nmi_kexec_failed, 0); + atomic_set(&uv_nmi_slave_continue, SLAVE_CLEAR); } uv_nmi_touch_watchdogs(); @@ -697,28 +827,57 @@ void uv_nmi_init(void) apic_write(APIC_LVT1, value); } -void uv_nmi_setup(void) +/* Setup hub nmi info */ +void __init uv_nmi_setup_common(bool hubbed) { int size = sizeof(void *) * (1 << NODES_SHIFT); - int cpu, nid; + int cpu; - /* Setup hub nmi info */ - uv_nmi_setup_mmrs(); uv_hub_nmi_list = kzalloc(size, GFP_KERNEL); - pr_info("UV: NMI hub list @ 0x%p (%d)\n", uv_hub_nmi_list, size); + nmi_debug("UV: NMI hub list @ 0x%p (%d)\n", uv_hub_nmi_list, size); BUG_ON(!uv_hub_nmi_list); size = sizeof(struct uv_hub_nmi_s); for_each_present_cpu(cpu) { - nid = cpu_to_node(cpu); + int nid = cpu_to_node(cpu); if (uv_hub_nmi_list[nid] == NULL) { uv_hub_nmi_list[nid] = kzalloc_node(size, GFP_KERNEL, nid); BUG_ON(!uv_hub_nmi_list[nid]); raw_spin_lock_init(&(uv_hub_nmi_list[nid]->nmi_lock)); atomic_set(&uv_hub_nmi_list[nid]->cpu_owner, -1); + uv_hub_nmi_list[nid]->hub_present = hubbed; + uv_hub_nmi_list[nid]->pch_owner = (nid == 0); } uv_hub_nmi_per(cpu) = uv_hub_nmi_list[nid]; } BUG_ON(!alloc_cpumask_var(&uv_nmi_cpu_mask, GFP_KERNEL)); +} + +/* setup for UV Hub systems */ +void __init uv_nmi_setup(void) +{ + uv_nmi_setup_mmrs(); + uv_nmi_setup_common(true); + uv_register_nmi_notifier(); + pr_info("UV: Hub NMI enabled\n"); +} + +/* setup for UV Hubless systems */ +void __init uv_nmi_setup_hubless(void) +{ + uv_nmi_setup_common(false); + + pch_base = xlate_dev_mem_ptr(PCH_PCR_GPIO_1_BASE); + nmi_debug("UV: PCH base:%p from 0x%lx, GPP_D_0\n", + pch_base, PCH_PCR_GPIO_1_BASE); + + uv_init_hubless_pch_io(GPI_NMI_ENA_GPP_D_0, + STS_GPP_D_0_MASK, STS_GPP_D_0_MASK); + + uv_nmi_setup_hubless_intr(); + + uv_reassert_nmi(); /* insure NMI enabled in proc inf reg */ + uv_register_nmi_notifier(); + pr_info("UV: Hubless NMI enabled\n"); } -- ^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: [PATCH 4/8] x86/platform/UV: Add Support for UV4 Hubless NMIs 2017-01-13 15:21 ` [PATCH 4/8] x86/platform/UV: Add Support for UV4 Hubless NMIs Mike Travis @ 2017-01-14 7:52 ` Ingo Molnar 2017-01-14 8:26 ` [PATCH] x86/platform/UV: Clean up the UV APIC code Ingo Molnar 0 siblings, 1 reply; 15+ messages in thread From: Ingo Molnar @ 2017-01-14 7:52 UTC (permalink / raw) To: Mike Travis Cc: Ingo Molnar, Thomas Gleixner, H. Peter Anvin, x86, linux-kernel, Dimitri Sivanich * Mike Travis <travis@sgi.com> wrote: > --- linux.orig/arch/x86/kernel/apic/x2apic_uv_x.c > +++ linux/arch/x86/kernel/apic/x2apic_uv_x.c > @@ -1529,6 +1529,8 @@ void __init uv_system_init(void) > uv_system_init_hub(); > > /* Initialize UV hubless system here */ > + else > + uv_nmi_setup_hubless(); That's not the proper coding style ... > +++ linux/arch/x86/platform/uv/uv_nmi.c > @@ -67,6 +67,19 @@ static struct uv_hub_nmi_s **uv_hub_nmi_ > DEFINE_PER_CPU(struct uv_cpu_nmi_s, uv_cpu_nmi); > EXPORT_PER_CPU_SYMBOL_GPL(uv_cpu_nmi); > > +/* UV hubless values */ > +static int *pch_base; > +#define NMI_CONTROL_PORT 0x70 Ditto... > +#define NMI_DUMMY_PORT 0x71 > +#define GPI_NMI_STS_GPP_D_0 0x164 > +#define GPI_NMI_ENA_GPP_D_0 0x174 > +#define STS_GPP_D_0_MASK 0x1 > +#define PAD_CFG_DW0_GPP_D_0 0x4c0 > +#define GPIROUTNMI (1ul << 17) > +#define PCH_PCR_GPIO_1_BASE 0xfdae0000ul > +#define PCH_PCR_GPIO_ADDRESS(offset) \ > + (int *)((unsigned long long)pch_base | offset) > +static void uv_init_hubless_pch_io(int offset, int mask, int data) > +{ > + int *addr = PCH_PCR_GPIO_ADDRESS(offset); > + int readd = readl(addr); > + > + if (mask) { /* OR in new data */ > + int writed = (readd & ~mask) | data; > + > + nmi_debug("UV:PCH: %p = %x & %x | %x (%x)\n", > + addr, readd, ~mask, data, writed); No unnecessary linebreaks please. > + > /* wait a moment for the hub nmi locker to set flag */ > - cpu_relax(); Please capitalize 'NMI' consistently in all the patches. > + /* PCH NMI causes only one cpu to respond */ Same for 'CPU'. > else if (uv_nmi_action_is("kdb") || uv_nmi_action_is("kgdb")) > uv_call_kgdb_kdb(cpu, regs, master); > > + /* Unknown NMI action */ > + else { Sigh ... > +/* setup for UV Hubless systems */ > +void __init uv_nmi_setup_hubless(void) > +{ > + uv_nmi_setup_common(false); > + > + pch_base = xlate_dev_mem_ptr(PCH_PCR_GPIO_1_BASE); > + nmi_debug("UV: PCH base:%p from 0x%lx, GPP_D_0\n", > + pch_base, PCH_PCR_GPIO_1_BASE); > + > + uv_init_hubless_pch_io(GPI_NMI_ENA_GPP_D_0, > + STS_GPP_D_0_MASK, STS_GPP_D_0_MASK); > + > + uv_nmi_setup_hubless_intr(); > + > + uv_reassert_nmi(); /* insure NMI enabled in proc inf reg */ And typos ... Much cleaner patches please! I haven't checked the others in the series but please review them for similar mishaps. Thanks, Ingo ^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH] x86/platform/UV: Clean up the UV APIC code 2017-01-14 7:52 ` Ingo Molnar @ 2017-01-14 8:26 ` Ingo Molnar 2017-01-14 13:11 ` Joe Perches 2017-02-01 10:09 ` [tip:x86/platform] " tip-bot for Ingo Molnar 0 siblings, 2 replies; 15+ messages in thread From: Ingo Molnar @ 2017-01-14 8:26 UTC (permalink / raw) To: Mike Travis Cc: Ingo Molnar, Thomas Gleixner, H. Peter Anvin, x86, linux-kernel, Dimitri Sivanich * Ingo Molnar <mingo@kernel.org> wrote: > > + uv_reassert_nmi(); /* insure NMI enabled in proc inf reg */ > > And typos ... > > Much cleaner patches please! I haven't checked the others in the series but please > review them for similar mishaps. So I had a look at the latest arch/x86/kernel/apic/x2apic_uv_x.c and looks like these cleanliness problems are widespread. I applied the first two patches which were (mostly...) fine, but after that the code needs some major love. Below is an attempt to clean it up some. Only build tested. In particular the random line breaks, apparently to pacify checkpatch, were horribly widespread. I just made the lines longer - that's still more readable than the multi-line horror that was there before. But there were a lot of other style problems as well, all around the place. If this patch works on real hardware as well then please integrate it into your patchset and base the remaining patches on top of it. Thanks, Ingo ==================> Subject: x86/platform/UV: Clean up the UV APIC code From: Ingo Molnar <mingo@kernel.org> Date: Sat Jan 14 09:22:02 CET 2017 Signed-off-by: Ingo Molnar <mingo@kernel.org> --- arch/x86/kernel/apic/x2apic_uv_x.c | 500 ++++++++++++++++++------------------- 1 file changed, 244 insertions(+), 256 deletions(-) Index: tip/arch/x86/kernel/apic/x2apic_uv_x.c =================================================================== --- tip.orig/arch/x86/kernel/apic/x2apic_uv_x.c +++ tip/arch/x86/kernel/apic/x2apic_uv_x.c @@ -41,15 +41,13 @@ DEFINE_PER_CPU(int, x2apic_extra_bits); -#define PR_DEVEL(fmt, args...) pr_devel("%s: " fmt, __func__, args) +static enum uv_system_type uv_system_type; +static u64 gru_start_paddr, gru_end_paddr; +static u64 gru_dist_base, gru_first_node_paddr = -1LL, gru_last_node_paddr; +static u64 gru_dist_lmask, gru_dist_umask; +static union uvh_apicid uvh_apicid; -static enum uv_system_type uv_system_type; -static u64 gru_start_paddr, gru_end_paddr; -static u64 gru_dist_base, gru_first_node_paddr = -1LL, gru_last_node_paddr; -static u64 gru_dist_lmask, gru_dist_umask; -static union uvh_apicid uvh_apicid; - -/* info derived from CPUID */ +/* Information derived from CPUID: */ static struct { unsigned int apicid_shift; unsigned int apicid_mask; @@ -61,21 +59,25 @@ static struct { int uv_min_hub_revision_id; EXPORT_SYMBOL_GPL(uv_min_hub_revision_id); + unsigned int uv_apicid_hibits; EXPORT_SYMBOL_GPL(uv_apicid_hibits); static struct apic apic_x2apic_uv_x; static struct uv_hub_info_s uv_hub_info_node0; -/* Set this to use hardware error handler instead of kernel panic */ +/* Set this to use hardware error handler instead of kernel panic: */ static int disable_uv_undefined_panic = 1; + unsigned long uv_undefined(char *str) { if (likely(!disable_uv_undefined_panic)) panic("UV: error: undefined MMR: %s\n", str); else pr_crit("UV: error: undefined MMR: %s\n", str); - return ~0ul; /* cause a machine fault */ + + /* Cause a machine fault: */ + return ~0ul; } EXPORT_SYMBOL(uv_undefined); @@ -86,18 +88,19 @@ static unsigned long __init uv_early_rea mmr = early_ioremap(UV_LOCAL_MMR_BASE | addr, sizeof(*mmr)); val = *mmr; early_iounmap(mmr, sizeof(*mmr)); + return val; } static inline bool is_GRU_range(u64 start, u64 end) { if (gru_dist_base) { - u64 su = start & gru_dist_umask; /* upper (incl pnode) bits */ - u64 sl = start & gru_dist_lmask; /* base offset bits */ + u64 su = start & gru_dist_umask; /* Upper (incl pnode) bits */ + u64 sl = start & gru_dist_lmask; /* Base offset bits */ u64 eu = end & gru_dist_umask; u64 el = end & gru_dist_lmask; - /* Must reside completely within a single GRU range */ + /* Must reside completely within a single GRU range: */ return (sl == gru_dist_base && el == gru_dist_base && su >= gru_first_node_paddr && su <= gru_last_node_paddr && @@ -141,7 +144,7 @@ static int __init early_get_pnodeid(void uv_hub_info->hub_revision = uv_min_hub_revision_id; uv_cpuid.pnode_mask = (1 << m_n_config.s.n_skt) - 1; pnode = (node_id.s.node_id >> 1) & uv_cpuid.pnode_mask; - uv_cpuid.gpa_shift = 46; /* default unless changed */ + uv_cpuid.gpa_shift = 46; /* Default unless changed */ pr_info("UV: rev:%d part#:%x nodeid:%04x n_skt:%d pnmsk:%x pn:%x\n", node_id.s.revision, node_id.s.part_number, node_id.s.node_id, @@ -149,11 +152,12 @@ static int __init early_get_pnodeid(void return pnode; } -/* [copied from arch/x86/kernel/cpu/topology.c:detect_extended_topology()] */ -#define SMT_LEVEL 0 /* leaf 0xb SMT level */ -#define INVALID_TYPE 0 /* leaf 0xb sub-leaf types */ -#define SMT_TYPE 1 -#define CORE_TYPE 2 +/* [Copied from arch/x86/kernel/cpu/topology.c:detect_extended_topology()] */ + +#define SMT_LEVEL 0 /* Leaf 0xb SMT level */ +#define INVALID_TYPE 0 /* Leaf 0xb sub-leaf types */ +#define SMT_TYPE 1 +#define CORE_TYPE 2 #define LEAFB_SUBTYPE(ecx) (((ecx) >> 8) & 0xff) #define BITS_SHIFT_NEXT_LEVEL(eax) ((eax) & 0x1f) @@ -167,11 +171,13 @@ static void set_x2apic_bits(void) pr_info("UV: CPU does not have CPUID.11\n"); return; } + cpuid_count(0xb, SMT_LEVEL, &eax, &ebx, &ecx, &edx); if (ebx == 0 || (LEAFB_SUBTYPE(ecx) != SMT_TYPE)) { pr_info("UV: CPUID.11 not implemented\n"); return; } + sid_shift = BITS_SHIFT_NEXT_LEVEL(eax); sub_index = 1; do { @@ -182,8 +188,9 @@ static void set_x2apic_bits(void) } sub_index++; } while (LEAFB_SUBTYPE(ecx) != INVALID_TYPE); - uv_cpuid.apicid_shift = 0; - uv_cpuid.apicid_mask = (~(-1 << sid_shift)); + + uv_cpuid.apicid_shift = 0; + uv_cpuid.apicid_mask = (~(-1 << sid_shift)); uv_cpuid.socketid_shift = sid_shift; } @@ -194,10 +201,8 @@ static void __init early_get_apic_socket set_x2apic_bits(); - pr_info("UV: apicid_shift:%d apicid_mask:0x%x\n", - uv_cpuid.apicid_shift, uv_cpuid.apicid_mask); - pr_info("UV: socketid_shift:%d pnode_mask:0x%x\n", - uv_cpuid.socketid_shift, uv_cpuid.pnode_mask); + pr_info("UV: apicid_shift:%d apicid_mask:0x%x\n", uv_cpuid.apicid_shift, uv_cpuid.apicid_mask); + pr_info("UV: socketid_shift:%d pnode_mask:0x%x\n", uv_cpuid.socketid_shift, uv_cpuid.pnode_mask); } /* @@ -210,10 +215,8 @@ static void __init uv_set_apicid_hibit(v union uv1h_lb_target_physical_apic_id_mask_u apicid_mask; if (is_uv1_hub()) { - apicid_mask.v = - uv_early_read_mmr(UV1H_LB_TARGET_PHYSICAL_APIC_ID_MASK); - uv_apicid_hibits = - apicid_mask.s1.bit_enables & UV_APICID_HIBIT_MASK; + apicid_mask.v = uv_early_read_mmr(UV1H_LB_TARGET_PHYSICAL_APIC_ID_MASK); + uv_apicid_hibits = apicid_mask.s1.bit_enables & UV_APICID_HIBIT_MASK; } } @@ -230,12 +233,12 @@ static int __init uv_acpi_madt_oem_check return 0; } - /* Setup early hub type field in uv_hub_info for Node 0 */ + /* Set up early hub type field in uv_hub_info for Node 0 */ uv_cpu_info->p_uv_hub_info = &uv_hub_info_node0; /* * Determine UV arch type. - * SGI: UV100/1000 + * SGI: UV100/1000 * SGI2: UV2000/3000 * SGI3: UV300 (truncated to 4 chars because of different varieties) * SGI4: UV400 (truncated to 4 chars because of different varieties) @@ -251,31 +254,32 @@ static int __init uv_acpi_madt_oem_check pnodeid = early_get_pnodeid(); early_get_apic_socketid_shift(); - x86_platform.is_untracked_pat_range = uv_is_untracked_pat_range; + + x86_platform.is_untracked_pat_range = uv_is_untracked_pat_range; x86_platform.nmi_init = uv_nmi_init; - if (!strcmp(oem_table_id, "UVX")) { /* most common */ + if (!strcmp(oem_table_id, "UVX")) { + /* This is the most common hardware variant: */ uv_system_type = UV_X2APIC; uv_apic = 0; - } else if (!strcmp(oem_table_id, "UVH")) { /* only UV1 systems */ + } else if (!strcmp(oem_table_id, "UVH")) { + /* Only UV1 systems: */ uv_system_type = UV_NON_UNIQUE_APIC; - __this_cpu_write(x2apic_extra_bits, - pnodeid << uvh_apicid.s.pnode_shift); + __this_cpu_write(x2apic_extra_bits, pnodeid << uvh_apicid.s.pnode_shift); uv_set_apicid_hibit(); uv_apic = 1; - } else if (!strcmp(oem_table_id, "UVL")) { /* only used for */ - uv_system_type = UV_LEGACY_APIC; /* very small systems */ + } else if (!strcmp(oem_table_id, "UVL")) { + /* Only used for very small systems: */ + uv_system_type = UV_LEGACY_APIC; uv_apic = 0; } else { goto badbios; } - pr_info("UV: OEM IDs %s/%s, System/HUB Types %d/%d, uv_apic %d\n", - oem_id, oem_table_id, uv_system_type, - uv_min_hub_revision_id, uv_apic); + pr_info("UV: OEM IDs %s/%s, System/HUB Types %d/%d, uv_apic %d\n", oem_id, oem_table_id, uv_system_type, uv_min_hub_revision_id, uv_apic); return uv_apic; @@ -308,16 +312,18 @@ EXPORT_SYMBOL_GPL(uv_possible_blades); unsigned long sn_rtc_cycles_per_second; EXPORT_SYMBOL(sn_rtc_cycles_per_second); -/* the following values are used for the per node hub info struct */ -static __initdata unsigned short *_node_to_pnode; -static __initdata unsigned short _min_socket, _max_socket; -static __initdata unsigned short _min_pnode, _max_pnode, _gr_table_len; -static __initdata struct uv_gam_range_entry *uv_gre_table; -static __initdata struct uv_gam_parameters *uv_gp_table; -static __initdata unsigned short *_socket_to_node; -static __initdata unsigned short *_socket_to_pnode; -static __initdata unsigned short *_pnode_to_socket; -static __initdata struct uv_gam_range_s *_gr_table; +/* The following values are used for the per node hub info struct */ +static __initdata unsigned short *_node_to_pnode; +static __initdata unsigned short _min_socket, _max_socket; +static __initdata unsigned short _min_pnode, _max_pnode, _gr_table_len; +static __initdata struct uv_gam_range_entry *uv_gre_table; +static __initdata struct uv_gam_parameters *uv_gp_table; +static __initdata unsigned short *_socket_to_node; +static __initdata unsigned short *_socket_to_pnode; +static __initdata unsigned short *_pnode_to_socket; + +static __initdata struct uv_gam_range_s *_gr_table; + #define SOCK_EMPTY ((unsigned short)~0) extern int uv_hub_info_version(void) @@ -326,7 +332,7 @@ extern int uv_hub_info_version(void) } EXPORT_SYMBOL(uv_hub_info_version); -/* Build GAM range lookup table */ +/* Build GAM range lookup table: */ static __init void build_uv_gr_table(void) { struct uv_gam_range_entry *gre = uv_gre_table; @@ -344,25 +350,24 @@ static __init void build_uv_gr_table(voi for (; gre->type != UV_GAM_RANGE_TYPE_UNUSED; gre++) { if (gre->type == UV_GAM_RANGE_TYPE_HOLE) { - if (!ram_limit) { /* mark hole between ram/non-ram */ + if (!ram_limit) { + /* Mark hole between RAM/non-RAM: */ ram_limit = last_limit; last_limit = gre->limit; lsid++; continue; } last_limit = gre->limit; - pr_info("UV: extra hole in GAM RE table @%d\n", - (int)(gre - uv_gre_table)); + pr_info("UV: extra hole in GAM RE table @%d\n", (int)(gre - uv_gre_table)); continue; } if (_max_socket < gre->sockid) { - pr_err("UV: GAM table sockid(%d) too large(>%d) @%d\n", - gre->sockid, _max_socket, - (int)(gre - uv_gre_table)); + pr_err("UV: GAM table sockid(%d) too large(>%d) @%d\n", gre->sockid, _max_socket, (int)(gre - uv_gre_table)); continue; } sid = gre->sockid - _min_socket; - if (lsid < sid) { /* new range */ + if (lsid < sid) { + /* New range: */ grt = &_gr_table[indx]; grt->base = lindx; grt->nasid = gre->nasid; @@ -371,27 +376,32 @@ static __init void build_uv_gr_table(voi lindx = indx++; continue; } - if (lsid == sid && !ram_limit) { /* update range */ - if (grt->limit == last_limit) { /* .. if contiguous */ + /* Update range: */ + if (lsid == sid && !ram_limit) { + /* .. if contiguous: */ + if (grt->limit == last_limit) { grt->limit = last_limit = gre->limit; continue; } } - if (!ram_limit) { /* non-contiguous ram range */ + /* Non-contiguous RAM range: */ + if (!ram_limit) { grt++; grt->base = lindx; grt->nasid = gre->nasid; grt->limit = last_limit = gre->limit; continue; } - grt++; /* non-contiguous/non-ram */ - grt->base = grt - _gr_table; /* base is this entry */ + /* Non-contiguous/non-RAM: */ + grt++; + /* base is this entry */ + grt->base = grt - _gr_table; grt->nasid = gre->nasid; grt->limit = last_limit = gre->limit; lsid++; } - /* shorten table if possible */ + /* Shorten table if possible */ grt++; i = grt - _gr_table; if (i < _gr_table_len) { @@ -405,16 +415,15 @@ static __init void build_uv_gr_table(voi } } - /* display resultant gam range table */ + /* Display resultant GAM range table: */ for (i = 0, grt = _gr_table; i < _gr_table_len; i++, grt++) { + unsigned long start, end; int gb = grt->base; - unsigned long start = gb < 0 ? 0 : - (unsigned long)_gr_table[gb].limit << UV_GAM_RANGE_SHFT; - unsigned long end = - (unsigned long)grt->limit << UV_GAM_RANGE_SHFT; - pr_info("UV: GAM Range %2d %04x 0x%013lx-0x%013lx (%d)\n", - i, grt->nasid, start, end, gb); + start = gb < 0 ? 0 : (unsigned long)_gr_table[gb].limit << UV_GAM_RANGE_SHFT; + end = (unsigned long)grt->limit << UV_GAM_RANGE_SHFT; + + pr_info("UV: GAM Range %2d %04x 0x%013lx-0x%013lx (%d)\n", i, grt->nasid, start, end, gb); } } @@ -425,16 +434,19 @@ static int uv_wakeup_secondary(int phys_ pnode = uv_apicid_to_pnode(phys_apicid); phys_apicid |= uv_apicid_hibits; + val = (1UL << UVH_IPI_INT_SEND_SHFT) | (phys_apicid << UVH_IPI_INT_APIC_ID_SHFT) | ((start_rip << UVH_IPI_INT_VECTOR_SHFT) >> 12) | APIC_DM_INIT; + uv_write_global_mmr64(pnode, UVH_IPI_INT, val); val = (1UL << UVH_IPI_INT_SEND_SHFT) | (phys_apicid << UVH_IPI_INT_APIC_ID_SHFT) | ((start_rip << UVH_IPI_INT_VECTOR_SHFT) >> 12) | APIC_DM_STARTUP; + uv_write_global_mmr64(pnode, UVH_IPI_INT, val); return 0; @@ -568,7 +580,7 @@ static struct apic apic_x2apic_uv_x __ro .apic_id_registered = uv_apic_id_registered, .irq_delivery_mode = dest_Fixed, - .irq_dest_mode = 0, /* physical */ + .irq_dest_mode = 0, /* Physical */ .target_cpus = online_target_cpus, .disable_esr = 0, @@ -629,23 +641,22 @@ static __init void get_lowmem_redirect(u switch (i) { case 0: m_redirect = UVH_RH_GAM_ALIAS210_REDIRECT_CONFIG_0_MMR; - m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_0_MMR; + m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_0_MMR; break; case 1: m_redirect = UVH_RH_GAM_ALIAS210_REDIRECT_CONFIG_1_MMR; - m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_1_MMR; + m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_1_MMR; break; case 2: m_redirect = UVH_RH_GAM_ALIAS210_REDIRECT_CONFIG_2_MMR; - m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_2_MMR; + m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_2_MMR; break; } alias.v = uv_read_local_mmr(m_overlay); if (alias.s.enable && alias.s.base == 0) { *size = (1UL << alias.s.m_alias); redirect.v = uv_read_local_mmr(m_redirect); - *base = (unsigned long)redirect.s.dest_base - << DEST_SHIFT; + *base = (unsigned long)redirect.s.dest_base << DEST_SHIFT; return; } } @@ -654,8 +665,7 @@ static __init void get_lowmem_redirect(u enum map_type {map_wb, map_uc}; -static __init void map_high(char *id, unsigned long base, int pshift, - int bshift, int max_pnode, enum map_type map_type) +static __init void map_high(char *id, unsigned long base, int pshift, int bshift, int max_pnode, enum map_type map_type) { unsigned long bytes, paddr; @@ -680,16 +690,19 @@ static __init void map_gru_distributed(u int nid; gru.v = c; - /* only base bits 42:28 relevant in dist mode */ + + /* Only base bits 42:28 relevant in dist mode */ gru_dist_base = gru.v & 0x000007fff0000000UL; if (!gru_dist_base) { pr_info("UV: Map GRU_DIST base address NULL\n"); return; } + bytes = 1UL << UVH_RH_GAM_GRU_OVERLAY_CONFIG_MMR_BASE_SHFT; gru_dist_lmask = ((1UL << uv_hub_info->m_val) - 1) & ~(bytes - 1); gru_dist_umask = ~((1UL << uv_hub_info->m_val) - 1); gru_dist_base &= gru_dist_lmask; /* Clear bits above M */ + for_each_online_node(nid) { paddr = ((u64)uv_node_to_pnode(nid) << uv_hub_info->m_val) | gru_dist_base; @@ -697,11 +710,12 @@ static __init void map_gru_distributed(u gru_first_node_paddr = min(paddr, gru_first_node_paddr); gru_last_node_paddr = max(paddr, gru_last_node_paddr); } + /* Save upper (63:M) bits of address only for is_GRU_range */ gru_first_node_paddr &= gru_dist_umask; gru_last_node_paddr &= gru_dist_umask; - pr_debug("UV: Map GRU_DIST base 0x%016llx 0x%016llx - 0x%016llx\n", - gru_dist_base, gru_first_node_paddr, gru_last_node_paddr); + + pr_debug("UV: Map GRU_DIST base 0x%016llx 0x%016llx - 0x%016llx\n", gru_dist_base, gru_first_node_paddr, gru_last_node_paddr); } static __init void map_gru_high(int max_pnode) @@ -721,6 +735,7 @@ static __init void map_gru_high(int max_ map_gru_distributed(gru.v); return; } + base = (gru.v & mask) >> shift; map_high("GRU", base, shift, shift, max_pnode, map_wb); gru_start_paddr = ((u64)base << shift); @@ -774,8 +789,8 @@ static __init void map_mmioh_high_uv3(in id = mmiohs[index].id; overlay.v = uv_read_local_mmr(mmiohs[index].overlay); - pr_info("UV: %s overlay 0x%lx base:0x%x m_io:%d\n", - id, overlay.v, overlay.s3.base, overlay.s3.m_io); + + pr_info("UV: %s overlay 0x%lx base:0x%x m_io:%d\n", id, overlay.v, overlay.s3.base, overlay.s3.m_io); if (!overlay.s3.enable) { pr_info("UV: %s disabled\n", id); return; @@ -786,7 +801,8 @@ static __init void map_mmioh_high_uv3(in m_io = overlay.s3.m_io; mmr = mmiohs[index].redirect; n = UV3H_RH_GAM_MMIOH_REDIRECT_CONFIG0_MMR_DEPTH; - min_pnode *= 2; /* convert to NASID */ + /* Convert to NASID: */ + min_pnode *= 2; max_pnode *= 2; max_io = lnasid = fi = li = -1; @@ -795,16 +811,18 @@ static __init void map_mmioh_high_uv3(in redirect.v = uv_read_local_mmr(mmr + i * 8); nasid = redirect.s3.nasid; + /* Invalid NASID: */ if (nasid < min_pnode || max_pnode < nasid) - nasid = -1; /* invalid NASID */ + nasid = -1; if (nasid == lnasid) { li = i; - if (i != n-1) /* last entry check */ + /* Last entry check: */ + if (i != n-1) continue; } - /* check if we have a cached (or last) redirect to print */ + /* Check if we have a cached (or last) redirect to print: */ if (lnasid != -1 || (i == n-1 && nasid != -1)) { unsigned long addr1, addr2; int f, l; @@ -816,12 +834,9 @@ static __init void map_mmioh_high_uv3(in f = fi; l = li; } - addr1 = (base << shift) + - f * (1ULL << m_io); - addr2 = (base << shift) + - (l + 1) * (1ULL << m_io); - pr_info("UV: %s[%03d..%03d] NASID 0x%04x ADDR 0x%016lx - 0x%016lx\n", - id, fi, li, lnasid, addr1, addr2); + addr1 = (base << shift) + f * (1ULL << m_io); + addr2 = (base << shift) + (l + 1) * (1ULL << m_io); + pr_info("UV: %s[%03d..%03d] NASID 0x%04x ADDR 0x%016lx - 0x%016lx\n", id, fi, li, lnasid, addr1, addr2); if (max_io < l) max_io = l; } @@ -829,8 +844,7 @@ static __init void map_mmioh_high_uv3(in lnasid = nasid; } - pr_info("UV: %s base:0x%lx shift:%d M_IO:%d MAX_IO:%d\n", - id, base, shift, m_io, max_io); + pr_info("UV: %s base:0x%lx shift:%d M_IO:%d MAX_IO:%d\n", id, base, shift, m_io, max_io); if (max_io >= 0) map_high(id, base, shift, m_io, max_io, map_uc); @@ -843,36 +857,35 @@ static __init void map_mmioh_high(int mi int shift, enable, m_io, n_io; if (is_uv3_hub() || is_uv4_hub()) { - /* Map both MMIOH Regions */ + /* Map both MMIOH regions: */ map_mmioh_high_uv3(0, min_pnode, max_pnode); map_mmioh_high_uv3(1, min_pnode, max_pnode); return; } if (is_uv1_hub()) { - mmr = UV1H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR; - shift = UV1H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR_BASE_SHFT; - mmioh.v = uv_read_local_mmr(mmr); - enable = !!mmioh.s1.enable; - base = mmioh.s1.base; - m_io = mmioh.s1.m_io; - n_io = mmioh.s1.n_io; + mmr = UV1H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR; + shift = UV1H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR_BASE_SHFT; + mmioh.v = uv_read_local_mmr(mmr); + enable = !!mmioh.s1.enable; + base = mmioh.s1.base; + m_io = mmioh.s1.m_io; + n_io = mmioh.s1.n_io; } else if (is_uv2_hub()) { - mmr = UV2H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR; - shift = UV2H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR_BASE_SHFT; - mmioh.v = uv_read_local_mmr(mmr); - enable = !!mmioh.s2.enable; - base = mmioh.s2.base; - m_io = mmioh.s2.m_io; - n_io = mmioh.s2.n_io; - } else + mmr = UV2H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR; + shift = UV2H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR_BASE_SHFT; + mmioh.v = uv_read_local_mmr(mmr); + enable = !!mmioh.s2.enable; + base = mmioh.s2.base; + m_io = mmioh.s2.m_io; + n_io = mmioh.s2.n_io; + } else { return; + } if (enable) { max_pnode &= (1 << n_io) - 1; - pr_info( - "UV: base:0x%lx shift:%d N_IO:%d M_IO:%d max_pnode:0x%x\n", - base, shift, m_io, n_io, max_pnode); + pr_info("UV: base:0x%lx shift:%d N_IO:%d M_IO:%d max_pnode:0x%x\n", base, shift, m_io, n_io, max_pnode); map_high("MMIOH", base, shift, m_io, max_pnode, map_uc); } else { pr_info("UV: MMIOH disabled\n"); @@ -890,16 +903,16 @@ static __init void uv_rtc_init(void) long status; u64 ticks_per_sec; - status = uv_bios_freq_base(BIOS_FREQ_BASE_REALTIME_CLOCK, - &ticks_per_sec); + status = uv_bios_freq_base(BIOS_FREQ_BASE_REALTIME_CLOCK, &ticks_per_sec); + if (status != BIOS_STATUS_SUCCESS || ticks_per_sec < 100000) { - printk(KERN_WARNING - "unable to determine platform RTC clock frequency, " - "guessing.\n"); - /* BIOS gives wrong value for clock freq. so guess */ + pr_warn("UV: unable to determine platform RTC clock frequency, guessing.\n"); + + /* BIOS gives wrong value for clock frequency, so guess: */ sn_rtc_cycles_per_second = 1000000000000UL / 30000UL; - } else + } else { sn_rtc_cycles_per_second = ticks_per_sec; + } } /* @@ -910,19 +923,19 @@ static void uv_heartbeat(unsigned long i struct timer_list *timer = &uv_scir_info->timer; unsigned char bits = uv_scir_info->state; - /* flip heartbeat bit */ + /* Flip heartbeat bit: */ bits ^= SCIR_CPU_HEARTBEAT; - /* is this cpu idle? */ + /* Is this CPU idle? */ if (idle_cpu(raw_smp_processor_id())) bits &= ~SCIR_CPU_ACTIVITY; else bits |= SCIR_CPU_ACTIVITY; - /* update system controller interface reg */ + /* Update system controller interface reg: */ uv_set_scir_bits(bits); - /* enable next timer period */ + /* Enable next timer period: */ mod_timer(timer, jiffies + SCIR_CPU_HB_INTERVAL); } @@ -937,7 +950,7 @@ static int uv_heartbeat_enable(unsigned add_timer_on(timer, cpu); uv_cpu_scir_info(cpu)->enabled = 1; - /* also ensure that boot cpu is enabled */ + /* Also ensure that boot CPU is enabled: */ cpu = 0; } return 0; @@ -970,9 +983,11 @@ static __init int uv_init_heartbeat(void { int cpu; - if (is_uv_system()) + if (is_uv_system()) { for_each_online_cpu(cpu) uv_heartbeat_enable(cpu); + } + return 0; } @@ -981,14 +996,10 @@ late_initcall(uv_init_heartbeat); #endif /* !CONFIG_HOTPLUG_CPU */ /* Direct Legacy VGA I/O traffic to designated IOH */ -int uv_set_vga_state(struct pci_dev *pdev, bool decode, - unsigned int command_bits, u32 flags) +int uv_set_vga_state(struct pci_dev *pdev, bool decode, unsigned int command_bits, u32 flags) { int domain, bus, rc; - PR_DEVEL("devfn %x decode %d cmd %x flags %d\n", - pdev->devfn, decode, command_bits, flags); - if (!(flags & PCI_VGA_STATE_CHANGE_BRIDGE)) return 0; @@ -999,13 +1010,12 @@ int uv_set_vga_state(struct pci_dev *pde bus = pdev->bus->number; rc = uv_bios_set_legacy_vga_target(decode, domain, bus); - PR_DEVEL("vga decode %d %x:%x, rc: %d\n", decode, domain, bus, rc); return rc; } /* - * Called on each cpu to initialize the per_cpu UV data area. + * Called on each CPU to initialize the per_cpu UV data area. * FIXME: hotplug not supported yet */ void uv_cpu_init(void) @@ -1032,92 +1042,79 @@ static void get_mn(struct mn *mnp) union uvh_rh_gam_config_mmr_u m_n_config; union uv3h_gr0_gam_gr_config_u m_gr_config; - m_n_config.v = uv_read_local_mmr(UVH_RH_GAM_CONFIG_MMR); - mnp->n_val = m_n_config.s.n_skt; + /* Make sure the whole structure is well initialized: */ + memset(mnp, 0, sizeof(*mnp)); + + m_n_config.v = uv_read_local_mmr(UVH_RH_GAM_CONFIG_MMR); + mnp->n_val = m_n_config.s.n_skt; + if (is_uv4_hub()) { - mnp->m_val = 0; - mnp->n_lshift = 0; + mnp->m_val = 0; + mnp->n_lshift = 0; } else if (is_uv3_hub()) { - mnp->m_val = m_n_config.s3.m_skt; - m_gr_config.v = uv_read_local_mmr(UV3H_GR0_GAM_GR_CONFIG); - mnp->n_lshift = m_gr_config.s3.m_skt; + mnp->m_val = m_n_config.s3.m_skt; + m_gr_config.v = uv_read_local_mmr(UV3H_GR0_GAM_GR_CONFIG); + mnp->n_lshift = m_gr_config.s3.m_skt; } else if (is_uv2_hub()) { - mnp->m_val = m_n_config.s2.m_skt; - mnp->n_lshift = mnp->m_val == 40 ? 40 : 39; + mnp->m_val = m_n_config.s2.m_skt; + mnp->n_lshift = mnp->m_val == 40 ? 40 : 39; } else if (is_uv1_hub()) { - mnp->m_val = m_n_config.s1.m_skt; - mnp->n_lshift = mnp->m_val; + mnp->m_val = m_n_config.s1.m_skt; + mnp->n_lshift = mnp->m_val; } mnp->m_shift = mnp->m_val ? 64 - mnp->m_val : 0; } -void __init uv_init_hub_info(struct uv_hub_info_s *hub_info) +void __init uv_init_hub_info(struct uv_hub_info_s *hi) { - struct mn mn = {0}; /* avoid unitialized warnings */ union uvh_node_id_u node_id; + struct mn mn; get_mn(&mn); - hub_info->m_val = mn.m_val; - hub_info->n_val = mn.n_val; - hub_info->m_shift = mn.m_shift; - hub_info->n_lshift = mn.n_lshift ? mn.n_lshift : 0; - - hub_info->hub_revision = uv_hub_info->hub_revision; - hub_info->pnode_mask = uv_cpuid.pnode_mask; - hub_info->min_pnode = _min_pnode; - hub_info->min_socket = _min_socket; - hub_info->pnode_to_socket = _pnode_to_socket; - hub_info->socket_to_node = _socket_to_node; - hub_info->socket_to_pnode = _socket_to_pnode; - hub_info->gr_table_len = _gr_table_len; - hub_info->gr_table = _gr_table; - hub_info->gpa_mask = mn.m_val ? + hi->gpa_mask = mn.m_val ? (1UL << (mn.m_val + mn.n_val)) - 1 : (1UL << uv_cpuid.gpa_shift) - 1; - node_id.v = uv_read_local_mmr(UVH_NODE_ID); - uv_cpuid.gnode_shift = max_t(unsigned int, - uv_cpuid.gnode_shift, mn.n_val); - hub_info->gnode_extra = - (node_id.s.node_id & ~((1 << uv_cpuid.gnode_shift) - 1)) >> 1; - - hub_info->gnode_upper = - ((unsigned long)hub_info->gnode_extra << mn.m_val); + hi->m_val = mn.m_val; + hi->n_val = mn.n_val; + hi->m_shift = mn.m_shift; + hi->n_lshift = mn.n_lshift ? mn.n_lshift : 0; + hi->hub_revision = uv_hub_info->hub_revision; + hi->pnode_mask = uv_cpuid.pnode_mask; + hi->min_pnode = _min_pnode; + hi->min_socket = _min_socket; + hi->pnode_to_socket = _pnode_to_socket; + hi->socket_to_node = _socket_to_node; + hi->socket_to_pnode = _socket_to_pnode; + hi->gr_table_len = _gr_table_len; + hi->gr_table = _gr_table; + + node_id.v = uv_read_local_mmr(UVH_NODE_ID); + uv_cpuid.gnode_shift = max_t(unsigned int, uv_cpuid.gnode_shift, mn.n_val); + hi->gnode_extra = (node_id.s.node_id & ~((1 << uv_cpuid.gnode_shift) - 1)) >> 1; + hi->gnode_upper = (unsigned long)hi->gnode_extra << mn.m_val; if (uv_gp_table) { - hub_info->global_mmr_base = uv_gp_table->mmr_base; - hub_info->global_mmr_shift = uv_gp_table->mmr_shift; - hub_info->global_gru_base = uv_gp_table->gru_base; - hub_info->global_gru_shift = uv_gp_table->gru_shift; - hub_info->gpa_shift = uv_gp_table->gpa_shift; - hub_info->gpa_mask = (1UL << hub_info->gpa_shift) - 1; + hi->global_mmr_base = uv_gp_table->mmr_base; + hi->global_mmr_shift = uv_gp_table->mmr_shift; + hi->global_gru_base = uv_gp_table->gru_base; + hi->global_gru_shift = uv_gp_table->gru_shift; + hi->gpa_shift = uv_gp_table->gpa_shift; + hi->gpa_mask = (1UL << hi->gpa_shift) - 1; } else { - hub_info->global_mmr_base = - uv_read_local_mmr(UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR) & - ~UV_MMR_ENABLE; - hub_info->global_mmr_shift = _UV_GLOBAL_MMR64_PNODE_SHIFT; + hi->global_mmr_base = uv_read_local_mmr(UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR) & ~UV_MMR_ENABLE; + hi->global_mmr_shift = _UV_GLOBAL_MMR64_PNODE_SHIFT; } - get_lowmem_redirect( - &hub_info->lowmem_remap_base, &hub_info->lowmem_remap_top); - - hub_info->apic_pnode_shift = uv_cpuid.socketid_shift; + get_lowmem_redirect(&hi->lowmem_remap_base, &hi->lowmem_remap_top); - /* show system specific info */ - pr_info("UV: N:%d M:%d m_shift:%d n_lshift:%d\n", - hub_info->n_val, hub_info->m_val, - hub_info->m_shift, hub_info->n_lshift); + hi->apic_pnode_shift = uv_cpuid.socketid_shift; - pr_info("UV: gpa_mask/shift:0x%lx/%d pnode_mask:0x%x apic_pns:%d\n", - hub_info->gpa_mask, hub_info->gpa_shift, - hub_info->pnode_mask, hub_info->apic_pnode_shift); - - pr_info("UV: mmr_base/shift:0x%lx/%ld gru_base/shift:0x%lx/%ld\n", - hub_info->global_mmr_base, hub_info->global_mmr_shift, - hub_info->global_gru_base, hub_info->global_gru_shift); - - pr_info("UV: gnode_upper:0x%lx gnode_extra:0x%x\n", - hub_info->gnode_upper, hub_info->gnode_extra); + /* Show system specific info: */ + pr_info("UV: N:%d M:%d m_shift:%d n_lshift:%d\n", hi->n_val, hi->m_val, hi->m_shift, hi->n_lshift); + pr_info("UV: gpa_mask/shift:0x%lx/%d pnode_mask:0x%x apic_pns:%d\n", hi->gpa_mask, hi->gpa_shift, hi->pnode_mask, hi->apic_pnode_shift); + pr_info("UV: mmr_base/shift:0x%lx/%ld gru_base/shift:0x%lx/%ld\n", hi->global_mmr_base, hi->global_mmr_shift, hi->global_gru_base, hi->global_gru_shift); + pr_info("UV: gnode_upper:0x%lx gnode_extra:0x%x\n", hi->gnode_upper, hi->gnode_extra); } static void __init decode_gam_params(unsigned long ptr) @@ -1143,12 +1140,9 @@ static void __init decode_gam_rng_tbl(un for (; gre->type != UV_GAM_RANGE_TYPE_UNUSED; gre++) { if (!index) { pr_info("UV: GAM Range Table...\n"); - pr_info("UV: # %20s %14s %5s %4s %5s %3s %2s\n", - "Range", "", "Size", "Type", "NASID", - "SID", "PN"); + pr_info("UV: # %20s %14s %5s %4s %5s %3s %2s\n", "Range", "", "Size", "Type", "NASID", "SID", "PN"); } - pr_info( - "UV: %2d: 0x%014lx-0x%014lx %5luG %3d %04x %02x %02x\n", + pr_info("UV: %2d: 0x%014lx-0x%014lx %5luG %3d %04x %02x %02x\n", index++, (unsigned long)lgre << UV_GAM_RANGE_SHFT, (unsigned long)gre->limit << UV_GAM_RANGE_SHFT, @@ -1166,14 +1160,13 @@ static void __init decode_gam_rng_tbl(un if (pnode_max < gre->pnode) pnode_max = gre->pnode; } - _min_socket = sock_min; - _max_socket = sock_max; - _min_pnode = pnode_min; - _max_pnode = pnode_max; - _gr_table_len = index; - pr_info( - "UV: GRT: %d entries, sockets(min:%x,max:%x) pnodes(min:%x,max:%x)\n", - index, _min_socket, _max_socket, _min_pnode, _max_pnode); + _min_socket = sock_min; + _max_socket = sock_max; + _min_pnode = pnode_min; + _max_pnode = pnode_max; + _gr_table_len = index; + + pr_info("UV: GRT: %d entries, sockets(min:%x,max:%x) pnodes(min:%x,max:%x)\n", index, _min_socket, _max_socket, _min_pnode, _max_pnode); } static int __init decode_uv_systab(void) @@ -1188,12 +1181,10 @@ static int __init decode_uv_systab(void) if ((!st) || (st->revision < UV_SYSTAB_VERSION_UV4_LATEST)) { int rev = st ? st->revision : 0; - pr_err( - "UV: BIOS UVsystab version(%x) mismatch, expecting(%x)\n", - rev, UV_SYSTAB_VERSION_UV4_LATEST); - pr_err( - "UV: Cannot support UV operations, switching to generic PC\n"); + pr_err("UV: BIOS UVsystab version(%x) mismatch, expecting(%x)\n", rev, UV_SYSTAB_VERSION_UV4_LATEST); + pr_err("UV: Cannot support UV operations, switching to generic PC\n"); uv_system_type = UV_NONE; + return -EINVAL; } @@ -1219,7 +1210,7 @@ static int __init decode_uv_systab(void) } /* - * Setup physical blade translations from UVH_NODE_PRESENT_TABLE + * Set up physical blade translations from UVH_NODE_PRESENT_TABLE * .. NB: UVH_NODE_PRESENT_TABLE is going away, * .. being replaced by GAM Range Table */ @@ -1255,14 +1246,13 @@ static void __init build_socket_tables(v if (!gre) { if (is_uv1_hub() || is_uv2_hub() || is_uv3_hub()) { pr_info("UV: No UVsystab socket table, ignoring\n"); - return; /* not required */ + return; } - pr_crit( - "UV: Error: UVsystab address translations not available!\n"); + pr_crit("UV: Error: UVsystab address translations not available!\n"); BUG(); } - /* build socket id -> node id, pnode */ + /* Build socket id -> node id, pnode */ num = maxsock - minsock + 1; bytes = num * sizeof(_socket_to_node[0]); _socket_to_node = kmalloc(bytes, GFP_KERNEL); @@ -1279,27 +1269,27 @@ static void __init build_socket_tables(v for (i = 0; i < nump; i++) _pnode_to_socket[i] = SOCK_EMPTY; - /* fill in pnode/node/addr conversion list values */ + /* Fill in pnode/node/addr conversion list values: */ pr_info("UV: GAM Building socket/pnode conversion tables\n"); for (; gre->type != UV_GAM_RANGE_TYPE_UNUSED; gre++) { if (gre->type == UV_GAM_RANGE_TYPE_HOLE) continue; i = gre->sockid - minsock; + /* Duplicate: */ if (_socket_to_pnode[i] != SOCK_EMPTY) - continue; /* duplicate */ + continue; _socket_to_pnode[i] = gre->pnode; i = gre->pnode - minpnode; _pnode_to_socket[i] = gre->sockid; - pr_info( - "UV: sid:%02x type:%d nasid:%04x pn:%02x pn2s:%2x\n", + pr_info("UV: sid:%02x type:%d nasid:%04x pn:%02x pn2s:%2x\n", gre->sockid, gre->type, gre->nasid, _socket_to_pnode[gre->sockid - minsock], _pnode_to_socket[gre->pnode - minpnode]); } - /* Set socket -> node values */ + /* Set socket -> node values: */ lnid = -1; for_each_present_cpu(cpu) { int nid = cpu_to_node(cpu); @@ -1315,7 +1305,7 @@ static void __init build_socket_tables(v sockid, apicid, nid); } - /* Setup physical blade to pnode translation from GAM Range Table */ + /* Set up physical blade to pnode translation from GAM Range Table: */ bytes = num_possible_nodes() * sizeof(_node_to_pnode[0]); _node_to_pnode = kmalloc(bytes, GFP_KERNEL); BUG_ON(!_node_to_pnode); @@ -1325,8 +1315,7 @@ static void __init build_socket_tables(v for (sockid = minsock; sockid <= maxsock; sockid++) { if (lnid == _socket_to_node[sockid - minsock]) { - _node_to_pnode[lnid] = - _socket_to_pnode[sockid - minsock]; + _node_to_pnode[lnid] = _socket_to_pnode[sockid - minsock]; break; } } @@ -1343,8 +1332,7 @@ static void __init build_socket_tables(v pr_info("UV: Checking socket->node/pnode for identity maps\n"); if (minsock == 0) { for (i = 0; i < num; i++) - if (_socket_to_node[i] == SOCK_EMPTY || - i != _socket_to_node[i]) + if (_socket_to_node[i] == SOCK_EMPTY || i != _socket_to_node[i]) break; if (i >= num) { kfree(_socket_to_node); @@ -1383,9 +1371,13 @@ void __init uv_system_init(void) map_low_mmrs(); - uv_bios_init(); /* get uv_systab for decoding */ + /* Get uv_systab for decoding: */ + uv_bios_init(); + + /* If there's an UVsystab problem then abort UV init: */ if (decode_uv_systab() < 0) - return; /* UVsystab problem, abort UV init */ + return; + build_socket_tables(); build_uv_gr_table(); uv_init_hub_info(&hub_info); @@ -1393,14 +1385,10 @@ void __init uv_system_init(void) if (!_node_to_pnode) boot_init_possible_blades(&hub_info); - /* uv_num_possible_blades() is really the hub count */ - pr_info("UV: Found %d hubs, %d nodes, %d cpus\n", - uv_num_possible_blades(), - num_possible_nodes(), - num_possible_cpus()); + /* uv_num_possible_blades() is really the hub count: */ + pr_info("UV: Found %d hubs, %d nodes, %d CPUs\n", uv_num_possible_blades(), num_possible_nodes(), num_possible_cpus()); - uv_bios_get_sn_info(0, &uv_type, &sn_partition_id, &sn_coherency_id, - &sn_region_size, &system_serial_number); + uv_bios_get_sn_info(0, &uv_type, &sn_partition_id, &sn_coherency_id, &sn_region_size, &system_serial_number); hub_info.coherency_domain_number = sn_coherency_id; uv_rtc_init(); @@ -1413,33 +1401,31 @@ void __init uv_system_init(void) struct uv_hub_info_s *new_hub; if (__uv_hub_info_list[nodeid]) { - pr_err("UV: Node %d UV HUB already initialized!?\n", - nodeid); + pr_err("UV: Node %d UV HUB already initialized!?\n", nodeid); BUG(); } /* Allocate new per hub info list */ - new_hub = (nodeid == 0) ? - &uv_hub_info_node0 : - kzalloc_node(bytes, GFP_KERNEL, nodeid); + new_hub = (nodeid == 0) ? &uv_hub_info_node0 : kzalloc_node(bytes, GFP_KERNEL, nodeid); BUG_ON(!new_hub); __uv_hub_info_list[nodeid] = new_hub; new_hub = uv_hub_info_list(nodeid); BUG_ON(!new_hub); *new_hub = hub_info; - /* Use information from GAM table if available */ + /* Use information from GAM table if available: */ if (_node_to_pnode) new_hub->pnode = _node_to_pnode[nodeid]; - else /* Fill in during cpu loop */ + else /* Or fill in during CPU loop: */ new_hub->pnode = 0xffff; + new_hub->numa_blade_id = uv_node_to_blade_id(nodeid); new_hub->memory_nid = -1; new_hub->nr_possible_cpus = 0; new_hub->nr_online_cpus = 0; } - /* Initialize per cpu info */ + /* Initialize per CPU info: */ for_each_possible_cpu(cpu) { int apicid = per_cpu(x86_cpu_to_apicid, cpu); int numa_node_id; @@ -1450,22 +1436,24 @@ void __init uv_system_init(void) pnode = uv_apicid_to_pnode(apicid); uv_cpu_info_per(cpu)->p_uv_hub_info = uv_hub_info_list(nodeid); - uv_cpu_info_per(cpu)->blade_cpu_id = - uv_cpu_hub_info(cpu)->nr_possible_cpus++; + uv_cpu_info_per(cpu)->blade_cpu_id = uv_cpu_hub_info(cpu)->nr_possible_cpus++; if (uv_cpu_hub_info(cpu)->memory_nid == -1) uv_cpu_hub_info(cpu)->memory_nid = cpu_to_node(cpu); - if (nodeid != numa_node_id && /* init memoryless node */ + + /* Init memoryless node: */ + if (nodeid != numa_node_id && uv_hub_info_list(numa_node_id)->pnode == 0xffff) uv_hub_info_list(numa_node_id)->pnode = pnode; else if (uv_cpu_hub_info(cpu)->pnode == 0xffff) uv_cpu_hub_info(cpu)->pnode = pnode; + uv_cpu_scir_info(cpu)->offset = uv_scir_offset(apicid); } for_each_node(nodeid) { unsigned short pnode = uv_hub_info_list(nodeid)->pnode; - /* Add pnode info for pre-GAM list nodes without cpus */ + /* Add pnode info for pre-GAM list nodes without CPUs: */ if (pnode == 0xffff) { unsigned long paddr; @@ -1491,12 +1479,12 @@ void __init uv_system_init(void) uv_scir_register_cpu_notifier(); proc_mkdir("sgi_uv", NULL); - /* register Legacy VGA I/O redirection handler */ + /* Register Legacy VGA I/O redirection handler: */ pci_register_set_vga_state(uv_set_vga_state); /* * For a kdump kernel the reset must be BOOT_ACPI, not BOOT_EFI, as - * EFI is not enabled in the kdump kernel. + * EFI is not enabled in the kdump kernel: */ if (is_kdump_kernel()) reboot_type = BOOT_ACPI; ^ permalink raw reply [flat|nested] 15+ messages in thread
* Re: [PATCH] x86/platform/UV: Clean up the UV APIC code 2017-01-14 8:26 ` [PATCH] x86/platform/UV: Clean up the UV APIC code Ingo Molnar @ 2017-01-14 13:11 ` Joe Perches 2017-02-01 10:09 ` [tip:x86/platform] " tip-bot for Ingo Molnar 1 sibling, 0 replies; 15+ messages in thread From: Joe Perches @ 2017-01-14 13:11 UTC (permalink / raw) To: Ingo Molnar, Mike Travis Cc: Ingo Molnar, Thomas Gleixner, H. Peter Anvin, x86, linux-kernel, Dimitri Sivanich, Linus Torvalds On Sat, 2017-01-14 at 09:26 +0100, Ingo Molnar wrote: > * Ingo Molnar <mingo@kernel.org> wrote: [] > In particular the random line breaks, apparently to pacify checkpatch, were > horribly widespread. I just made the lines longer - that's still more readable > than the multi-line horror that was there before. [] > arch/x86/kernel/apic/x2apic_uv_x.c | 500 ++++++++++++++++++------------------- [] > - pr_info("UV: OEM IDs %s/%s, System/HUB Types %d/%d, uv_apic %d\n", > - oem_id, oem_table_id, uv_system_type, > - uv_min_hub_revision_id, uv_apic); > + pr_info("UV: OEM IDs %s/%s, System/HUB Types %d/%d, uv_apic %d\n", oem_id, oem_table_id, uv_system_type, uv_min_hub_revision_id, uv_apic); 140+ chars on a single line. Is this really better or easier to read? Long lines make it harder to humans to verify format and arguments used in printf type calls. ^ permalink raw reply [flat|nested] 15+ messages in thread
* [tip:x86/platform] x86/platform/UV: Clean up the UV APIC code 2017-01-14 8:26 ` [PATCH] x86/platform/UV: Clean up the UV APIC code Ingo Molnar 2017-01-14 13:11 ` Joe Perches @ 2017-02-01 10:09 ` tip-bot for Ingo Molnar 1 sibling, 0 replies; 15+ messages in thread From: tip-bot for Ingo Molnar @ 2017-02-01 10:09 UTC (permalink / raw) To: linux-tip-commits Cc: peterz, linux-kernel, tglx, torvalds, sivanich, travis, mingo, hpa Commit-ID: 7243e10689fd17a3e151f41216569295cefa2958 Gitweb: http://git.kernel.org/tip/7243e10689fd17a3e151f41216569295cefa2958 Author: Ingo Molnar <mingo@kernel.org> AuthorDate: Sat, 14 Jan 2017 09:26:12 +0100 Committer: Ingo Molnar <mingo@kernel.org> CommitDate: Wed, 1 Feb 2017 10:20:59 +0100 x86/platform/UV: Clean up the UV APIC code Make it more readable. Acked-by: Thomas Gleixner <tglx@linutronix.de> Cc: Dimitri Sivanich <sivanich@hpe.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Mike Travis <travis@sgi.com> Cc: Peter Zijlstra <peterz@infradead.org> Link: http://lkml.kernel.org/r/20170114082612.GA27842@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org> --- arch/x86/kernel/apic/x2apic_uv_x.c | 500 ++++++++++++++++++------------------- 1 file changed, 244 insertions(+), 256 deletions(-) diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index 97ea712..656994a 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -41,15 +41,13 @@ DEFINE_PER_CPU(int, x2apic_extra_bits); -#define PR_DEVEL(fmt, args...) pr_devel("%s: " fmt, __func__, args) +static enum uv_system_type uv_system_type; +static u64 gru_start_paddr, gru_end_paddr; +static u64 gru_dist_base, gru_first_node_paddr = -1LL, gru_last_node_paddr; +static u64 gru_dist_lmask, gru_dist_umask; +static union uvh_apicid uvh_apicid; -static enum uv_system_type uv_system_type; -static u64 gru_start_paddr, gru_end_paddr; -static u64 gru_dist_base, gru_first_node_paddr = -1LL, gru_last_node_paddr; -static u64 gru_dist_lmask, gru_dist_umask; -static union uvh_apicid uvh_apicid; - -/* info derived from CPUID */ +/* Information derived from CPUID: */ static struct { unsigned int apicid_shift; unsigned int apicid_mask; @@ -61,21 +59,25 @@ static struct { int uv_min_hub_revision_id; EXPORT_SYMBOL_GPL(uv_min_hub_revision_id); + unsigned int uv_apicid_hibits; EXPORT_SYMBOL_GPL(uv_apicid_hibits); static struct apic apic_x2apic_uv_x; static struct uv_hub_info_s uv_hub_info_node0; -/* Set this to use hardware error handler instead of kernel panic */ +/* Set this to use hardware error handler instead of kernel panic: */ static int disable_uv_undefined_panic = 1; + unsigned long uv_undefined(char *str) { if (likely(!disable_uv_undefined_panic)) panic("UV: error: undefined MMR: %s\n", str); else pr_crit("UV: error: undefined MMR: %s\n", str); - return ~0ul; /* cause a machine fault */ + + /* Cause a machine fault: */ + return ~0ul; } EXPORT_SYMBOL(uv_undefined); @@ -86,18 +88,19 @@ static unsigned long __init uv_early_read_mmr(unsigned long addr) mmr = early_ioremap(UV_LOCAL_MMR_BASE | addr, sizeof(*mmr)); val = *mmr; early_iounmap(mmr, sizeof(*mmr)); + return val; } static inline bool is_GRU_range(u64 start, u64 end) { if (gru_dist_base) { - u64 su = start & gru_dist_umask; /* upper (incl pnode) bits */ - u64 sl = start & gru_dist_lmask; /* base offset bits */ + u64 su = start & gru_dist_umask; /* Upper (incl pnode) bits */ + u64 sl = start & gru_dist_lmask; /* Base offset bits */ u64 eu = end & gru_dist_umask; u64 el = end & gru_dist_lmask; - /* Must reside completely within a single GRU range */ + /* Must reside completely within a single GRU range: */ return (sl == gru_dist_base && el == gru_dist_base && su >= gru_first_node_paddr && su <= gru_last_node_paddr && @@ -141,7 +144,7 @@ static int __init early_get_pnodeid(void) uv_hub_info->hub_revision = uv_min_hub_revision_id; uv_cpuid.pnode_mask = (1 << m_n_config.s.n_skt) - 1; pnode = (node_id.s.node_id >> 1) & uv_cpuid.pnode_mask; - uv_cpuid.gpa_shift = 46; /* default unless changed */ + uv_cpuid.gpa_shift = 46; /* Default unless changed */ pr_info("UV: rev:%d part#:%x nodeid:%04x n_skt:%d pnmsk:%x pn:%x\n", node_id.s.revision, node_id.s.part_number, node_id.s.node_id, @@ -149,11 +152,12 @@ static int __init early_get_pnodeid(void) return pnode; } -/* [copied from arch/x86/kernel/cpu/topology.c:detect_extended_topology()] */ -#define SMT_LEVEL 0 /* leaf 0xb SMT level */ -#define INVALID_TYPE 0 /* leaf 0xb sub-leaf types */ -#define SMT_TYPE 1 -#define CORE_TYPE 2 +/* [Copied from arch/x86/kernel/cpu/topology.c:detect_extended_topology()] */ + +#define SMT_LEVEL 0 /* Leaf 0xb SMT level */ +#define INVALID_TYPE 0 /* Leaf 0xb sub-leaf types */ +#define SMT_TYPE 1 +#define CORE_TYPE 2 #define LEAFB_SUBTYPE(ecx) (((ecx) >> 8) & 0xff) #define BITS_SHIFT_NEXT_LEVEL(eax) ((eax) & 0x1f) @@ -167,11 +171,13 @@ static void set_x2apic_bits(void) pr_info("UV: CPU does not have CPUID.11\n"); return; } + cpuid_count(0xb, SMT_LEVEL, &eax, &ebx, &ecx, &edx); if (ebx == 0 || (LEAFB_SUBTYPE(ecx) != SMT_TYPE)) { pr_info("UV: CPUID.11 not implemented\n"); return; } + sid_shift = BITS_SHIFT_NEXT_LEVEL(eax); sub_index = 1; do { @@ -182,8 +188,9 @@ static void set_x2apic_bits(void) } sub_index++; } while (LEAFB_SUBTYPE(ecx) != INVALID_TYPE); - uv_cpuid.apicid_shift = 0; - uv_cpuid.apicid_mask = (~(-1 << sid_shift)); + + uv_cpuid.apicid_shift = 0; + uv_cpuid.apicid_mask = (~(-1 << sid_shift)); uv_cpuid.socketid_shift = sid_shift; } @@ -194,10 +201,8 @@ static void __init early_get_apic_socketid_shift(void) set_x2apic_bits(); - pr_info("UV: apicid_shift:%d apicid_mask:0x%x\n", - uv_cpuid.apicid_shift, uv_cpuid.apicid_mask); - pr_info("UV: socketid_shift:%d pnode_mask:0x%x\n", - uv_cpuid.socketid_shift, uv_cpuid.pnode_mask); + pr_info("UV: apicid_shift:%d apicid_mask:0x%x\n", uv_cpuid.apicid_shift, uv_cpuid.apicid_mask); + pr_info("UV: socketid_shift:%d pnode_mask:0x%x\n", uv_cpuid.socketid_shift, uv_cpuid.pnode_mask); } /* @@ -210,10 +215,8 @@ static void __init uv_set_apicid_hibit(void) union uv1h_lb_target_physical_apic_id_mask_u apicid_mask; if (is_uv1_hub()) { - apicid_mask.v = - uv_early_read_mmr(UV1H_LB_TARGET_PHYSICAL_APIC_ID_MASK); - uv_apicid_hibits = - apicid_mask.s1.bit_enables & UV_APICID_HIBIT_MASK; + apicid_mask.v = uv_early_read_mmr(UV1H_LB_TARGET_PHYSICAL_APIC_ID_MASK); + uv_apicid_hibits = apicid_mask.s1.bit_enables & UV_APICID_HIBIT_MASK; } } @@ -230,12 +233,12 @@ static int __init uv_acpi_madt_oem_check(char *oem_id, char *oem_table_id) return 0; } - /* Setup early hub type field in uv_hub_info for Node 0 */ + /* Set up early hub type field in uv_hub_info for Node 0 */ uv_cpu_info->p_uv_hub_info = &uv_hub_info_node0; /* * Determine UV arch type. - * SGI: UV100/1000 + * SGI: UV100/1000 * SGI2: UV2000/3000 * SGI3: UV300 (truncated to 4 chars because of different varieties) * SGI4: UV400 (truncated to 4 chars because of different varieties) @@ -251,31 +254,32 @@ static int __init uv_acpi_madt_oem_check(char *oem_id, char *oem_table_id) pnodeid = early_get_pnodeid(); early_get_apic_socketid_shift(); - x86_platform.is_untracked_pat_range = uv_is_untracked_pat_range; + + x86_platform.is_untracked_pat_range = uv_is_untracked_pat_range; x86_platform.nmi_init = uv_nmi_init; - if (!strcmp(oem_table_id, "UVX")) { /* most common */ + if (!strcmp(oem_table_id, "UVX")) { + /* This is the most common hardware variant: */ uv_system_type = UV_X2APIC; uv_apic = 0; - } else if (!strcmp(oem_table_id, "UVH")) { /* only UV1 systems */ + } else if (!strcmp(oem_table_id, "UVH")) { + /* Only UV1 systems: */ uv_system_type = UV_NON_UNIQUE_APIC; - __this_cpu_write(x2apic_extra_bits, - pnodeid << uvh_apicid.s.pnode_shift); + __this_cpu_write(x2apic_extra_bits, pnodeid << uvh_apicid.s.pnode_shift); uv_set_apicid_hibit(); uv_apic = 1; - } else if (!strcmp(oem_table_id, "UVL")) { /* only used for */ - uv_system_type = UV_LEGACY_APIC; /* very small systems */ + } else if (!strcmp(oem_table_id, "UVL")) { + /* Only used for very small systems: */ + uv_system_type = UV_LEGACY_APIC; uv_apic = 0; } else { goto badbios; } - pr_info("UV: OEM IDs %s/%s, System/HUB Types %d/%d, uv_apic %d\n", - oem_id, oem_table_id, uv_system_type, - uv_min_hub_revision_id, uv_apic); + pr_info("UV: OEM IDs %s/%s, System/HUB Types %d/%d, uv_apic %d\n", oem_id, oem_table_id, uv_system_type, uv_min_hub_revision_id, uv_apic); return uv_apic; @@ -308,16 +312,18 @@ EXPORT_SYMBOL_GPL(uv_possible_blades); unsigned long sn_rtc_cycles_per_second; EXPORT_SYMBOL(sn_rtc_cycles_per_second); -/* the following values are used for the per node hub info struct */ -static __initdata unsigned short *_node_to_pnode; -static __initdata unsigned short _min_socket, _max_socket; -static __initdata unsigned short _min_pnode, _max_pnode, _gr_table_len; -static __initdata struct uv_gam_range_entry *uv_gre_table; -static __initdata struct uv_gam_parameters *uv_gp_table; -static __initdata unsigned short *_socket_to_node; -static __initdata unsigned short *_socket_to_pnode; -static __initdata unsigned short *_pnode_to_socket; -static __initdata struct uv_gam_range_s *_gr_table; +/* The following values are used for the per node hub info struct */ +static __initdata unsigned short *_node_to_pnode; +static __initdata unsigned short _min_socket, _max_socket; +static __initdata unsigned short _min_pnode, _max_pnode, _gr_table_len; +static __initdata struct uv_gam_range_entry *uv_gre_table; +static __initdata struct uv_gam_parameters *uv_gp_table; +static __initdata unsigned short *_socket_to_node; +static __initdata unsigned short *_socket_to_pnode; +static __initdata unsigned short *_pnode_to_socket; + +static __initdata struct uv_gam_range_s *_gr_table; + #define SOCK_EMPTY ((unsigned short)~0) extern int uv_hub_info_version(void) @@ -326,7 +332,7 @@ extern int uv_hub_info_version(void) } EXPORT_SYMBOL(uv_hub_info_version); -/* Build GAM range lookup table */ +/* Build GAM range lookup table: */ static __init void build_uv_gr_table(void) { struct uv_gam_range_entry *gre = uv_gre_table; @@ -344,25 +350,24 @@ static __init void build_uv_gr_table(void) for (; gre->type != UV_GAM_RANGE_TYPE_UNUSED; gre++) { if (gre->type == UV_GAM_RANGE_TYPE_HOLE) { - if (!ram_limit) { /* mark hole between ram/non-ram */ + if (!ram_limit) { + /* Mark hole between RAM/non-RAM: */ ram_limit = last_limit; last_limit = gre->limit; lsid++; continue; } last_limit = gre->limit; - pr_info("UV: extra hole in GAM RE table @%d\n", - (int)(gre - uv_gre_table)); + pr_info("UV: extra hole in GAM RE table @%d\n", (int)(gre - uv_gre_table)); continue; } if (_max_socket < gre->sockid) { - pr_err("UV: GAM table sockid(%d) too large(>%d) @%d\n", - gre->sockid, _max_socket, - (int)(gre - uv_gre_table)); + pr_err("UV: GAM table sockid(%d) too large(>%d) @%d\n", gre->sockid, _max_socket, (int)(gre - uv_gre_table)); continue; } sid = gre->sockid - _min_socket; - if (lsid < sid) { /* new range */ + if (lsid < sid) { + /* New range: */ grt = &_gr_table[indx]; grt->base = lindx; grt->nasid = gre->nasid; @@ -371,27 +376,32 @@ static __init void build_uv_gr_table(void) lindx = indx++; continue; } - if (lsid == sid && !ram_limit) { /* update range */ - if (grt->limit == last_limit) { /* .. if contiguous */ + /* Update range: */ + if (lsid == sid && !ram_limit) { + /* .. if contiguous: */ + if (grt->limit == last_limit) { grt->limit = last_limit = gre->limit; continue; } } - if (!ram_limit) { /* non-contiguous ram range */ + /* Non-contiguous RAM range: */ + if (!ram_limit) { grt++; grt->base = lindx; grt->nasid = gre->nasid; grt->limit = last_limit = gre->limit; continue; } - grt++; /* non-contiguous/non-ram */ - grt->base = grt - _gr_table; /* base is this entry */ + /* Non-contiguous/non-RAM: */ + grt++; + /* base is this entry */ + grt->base = grt - _gr_table; grt->nasid = gre->nasid; grt->limit = last_limit = gre->limit; lsid++; } - /* shorten table if possible */ + /* Shorten table if possible */ grt++; i = grt - _gr_table; if (i < _gr_table_len) { @@ -405,16 +415,15 @@ static __init void build_uv_gr_table(void) } } - /* display resultant gam range table */ + /* Display resultant GAM range table: */ for (i = 0, grt = _gr_table; i < _gr_table_len; i++, grt++) { + unsigned long start, end; int gb = grt->base; - unsigned long start = gb < 0 ? 0 : - (unsigned long)_gr_table[gb].limit << UV_GAM_RANGE_SHFT; - unsigned long end = - (unsigned long)grt->limit << UV_GAM_RANGE_SHFT; - pr_info("UV: GAM Range %2d %04x 0x%013lx-0x%013lx (%d)\n", - i, grt->nasid, start, end, gb); + start = gb < 0 ? 0 : (unsigned long)_gr_table[gb].limit << UV_GAM_RANGE_SHFT; + end = (unsigned long)grt->limit << UV_GAM_RANGE_SHFT; + + pr_info("UV: GAM Range %2d %04x 0x%013lx-0x%013lx (%d)\n", i, grt->nasid, start, end, gb); } } @@ -425,16 +434,19 @@ static int uv_wakeup_secondary(int phys_apicid, unsigned long start_rip) pnode = uv_apicid_to_pnode(phys_apicid); phys_apicid |= uv_apicid_hibits; + val = (1UL << UVH_IPI_INT_SEND_SHFT) | (phys_apicid << UVH_IPI_INT_APIC_ID_SHFT) | ((start_rip << UVH_IPI_INT_VECTOR_SHFT) >> 12) | APIC_DM_INIT; + uv_write_global_mmr64(pnode, UVH_IPI_INT, val); val = (1UL << UVH_IPI_INT_SEND_SHFT) | (phys_apicid << UVH_IPI_INT_APIC_ID_SHFT) | ((start_rip << UVH_IPI_INT_VECTOR_SHFT) >> 12) | APIC_DM_STARTUP; + uv_write_global_mmr64(pnode, UVH_IPI_INT, val); return 0; @@ -568,7 +580,7 @@ static struct apic apic_x2apic_uv_x __ro_after_init = { .apic_id_registered = uv_apic_id_registered, .irq_delivery_mode = dest_Fixed, - .irq_dest_mode = 0, /* physical */ + .irq_dest_mode = 0, /* Physical */ .target_cpus = online_target_cpus, .disable_esr = 0, @@ -629,23 +641,22 @@ static __init void get_lowmem_redirect(unsigned long *base, unsigned long *size) switch (i) { case 0: m_redirect = UVH_RH_GAM_ALIAS210_REDIRECT_CONFIG_0_MMR; - m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_0_MMR; + m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_0_MMR; break; case 1: m_redirect = UVH_RH_GAM_ALIAS210_REDIRECT_CONFIG_1_MMR; - m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_1_MMR; + m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_1_MMR; break; case 2: m_redirect = UVH_RH_GAM_ALIAS210_REDIRECT_CONFIG_2_MMR; - m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_2_MMR; + m_overlay = UVH_RH_GAM_ALIAS210_OVERLAY_CONFIG_2_MMR; break; } alias.v = uv_read_local_mmr(m_overlay); if (alias.s.enable && alias.s.base == 0) { *size = (1UL << alias.s.m_alias); redirect.v = uv_read_local_mmr(m_redirect); - *base = (unsigned long)redirect.s.dest_base - << DEST_SHIFT; + *base = (unsigned long)redirect.s.dest_base << DEST_SHIFT; return; } } @@ -654,8 +665,7 @@ static __init void get_lowmem_redirect(unsigned long *base, unsigned long *size) enum map_type {map_wb, map_uc}; -static __init void map_high(char *id, unsigned long base, int pshift, - int bshift, int max_pnode, enum map_type map_type) +static __init void map_high(char *id, unsigned long base, int pshift, int bshift, int max_pnode, enum map_type map_type) { unsigned long bytes, paddr; @@ -680,16 +690,19 @@ static __init void map_gru_distributed(unsigned long c) int nid; gru.v = c; - /* only base bits 42:28 relevant in dist mode */ + + /* Only base bits 42:28 relevant in dist mode */ gru_dist_base = gru.v & 0x000007fff0000000UL; if (!gru_dist_base) { pr_info("UV: Map GRU_DIST base address NULL\n"); return; } + bytes = 1UL << UVH_RH_GAM_GRU_OVERLAY_CONFIG_MMR_BASE_SHFT; gru_dist_lmask = ((1UL << uv_hub_info->m_val) - 1) & ~(bytes - 1); gru_dist_umask = ~((1UL << uv_hub_info->m_val) - 1); gru_dist_base &= gru_dist_lmask; /* Clear bits above M */ + for_each_online_node(nid) { paddr = ((u64)uv_node_to_pnode(nid) << uv_hub_info->m_val) | gru_dist_base; @@ -697,11 +710,12 @@ static __init void map_gru_distributed(unsigned long c) gru_first_node_paddr = min(paddr, gru_first_node_paddr); gru_last_node_paddr = max(paddr, gru_last_node_paddr); } + /* Save upper (63:M) bits of address only for is_GRU_range */ gru_first_node_paddr &= gru_dist_umask; gru_last_node_paddr &= gru_dist_umask; - pr_debug("UV: Map GRU_DIST base 0x%016llx 0x%016llx - 0x%016llx\n", - gru_dist_base, gru_first_node_paddr, gru_last_node_paddr); + + pr_debug("UV: Map GRU_DIST base 0x%016llx 0x%016llx - 0x%016llx\n", gru_dist_base, gru_first_node_paddr, gru_last_node_paddr); } static __init void map_gru_high(int max_pnode) @@ -721,6 +735,7 @@ static __init void map_gru_high(int max_pnode) map_gru_distributed(gru.v); return; } + base = (gru.v & mask) >> shift; map_high("GRU", base, shift, shift, max_pnode, map_wb); gru_start_paddr = ((u64)base << shift); @@ -774,8 +789,8 @@ static __init void map_mmioh_high_uv3(int index, int min_pnode, int max_pnode) id = mmiohs[index].id; overlay.v = uv_read_local_mmr(mmiohs[index].overlay); - pr_info("UV: %s overlay 0x%lx base:0x%x m_io:%d\n", - id, overlay.v, overlay.s3.base, overlay.s3.m_io); + + pr_info("UV: %s overlay 0x%lx base:0x%x m_io:%d\n", id, overlay.v, overlay.s3.base, overlay.s3.m_io); if (!overlay.s3.enable) { pr_info("UV: %s disabled\n", id); return; @@ -786,7 +801,8 @@ static __init void map_mmioh_high_uv3(int index, int min_pnode, int max_pnode) m_io = overlay.s3.m_io; mmr = mmiohs[index].redirect; n = UV3H_RH_GAM_MMIOH_REDIRECT_CONFIG0_MMR_DEPTH; - min_pnode *= 2; /* convert to NASID */ + /* Convert to NASID: */ + min_pnode *= 2; max_pnode *= 2; max_io = lnasid = fi = li = -1; @@ -795,16 +811,18 @@ static __init void map_mmioh_high_uv3(int index, int min_pnode, int max_pnode) redirect.v = uv_read_local_mmr(mmr + i * 8); nasid = redirect.s3.nasid; + /* Invalid NASID: */ if (nasid < min_pnode || max_pnode < nasid) - nasid = -1; /* invalid NASID */ + nasid = -1; if (nasid == lnasid) { li = i; - if (i != n-1) /* last entry check */ + /* Last entry check: */ + if (i != n-1) continue; } - /* check if we have a cached (or last) redirect to print */ + /* Check if we have a cached (or last) redirect to print: */ if (lnasid != -1 || (i == n-1 && nasid != -1)) { unsigned long addr1, addr2; int f, l; @@ -816,12 +834,9 @@ static __init void map_mmioh_high_uv3(int index, int min_pnode, int max_pnode) f = fi; l = li; } - addr1 = (base << shift) + - f * (1ULL << m_io); - addr2 = (base << shift) + - (l + 1) * (1ULL << m_io); - pr_info("UV: %s[%03d..%03d] NASID 0x%04x ADDR 0x%016lx - 0x%016lx\n", - id, fi, li, lnasid, addr1, addr2); + addr1 = (base << shift) + f * (1ULL << m_io); + addr2 = (base << shift) + (l + 1) * (1ULL << m_io); + pr_info("UV: %s[%03d..%03d] NASID 0x%04x ADDR 0x%016lx - 0x%016lx\n", id, fi, li, lnasid, addr1, addr2); if (max_io < l) max_io = l; } @@ -829,8 +844,7 @@ static __init void map_mmioh_high_uv3(int index, int min_pnode, int max_pnode) lnasid = nasid; } - pr_info("UV: %s base:0x%lx shift:%d M_IO:%d MAX_IO:%d\n", - id, base, shift, m_io, max_io); + pr_info("UV: %s base:0x%lx shift:%d M_IO:%d MAX_IO:%d\n", id, base, shift, m_io, max_io); if (max_io >= 0) map_high(id, base, shift, m_io, max_io, map_uc); @@ -843,36 +857,35 @@ static __init void map_mmioh_high(int min_pnode, int max_pnode) int shift, enable, m_io, n_io; if (is_uv3_hub() || is_uv4_hub()) { - /* Map both MMIOH Regions */ + /* Map both MMIOH regions: */ map_mmioh_high_uv3(0, min_pnode, max_pnode); map_mmioh_high_uv3(1, min_pnode, max_pnode); return; } if (is_uv1_hub()) { - mmr = UV1H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR; - shift = UV1H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR_BASE_SHFT; - mmioh.v = uv_read_local_mmr(mmr); - enable = !!mmioh.s1.enable; - base = mmioh.s1.base; - m_io = mmioh.s1.m_io; - n_io = mmioh.s1.n_io; + mmr = UV1H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR; + shift = UV1H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR_BASE_SHFT; + mmioh.v = uv_read_local_mmr(mmr); + enable = !!mmioh.s1.enable; + base = mmioh.s1.base; + m_io = mmioh.s1.m_io; + n_io = mmioh.s1.n_io; } else if (is_uv2_hub()) { - mmr = UV2H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR; - shift = UV2H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR_BASE_SHFT; - mmioh.v = uv_read_local_mmr(mmr); - enable = !!mmioh.s2.enable; - base = mmioh.s2.base; - m_io = mmioh.s2.m_io; - n_io = mmioh.s2.n_io; - } else + mmr = UV2H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR; + shift = UV2H_RH_GAM_MMIOH_OVERLAY_CONFIG_MMR_BASE_SHFT; + mmioh.v = uv_read_local_mmr(mmr); + enable = !!mmioh.s2.enable; + base = mmioh.s2.base; + m_io = mmioh.s2.m_io; + n_io = mmioh.s2.n_io; + } else { return; + } if (enable) { max_pnode &= (1 << n_io) - 1; - pr_info( - "UV: base:0x%lx shift:%d N_IO:%d M_IO:%d max_pnode:0x%x\n", - base, shift, m_io, n_io, max_pnode); + pr_info("UV: base:0x%lx shift:%d N_IO:%d M_IO:%d max_pnode:0x%x\n", base, shift, m_io, n_io, max_pnode); map_high("MMIOH", base, shift, m_io, max_pnode, map_uc); } else { pr_info("UV: MMIOH disabled\n"); @@ -890,16 +903,16 @@ static __init void uv_rtc_init(void) long status; u64 ticks_per_sec; - status = uv_bios_freq_base(BIOS_FREQ_BASE_REALTIME_CLOCK, - &ticks_per_sec); + status = uv_bios_freq_base(BIOS_FREQ_BASE_REALTIME_CLOCK, &ticks_per_sec); + if (status != BIOS_STATUS_SUCCESS || ticks_per_sec < 100000) { - printk(KERN_WARNING - "unable to determine platform RTC clock frequency, " - "guessing.\n"); - /* BIOS gives wrong value for clock freq. so guess */ + pr_warn("UV: unable to determine platform RTC clock frequency, guessing.\n"); + + /* BIOS gives wrong value for clock frequency, so guess: */ sn_rtc_cycles_per_second = 1000000000000UL / 30000UL; - } else + } else { sn_rtc_cycles_per_second = ticks_per_sec; + } } /* @@ -910,19 +923,19 @@ static void uv_heartbeat(unsigned long ignored) struct timer_list *timer = &uv_scir_info->timer; unsigned char bits = uv_scir_info->state; - /* flip heartbeat bit */ + /* Flip heartbeat bit: */ bits ^= SCIR_CPU_HEARTBEAT; - /* is this cpu idle? */ + /* Is this CPU idle? */ if (idle_cpu(raw_smp_processor_id())) bits &= ~SCIR_CPU_ACTIVITY; else bits |= SCIR_CPU_ACTIVITY; - /* update system controller interface reg */ + /* Update system controller interface reg: */ uv_set_scir_bits(bits); - /* enable next timer period */ + /* Enable next timer period: */ mod_timer(timer, jiffies + SCIR_CPU_HB_INTERVAL); } @@ -937,7 +950,7 @@ static int uv_heartbeat_enable(unsigned int cpu) add_timer_on(timer, cpu); uv_cpu_scir_info(cpu)->enabled = 1; - /* also ensure that boot cpu is enabled */ + /* Also ensure that boot CPU is enabled: */ cpu = 0; } return 0; @@ -970,9 +983,11 @@ static __init int uv_init_heartbeat(void) { int cpu; - if (is_uv_system()) + if (is_uv_system()) { for_each_online_cpu(cpu) uv_heartbeat_enable(cpu); + } + return 0; } @@ -981,14 +996,10 @@ late_initcall(uv_init_heartbeat); #endif /* !CONFIG_HOTPLUG_CPU */ /* Direct Legacy VGA I/O traffic to designated IOH */ -int uv_set_vga_state(struct pci_dev *pdev, bool decode, - unsigned int command_bits, u32 flags) +int uv_set_vga_state(struct pci_dev *pdev, bool decode, unsigned int command_bits, u32 flags) { int domain, bus, rc; - PR_DEVEL("devfn %x decode %d cmd %x flags %d\n", - pdev->devfn, decode, command_bits, flags); - if (!(flags & PCI_VGA_STATE_CHANGE_BRIDGE)) return 0; @@ -999,13 +1010,12 @@ int uv_set_vga_state(struct pci_dev *pdev, bool decode, bus = pdev->bus->number; rc = uv_bios_set_legacy_vga_target(decode, domain, bus); - PR_DEVEL("vga decode %d %x:%x, rc: %d\n", decode, domain, bus, rc); return rc; } /* - * Called on each cpu to initialize the per_cpu UV data area. + * Called on each CPU to initialize the per_cpu UV data area. * FIXME: hotplug not supported yet */ void uv_cpu_init(void) @@ -1032,92 +1042,79 @@ static void get_mn(struct mn *mnp) union uvh_rh_gam_config_mmr_u m_n_config; union uv3h_gr0_gam_gr_config_u m_gr_config; - m_n_config.v = uv_read_local_mmr(UVH_RH_GAM_CONFIG_MMR); - mnp->n_val = m_n_config.s.n_skt; + /* Make sure the whole structure is well initialized: */ + memset(mnp, 0, sizeof(*mnp)); + + m_n_config.v = uv_read_local_mmr(UVH_RH_GAM_CONFIG_MMR); + mnp->n_val = m_n_config.s.n_skt; + if (is_uv4_hub()) { - mnp->m_val = 0; - mnp->n_lshift = 0; + mnp->m_val = 0; + mnp->n_lshift = 0; } else if (is_uv3_hub()) { - mnp->m_val = m_n_config.s3.m_skt; - m_gr_config.v = uv_read_local_mmr(UV3H_GR0_GAM_GR_CONFIG); - mnp->n_lshift = m_gr_config.s3.m_skt; + mnp->m_val = m_n_config.s3.m_skt; + m_gr_config.v = uv_read_local_mmr(UV3H_GR0_GAM_GR_CONFIG); + mnp->n_lshift = m_gr_config.s3.m_skt; } else if (is_uv2_hub()) { - mnp->m_val = m_n_config.s2.m_skt; - mnp->n_lshift = mnp->m_val == 40 ? 40 : 39; + mnp->m_val = m_n_config.s2.m_skt; + mnp->n_lshift = mnp->m_val == 40 ? 40 : 39; } else if (is_uv1_hub()) { - mnp->m_val = m_n_config.s1.m_skt; - mnp->n_lshift = mnp->m_val; + mnp->m_val = m_n_config.s1.m_skt; + mnp->n_lshift = mnp->m_val; } mnp->m_shift = mnp->m_val ? 64 - mnp->m_val : 0; } -void __init uv_init_hub_info(struct uv_hub_info_s *hub_info) +void __init uv_init_hub_info(struct uv_hub_info_s *hi) { - struct mn mn = {0}; /* avoid unitialized warnings */ union uvh_node_id_u node_id; + struct mn mn; get_mn(&mn); - hub_info->m_val = mn.m_val; - hub_info->n_val = mn.n_val; - hub_info->m_shift = mn.m_shift; - hub_info->n_lshift = mn.n_lshift ? mn.n_lshift : 0; - - hub_info->hub_revision = uv_hub_info->hub_revision; - hub_info->pnode_mask = uv_cpuid.pnode_mask; - hub_info->min_pnode = _min_pnode; - hub_info->min_socket = _min_socket; - hub_info->pnode_to_socket = _pnode_to_socket; - hub_info->socket_to_node = _socket_to_node; - hub_info->socket_to_pnode = _socket_to_pnode; - hub_info->gr_table_len = _gr_table_len; - hub_info->gr_table = _gr_table; - hub_info->gpa_mask = mn.m_val ? + hi->gpa_mask = mn.m_val ? (1UL << (mn.m_val + mn.n_val)) - 1 : (1UL << uv_cpuid.gpa_shift) - 1; - node_id.v = uv_read_local_mmr(UVH_NODE_ID); - uv_cpuid.gnode_shift = max_t(unsigned int, - uv_cpuid.gnode_shift, mn.n_val); - hub_info->gnode_extra = - (node_id.s.node_id & ~((1 << uv_cpuid.gnode_shift) - 1)) >> 1; - - hub_info->gnode_upper = - ((unsigned long)hub_info->gnode_extra << mn.m_val); + hi->m_val = mn.m_val; + hi->n_val = mn.n_val; + hi->m_shift = mn.m_shift; + hi->n_lshift = mn.n_lshift ? mn.n_lshift : 0; + hi->hub_revision = uv_hub_info->hub_revision; + hi->pnode_mask = uv_cpuid.pnode_mask; + hi->min_pnode = _min_pnode; + hi->min_socket = _min_socket; + hi->pnode_to_socket = _pnode_to_socket; + hi->socket_to_node = _socket_to_node; + hi->socket_to_pnode = _socket_to_pnode; + hi->gr_table_len = _gr_table_len; + hi->gr_table = _gr_table; + + node_id.v = uv_read_local_mmr(UVH_NODE_ID); + uv_cpuid.gnode_shift = max_t(unsigned int, uv_cpuid.gnode_shift, mn.n_val); + hi->gnode_extra = (node_id.s.node_id & ~((1 << uv_cpuid.gnode_shift) - 1)) >> 1; + hi->gnode_upper = (unsigned long)hi->gnode_extra << mn.m_val; if (uv_gp_table) { - hub_info->global_mmr_base = uv_gp_table->mmr_base; - hub_info->global_mmr_shift = uv_gp_table->mmr_shift; - hub_info->global_gru_base = uv_gp_table->gru_base; - hub_info->global_gru_shift = uv_gp_table->gru_shift; - hub_info->gpa_shift = uv_gp_table->gpa_shift; - hub_info->gpa_mask = (1UL << hub_info->gpa_shift) - 1; + hi->global_mmr_base = uv_gp_table->mmr_base; + hi->global_mmr_shift = uv_gp_table->mmr_shift; + hi->global_gru_base = uv_gp_table->gru_base; + hi->global_gru_shift = uv_gp_table->gru_shift; + hi->gpa_shift = uv_gp_table->gpa_shift; + hi->gpa_mask = (1UL << hi->gpa_shift) - 1; } else { - hub_info->global_mmr_base = - uv_read_local_mmr(UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR) & - ~UV_MMR_ENABLE; - hub_info->global_mmr_shift = _UV_GLOBAL_MMR64_PNODE_SHIFT; + hi->global_mmr_base = uv_read_local_mmr(UVH_RH_GAM_MMR_OVERLAY_CONFIG_MMR) & ~UV_MMR_ENABLE; + hi->global_mmr_shift = _UV_GLOBAL_MMR64_PNODE_SHIFT; } - get_lowmem_redirect( - &hub_info->lowmem_remap_base, &hub_info->lowmem_remap_top); - - hub_info->apic_pnode_shift = uv_cpuid.socketid_shift; + get_lowmem_redirect(&hi->lowmem_remap_base, &hi->lowmem_remap_top); - /* show system specific info */ - pr_info("UV: N:%d M:%d m_shift:%d n_lshift:%d\n", - hub_info->n_val, hub_info->m_val, - hub_info->m_shift, hub_info->n_lshift); + hi->apic_pnode_shift = uv_cpuid.socketid_shift; - pr_info("UV: gpa_mask/shift:0x%lx/%d pnode_mask:0x%x apic_pns:%d\n", - hub_info->gpa_mask, hub_info->gpa_shift, - hub_info->pnode_mask, hub_info->apic_pnode_shift); - - pr_info("UV: mmr_base/shift:0x%lx/%ld gru_base/shift:0x%lx/%ld\n", - hub_info->global_mmr_base, hub_info->global_mmr_shift, - hub_info->global_gru_base, hub_info->global_gru_shift); - - pr_info("UV: gnode_upper:0x%lx gnode_extra:0x%x\n", - hub_info->gnode_upper, hub_info->gnode_extra); + /* Show system specific info: */ + pr_info("UV: N:%d M:%d m_shift:%d n_lshift:%d\n", hi->n_val, hi->m_val, hi->m_shift, hi->n_lshift); + pr_info("UV: gpa_mask/shift:0x%lx/%d pnode_mask:0x%x apic_pns:%d\n", hi->gpa_mask, hi->gpa_shift, hi->pnode_mask, hi->apic_pnode_shift); + pr_info("UV: mmr_base/shift:0x%lx/%ld gru_base/shift:0x%lx/%ld\n", hi->global_mmr_base, hi->global_mmr_shift, hi->global_gru_base, hi->global_gru_shift); + pr_info("UV: gnode_upper:0x%lx gnode_extra:0x%x\n", hi->gnode_upper, hi->gnode_extra); } static void __init decode_gam_params(unsigned long ptr) @@ -1143,12 +1140,9 @@ static void __init decode_gam_rng_tbl(unsigned long ptr) for (; gre->type != UV_GAM_RANGE_TYPE_UNUSED; gre++) { if (!index) { pr_info("UV: GAM Range Table...\n"); - pr_info("UV: # %20s %14s %5s %4s %5s %3s %2s\n", - "Range", "", "Size", "Type", "NASID", - "SID", "PN"); + pr_info("UV: # %20s %14s %5s %4s %5s %3s %2s\n", "Range", "", "Size", "Type", "NASID", "SID", "PN"); } - pr_info( - "UV: %2d: 0x%014lx-0x%014lx %5luG %3d %04x %02x %02x\n", + pr_info("UV: %2d: 0x%014lx-0x%014lx %5luG %3d %04x %02x %02x\n", index++, (unsigned long)lgre << UV_GAM_RANGE_SHFT, (unsigned long)gre->limit << UV_GAM_RANGE_SHFT, @@ -1166,14 +1160,13 @@ static void __init decode_gam_rng_tbl(unsigned long ptr) if (pnode_max < gre->pnode) pnode_max = gre->pnode; } - _min_socket = sock_min; - _max_socket = sock_max; - _min_pnode = pnode_min; - _max_pnode = pnode_max; - _gr_table_len = index; - pr_info( - "UV: GRT: %d entries, sockets(min:%x,max:%x) pnodes(min:%x,max:%x)\n", - index, _min_socket, _max_socket, _min_pnode, _max_pnode); + _min_socket = sock_min; + _max_socket = sock_max; + _min_pnode = pnode_min; + _max_pnode = pnode_max; + _gr_table_len = index; + + pr_info("UV: GRT: %d entries, sockets(min:%x,max:%x) pnodes(min:%x,max:%x)\n", index, _min_socket, _max_socket, _min_pnode, _max_pnode); } static int __init decode_uv_systab(void) @@ -1188,12 +1181,10 @@ static int __init decode_uv_systab(void) if ((!st) || (st->revision < UV_SYSTAB_VERSION_UV4_LATEST)) { int rev = st ? st->revision : 0; - pr_err( - "UV: BIOS UVsystab version(%x) mismatch, expecting(%x)\n", - rev, UV_SYSTAB_VERSION_UV4_LATEST); - pr_err( - "UV: Cannot support UV operations, switching to generic PC\n"); + pr_err("UV: BIOS UVsystab version(%x) mismatch, expecting(%x)\n", rev, UV_SYSTAB_VERSION_UV4_LATEST); + pr_err("UV: Cannot support UV operations, switching to generic PC\n"); uv_system_type = UV_NONE; + return -EINVAL; } @@ -1219,7 +1210,7 @@ static int __init decode_uv_systab(void) } /* - * Setup physical blade translations from UVH_NODE_PRESENT_TABLE + * Set up physical blade translations from UVH_NODE_PRESENT_TABLE * .. NB: UVH_NODE_PRESENT_TABLE is going away, * .. being replaced by GAM Range Table */ @@ -1255,14 +1246,13 @@ static void __init build_socket_tables(void) if (!gre) { if (is_uv1_hub() || is_uv2_hub() || is_uv3_hub()) { pr_info("UV: No UVsystab socket table, ignoring\n"); - return; /* not required */ + return; } - pr_crit( - "UV: Error: UVsystab address translations not available!\n"); + pr_crit("UV: Error: UVsystab address translations not available!\n"); BUG(); } - /* build socket id -> node id, pnode */ + /* Build socket id -> node id, pnode */ num = maxsock - minsock + 1; bytes = num * sizeof(_socket_to_node[0]); _socket_to_node = kmalloc(bytes, GFP_KERNEL); @@ -1279,27 +1269,27 @@ static void __init build_socket_tables(void) for (i = 0; i < nump; i++) _pnode_to_socket[i] = SOCK_EMPTY; - /* fill in pnode/node/addr conversion list values */ + /* Fill in pnode/node/addr conversion list values: */ pr_info("UV: GAM Building socket/pnode conversion tables\n"); for (; gre->type != UV_GAM_RANGE_TYPE_UNUSED; gre++) { if (gre->type == UV_GAM_RANGE_TYPE_HOLE) continue; i = gre->sockid - minsock; + /* Duplicate: */ if (_socket_to_pnode[i] != SOCK_EMPTY) - continue; /* duplicate */ + continue; _socket_to_pnode[i] = gre->pnode; i = gre->pnode - minpnode; _pnode_to_socket[i] = gre->sockid; - pr_info( - "UV: sid:%02x type:%d nasid:%04x pn:%02x pn2s:%2x\n", + pr_info("UV: sid:%02x type:%d nasid:%04x pn:%02x pn2s:%2x\n", gre->sockid, gre->type, gre->nasid, _socket_to_pnode[gre->sockid - minsock], _pnode_to_socket[gre->pnode - minpnode]); } - /* Set socket -> node values */ + /* Set socket -> node values: */ lnid = -1; for_each_present_cpu(cpu) { int nid = cpu_to_node(cpu); @@ -1315,7 +1305,7 @@ static void __init build_socket_tables(void) sockid, apicid, nid); } - /* Setup physical blade to pnode translation from GAM Range Table */ + /* Set up physical blade to pnode translation from GAM Range Table: */ bytes = num_possible_nodes() * sizeof(_node_to_pnode[0]); _node_to_pnode = kmalloc(bytes, GFP_KERNEL); BUG_ON(!_node_to_pnode); @@ -1325,8 +1315,7 @@ static void __init build_socket_tables(void) for (sockid = minsock; sockid <= maxsock; sockid++) { if (lnid == _socket_to_node[sockid - minsock]) { - _node_to_pnode[lnid] = - _socket_to_pnode[sockid - minsock]; + _node_to_pnode[lnid] = _socket_to_pnode[sockid - minsock]; break; } } @@ -1343,8 +1332,7 @@ static void __init build_socket_tables(void) pr_info("UV: Checking socket->node/pnode for identity maps\n"); if (minsock == 0) { for (i = 0; i < num; i++) - if (_socket_to_node[i] == SOCK_EMPTY || - i != _socket_to_node[i]) + if (_socket_to_node[i] == SOCK_EMPTY || i != _socket_to_node[i]) break; if (i >= num) { kfree(_socket_to_node); @@ -1383,9 +1371,13 @@ void __init uv_system_init(void) map_low_mmrs(); - uv_bios_init(); /* get uv_systab for decoding */ + /* Get uv_systab for decoding: */ + uv_bios_init(); + + /* If there's an UVsystab problem then abort UV init: */ if (decode_uv_systab() < 0) - return; /* UVsystab problem, abort UV init */ + return; + build_socket_tables(); build_uv_gr_table(); uv_init_hub_info(&hub_info); @@ -1393,14 +1385,10 @@ void __init uv_system_init(void) if (!_node_to_pnode) boot_init_possible_blades(&hub_info); - /* uv_num_possible_blades() is really the hub count */ - pr_info("UV: Found %d hubs, %d nodes, %d cpus\n", - uv_num_possible_blades(), - num_possible_nodes(), - num_possible_cpus()); + /* uv_num_possible_blades() is really the hub count: */ + pr_info("UV: Found %d hubs, %d nodes, %d CPUs\n", uv_num_possible_blades(), num_possible_nodes(), num_possible_cpus()); - uv_bios_get_sn_info(0, &uv_type, &sn_partition_id, &sn_coherency_id, - &sn_region_size, &system_serial_number); + uv_bios_get_sn_info(0, &uv_type, &sn_partition_id, &sn_coherency_id, &sn_region_size, &system_serial_number); hub_info.coherency_domain_number = sn_coherency_id; uv_rtc_init(); @@ -1413,33 +1401,31 @@ void __init uv_system_init(void) struct uv_hub_info_s *new_hub; if (__uv_hub_info_list[nodeid]) { - pr_err("UV: Node %d UV HUB already initialized!?\n", - nodeid); + pr_err("UV: Node %d UV HUB already initialized!?\n", nodeid); BUG(); } /* Allocate new per hub info list */ - new_hub = (nodeid == 0) ? - &uv_hub_info_node0 : - kzalloc_node(bytes, GFP_KERNEL, nodeid); + new_hub = (nodeid == 0) ? &uv_hub_info_node0 : kzalloc_node(bytes, GFP_KERNEL, nodeid); BUG_ON(!new_hub); __uv_hub_info_list[nodeid] = new_hub; new_hub = uv_hub_info_list(nodeid); BUG_ON(!new_hub); *new_hub = hub_info; - /* Use information from GAM table if available */ + /* Use information from GAM table if available: */ if (_node_to_pnode) new_hub->pnode = _node_to_pnode[nodeid]; - else /* Fill in during cpu loop */ + else /* Or fill in during CPU loop: */ new_hub->pnode = 0xffff; + new_hub->numa_blade_id = uv_node_to_blade_id(nodeid); new_hub->memory_nid = -1; new_hub->nr_possible_cpus = 0; new_hub->nr_online_cpus = 0; } - /* Initialize per cpu info */ + /* Initialize per CPU info: */ for_each_possible_cpu(cpu) { int apicid = per_cpu(x86_cpu_to_apicid, cpu); int numa_node_id; @@ -1450,22 +1436,24 @@ void __init uv_system_init(void) pnode = uv_apicid_to_pnode(apicid); uv_cpu_info_per(cpu)->p_uv_hub_info = uv_hub_info_list(nodeid); - uv_cpu_info_per(cpu)->blade_cpu_id = - uv_cpu_hub_info(cpu)->nr_possible_cpus++; + uv_cpu_info_per(cpu)->blade_cpu_id = uv_cpu_hub_info(cpu)->nr_possible_cpus++; if (uv_cpu_hub_info(cpu)->memory_nid == -1) uv_cpu_hub_info(cpu)->memory_nid = cpu_to_node(cpu); - if (nodeid != numa_node_id && /* init memoryless node */ + + /* Init memoryless node: */ + if (nodeid != numa_node_id && uv_hub_info_list(numa_node_id)->pnode == 0xffff) uv_hub_info_list(numa_node_id)->pnode = pnode; else if (uv_cpu_hub_info(cpu)->pnode == 0xffff) uv_cpu_hub_info(cpu)->pnode = pnode; + uv_cpu_scir_info(cpu)->offset = uv_scir_offset(apicid); } for_each_node(nodeid) { unsigned short pnode = uv_hub_info_list(nodeid)->pnode; - /* Add pnode info for pre-GAM list nodes without cpus */ + /* Add pnode info for pre-GAM list nodes without CPUs: */ if (pnode == 0xffff) { unsigned long paddr; @@ -1491,12 +1479,12 @@ void __init uv_system_init(void) uv_scir_register_cpu_notifier(); proc_mkdir("sgi_uv", NULL); - /* register Legacy VGA I/O redirection handler */ + /* Register Legacy VGA I/O redirection handler: */ pci_register_set_vga_state(uv_set_vga_state); /* * For a kdump kernel the reset must be BOOT_ACPI, not BOOT_EFI, as - * EFI is not enabled in the kdump kernel. + * EFI is not enabled in the kdump kernel: */ if (is_kdump_kernel()) reboot_type = BOOT_ACPI; ^ permalink raw reply related [flat|nested] 15+ messages in thread
* [PATCH 5/8] x86/platform/UV: Add basic CPU NMI health check 2017-01-13 15:21 [PATCH 0/8] x86/platform/UV: UV Update PatchSet 2 Mike Travis ` (3 preceding siblings ...) 2017-01-13 15:21 ` [PATCH 4/8] x86/platform/UV: Add Support for UV4 Hubless NMIs Mike Travis @ 2017-01-13 15:21 ` Mike Travis 2017-01-13 15:21 ` [PATCH 6/8] x86/platform/UV: Verify NMI action is valid, default is standard Mike Travis ` (2 subsequent siblings) 7 siblings, 0 replies; 15+ messages in thread From: Mike Travis @ 2017-01-13 15:21 UTC (permalink / raw) To: Ingo Molnar, Thomas Gleixner, H. Peter Anvin Cc: x86, linux-kernel, Dimitri Sivanich [-- Attachment #1: uv4_add-health-check --] [-- Type: text/plain, Size: 2005 bytes --] Add a low impact health check triggered by the system NMI command that essentially checks which CPUs are responding to external NMI's. Signed-off-by: Mike Travis <travis@sgi.com> Acked-by: Dimitri Sivanich <sivanich@hpe.com> Reviewed-by: Russ Anderson <rja@hpe.com> Reviewed-by: Alex Thorlton <athorlton@sgi.com> --- arch/x86/platform/uv/uv_nmi.c | 23 ++++++++++++++++++++++- 1 file changed, 22 insertions(+), 1 deletion(-) --- linux.orig/arch/x86/platform/uv/uv_nmi.c +++ linux/arch/x86/platform/uv/uv_nmi.c @@ -177,6 +177,7 @@ module_param_named(debug, uv_nmi_debug, * "kdump" - do crash dump * "kdb" - enter KDB (default) * "kgdb" - enter KGDB + * "health" - check if CPUs respond to NMI */ static char uv_nmi_action[8] = "kdb"; module_param_string(action, uv_nmi_action, sizeof(uv_nmi_action), 0644); @@ -573,6 +574,22 @@ static void uv_nmi_sync_exit(int master) } } +/* Current "health" check is which CPU's are responsive */ +static void uv_nmi_action_health(int cpu, struct pt_regs *regs, int master) +{ + if (master) { + int in = atomic_read(&uv_nmi_cpus_in_nmi); + int out = num_online_cpus() - in; + + pr_alert("UV: NMI CPU health check (non-responding:%d)\n", out); + atomic_set(&uv_nmi_slave_continue, SLAVE_EXIT); + } else { + while (!atomic_read(&uv_nmi_slave_continue)) + cpu_relax(); + } + uv_nmi_sync_exit(master); +} + /* Walk through cpu list and dump state of each */ static void uv_nmi_dump_state(int cpu, struct pt_regs *regs, int master) { @@ -748,8 +765,12 @@ int uv_handle_nmi(unsigned int reason, s /* Pause as all cpus enter the NMI handler */ uv_nmi_wait(master); + /* Health check for any processors that are 'stuck' */ + if (uv_nmi_action_is("health")) + uv_nmi_action_health(cpu, regs, master); + /* Dump state of each cpu */ - if (uv_nmi_action_is("ips") || uv_nmi_action_is("dump")) + else if (uv_nmi_action_is("ips") || uv_nmi_action_is("dump")) uv_nmi_dump_state(cpu, regs, master); /* Call KGDB/KDB if enabled */ -- ^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH 6/8] x86/platform/UV: Verify NMI action is valid, default is standard 2017-01-13 15:21 [PATCH 0/8] x86/platform/UV: UV Update PatchSet 2 Mike Travis ` (4 preceding siblings ...) 2017-01-13 15:21 ` [PATCH 5/8] x86/platform/UV: Add basic CPU NMI health check Mike Travis @ 2017-01-13 15:21 ` Mike Travis 2017-01-13 15:21 ` [PATCH 7/8] x86/platform/UV: Initialize PCH GPP_D_0 NMI Pin to be NMI source Mike Travis 2017-01-13 15:21 ` [PATCH 8/8] x86/platform/UV: Insure uv_system_init is called when necessary Mike Travis 7 siblings, 0 replies; 15+ messages in thread From: Mike Travis @ 2017-01-13 15:21 UTC (permalink / raw) To: Ingo Molnar, Thomas Gleixner, H. Peter Anvin Cc: x86, linux-kernel, Dimitri Sivanich [-- Attachment #1: uv4_add-check-action --] [-- Type: text/plain, Size: 2989 bytes --] Verify that the NMI action being set is valid. The default NMI action changes from the non-standard 'kdb' to the more standard 'dump'. Signed-off-by: Mike Travis <travis@sgi.com> Acked-by: Dimitri Sivanich <sivanich@hpe.com> Reviewed-by: Russ Anderson <rja@hpe.com> Reviewed-by: Alex Thorlton <athorlton@sgi.com> --- Note: this patch fails checkpatch with the following error: WARNING: do not add new typedefs #44: FILE: arch/x86/platform/uv/uv_nmi.c:186: +typedef char action_t[ACTION_LEN]; But I do not know another way to create the correct parameter for this check: +#define param_check_action(name, p) __param_check(name, p, action_t) --- arch/x86/platform/uv/uv_nmi.c | 69 +++++++++++++++++++++++++++++++++++------- 1 file changed, 58 insertions(+), 11 deletions(-) --- linux.orig/arch/x86/platform/uv/uv_nmi.c +++ linux/arch/x86/platform/uv/uv_nmi.c @@ -170,17 +170,64 @@ module_param_named(debug, uv_nmi_debug, pr_info(fmt, ##__VA_ARGS__); \ } while (0) -/* - * Valid NMI Actions: - * "dump" - dump process stack for each cpu - * "ips" - dump IP info for each cpu - * "kdump" - do crash dump - * "kdb" - enter KDB (default) - * "kgdb" - enter KGDB - * "health" - check if CPUs respond to NMI - */ -static char uv_nmi_action[8] = "kdb"; -module_param_string(action, uv_nmi_action, sizeof(uv_nmi_action), 0644); +/* Valid NMI Actions */ +#define ACTION_LEN 16 +static struct nmi_action { + char *action; + char *desc; +} valid_acts[] = { + { "kdump", "do kernel crash dump" }, + { "dump", "dump process stack for each cpu" }, + { "ips", "dump Inst Ptr info for each cpu" }, + { "kdb", "enter KDB (needs kgdboc= assignment)" }, + { "kgdb", "enter KGDB (needs gdb target remote)" }, + { "health", "check if CPUs respond to NMI" }, +}; +typedef char action_t[ACTION_LEN]; +static action_t uv_nmi_action = { "dump" }; + +static int param_get_action(char *buffer, const struct kernel_param *kp) +{ + return sprintf(buffer, "%s\n", uv_nmi_action); +} + +static int param_set_action(const char *val, const struct kernel_param *kp) +{ + int i; + int n = ARRAY_SIZE(valid_acts); + char arg[ACTION_LEN], *p; + + /* (remove possible '\n') */ + strncpy(arg, val, ACTION_LEN - 1); + arg[ACTION_LEN - 1] = '\0'; + p = strchr(arg, '\n'); + if (p) + *p = '\0'; + + for (i = 0; i < n; i++) + if (!strcmp(arg, valid_acts[i].action)) + break; + + if (i < n) { + strcpy(uv_nmi_action, arg); + pr_info("UV: New NMI action:%s\n", uv_nmi_action); + return 0; + } + + pr_err("UV: Invalid NMI action:%s, valid actions are:\n", arg); + for (i = 0; i < n; i++) + pr_err("UV: %-8s - %s\n", + valid_acts[i].action, valid_acts[i].desc); + return -EINVAL; +} + +static const struct kernel_param_ops param_ops_action = { + .get = param_get_action, + .set = param_set_action, +}; +#define param_check_action(name, p) __param_check(name, p, action_t) + +module_param_named(action, uv_nmi_action, action, 0644); static inline bool uv_nmi_action_is(const char *action) { -- ^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH 7/8] x86/platform/UV: Initialize PCH GPP_D_0 NMI Pin to be NMI source 2017-01-13 15:21 [PATCH 0/8] x86/platform/UV: UV Update PatchSet 2 Mike Travis ` (5 preceding siblings ...) 2017-01-13 15:21 ` [PATCH 6/8] x86/platform/UV: Verify NMI action is valid, default is standard Mike Travis @ 2017-01-13 15:21 ` Mike Travis 2017-01-13 15:21 ` [PATCH 8/8] x86/platform/UV: Insure uv_system_init is called when necessary Mike Travis 7 siblings, 0 replies; 15+ messages in thread From: Mike Travis @ 2017-01-13 15:21 UTC (permalink / raw) To: Ingo Molnar, Thomas Gleixner, H. Peter Anvin Cc: x86, linux-kernel, Dimitri Sivanich [-- Attachment #1: uv4_add-hubless-nmi-init-pch --] [-- Type: text/plain, Size: 4823 bytes --] The initialize PCH NMI I/O function is separate and may be moved to BIOS for security reasons. This function detects whether the PCH NMI config has already been done and if not, it will then initialize the PCH here. Signed-off-by: Mike Travis <travis@sgi.com> Acked-by: Dimitri Sivanich <sivanich@hpe.com> Reviewed-by: Russ Anderson <rja@hpe.com> --- arch/x86/platform/uv/uv_nmi.c | 128 ++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 128 insertions(+) --- linux.orig/arch/x86/platform/uv/uv_nmi.c +++ linux/arch/x86/platform/uv/uv_nmi.c @@ -71,6 +71,7 @@ EXPORT_PER_CPU_SYMBOL_GPL(uv_cpu_nmi); static int *pch_base; #define NMI_CONTROL_PORT 0x70 #define NMI_DUMMY_PORT 0x71 +#define PAD_OWN_GPP_D_0 0x2c #define GPI_NMI_STS_GPP_D_0 0x164 #define GPI_NMI_ENA_GPP_D_0 0x174 #define STS_GPP_D_0_MASK 0x1 @@ -161,6 +162,9 @@ static bool uv_pch_intr_enable = true; static bool uv_pch_intr_now_enabled; module_param_named(pch_intr_enable, uv_pch_intr_enable, bool, 0644); +static bool uv_pch_init_enable = true; +module_param_named(pch_init_enable, uv_pch_init_enable, bool, 0644); + static int uv_nmi_debug; module_param_named(debug, uv_nmi_debug, int, 0644); @@ -309,6 +313,127 @@ static void uv_nmi_setup_hubless_intr(vo uv_pch_intr_now_enabled ? "enabled" : "disabled"); } +static struct init_nmi { + unsigned int offset; + unsigned int mask; + unsigned int data; +} init_nmi[] = { + { /* HOSTSW_OWN_GPP_D_0 */ + .offset = 0x84, + .mask = 0x1, + .data = 0x0, /* ACPI Mode */ + }, + +/* clear status */ + { /* GPI_INT_STS_GPP_D_0 */ + .offset = 0x104, + .mask = 0x0, + .data = 0x1, /* Clear Status */ + }, + { /* GPI_GPE_STS_GPP_D_0 */ + .offset = 0x124, + .mask = 0x0, + .data = 0x1, /* Clear Status */ + }, + { /* GPI_SMI_STS_GPP_D_0 */ + .offset = 0x144, + .mask = 0x0, + .data = 0x1, /* Clear Status */ + }, + { /* GPI_NMI_STS_GPP_D_0 */ + .offset = 0x164, + .mask = 0x0, + .data = 0x1, /* Clear Status */ + }, + +/* disable interrupts */ + { /* GPI_INT_EN_GPP_D_0 */ + .offset = 0x114, + .mask = 0x1, + .data = 0x0, /* disable interrupt generation */ + }, + { /* GPI_GPE_EN_GPP_D_0 */ + .offset = 0x134, + .mask = 0x1, + .data = 0x0, /* disable interrupt generation */ + }, + { /* GPI_SMI_EN_GPP_D_0 */ + .offset = 0x154, + .mask = 0x1, + .data = 0x0, /* disable interrupt generation */ + }, + { /* GPI_NMI_EN_GPP_D_0 */ + .offset = 0x174, + .mask = 0x1, + .data = 0x0, /* disable interrupt generation */ + }, + +/* setup GPP_D_0 Pad Config */ + { /* PAD_CFG_DW0_GPP_D_0 */ + .offset = 0x4c0, + .mask = 0xffffffff, + .data = 0x82020100, +/* + * 31:30 Pad Reset Config (PADRSTCFG): = 2h # PLTRST# (default) + * + * 29 RX Pad State Select (RXPADSTSEL): = 0 # Raw RX pad state directly + * from RX buffer (default) + * + * 28 RX Raw Override to '1' (RXRAW1): = 0 # No Override + * + * 26:25 RX Level/Edge Configuration (RXEVCFG): + * = 0h # Level + * = 1h # Edge + * + * 23 RX Invert (RXINV): = 0 # No Inversion (signal active high) + * + * 20 GPIO Input Route IOxAPIC (GPIROUTIOXAPIC): + * = 0 # Routing does not cause peripheral IRQ... + * # (we want an NMI not an IRQ) + * + * 19 GPIO Input Route SCI (GPIROUTSCI): = 0 # Routing does not cause SCI. + * 18 GPIO Input Route SMI (GPIROUTSMI): = 0 # Routing does not cause SMI. + * 17 GPIO Input Route NMI (GPIROUTNMI): = 1 # Routing can cause NMI. + * + * 11:10 Pad Mode (PMODE1/0): = 0h = GPIO control the Pad. + * 9 GPIO RX Disable (GPIORXDIS): + * = 0 # Enable the input buffer (active low enable) + * + * 8 GPIO TX Disable (GPIOTXDIS): + * = 1 # Disable the output buffer; i.e. Hi-Z + * + * 1 GPIO RX State (GPIORXSTATE): This is the current internal RX pad state.. + * 0 GPIO TX State (GPIOTXSTATE): + * = 0 # (Leave at default) + */ + }, + +/* Pad Config DW1 */ + { /* PAD_CFG_DW1_GPP_D_0 */ + .offset = 0x4c4, + .mask = 0x3c00, + .data = 0, /* Termination = none (default) */ + }, +}; + +static void uv_init_hubless_pch_d0(void) +{ + int i, read; + + read = *PCH_PCR_GPIO_ADDRESS(PAD_OWN_GPP_D_0); + if (read != 0) { + pr_info("UV: Hubless NMI already configured\n"); + return; + } + + nmi_debug("UV: Initializing UV Hubless NMI on PCH\n"); + for (i = 0; i < ARRAY_SIZE(init_nmi); i++) { + uv_init_hubless_pch_io(init_nmi[i].offset, + init_nmi[i].mask, + init_nmi[i].data); + } +} + static int uv_nmi_test_hubless(struct uv_hub_nmi_s *hub_nmi) { int *pstat = PCH_PCR_GPIO_ADDRESS(GPI_NMI_STS_GPP_D_0); @@ -939,6 +1064,9 @@ void __init uv_nmi_setup_hubless(void) nmi_debug("UV: PCH base:%p from 0x%lx, GPP_D_0\n", pch_base, PCH_PCR_GPIO_1_BASE); + if (uv_pch_init_enable) + uv_init_hubless_pch_d0(); + uv_init_hubless_pch_io(GPI_NMI_ENA_GPP_D_0, STS_GPP_D_0_MASK, STS_GPP_D_0_MASK); -- ^ permalink raw reply [flat|nested] 15+ messages in thread
* [PATCH 8/8] x86/platform/UV: Insure uv_system_init is called when necessary 2017-01-13 15:21 [PATCH 0/8] x86/platform/UV: UV Update PatchSet 2 Mike Travis ` (6 preceding siblings ...) 2017-01-13 15:21 ` [PATCH 7/8] x86/platform/UV: Initialize PCH GPP_D_0 NMI Pin to be NMI source Mike Travis @ 2017-01-13 15:21 ` Mike Travis 7 siblings, 0 replies; 15+ messages in thread From: Mike Travis @ 2017-01-13 15:21 UTC (permalink / raw) To: Ingo Molnar, Thomas Gleixner, H. Peter Anvin Cc: x86, linux-kernel, Dimitri Sivanich [-- Attachment #1: uv4_call_uv_system_init --] [-- Type: text/plain, Size: 964 bytes --] Move the check to whether this is a UV system that needs initialization from is_uv_system() to the internal uv_system_init() function. This is because on a UV system without a HUB the is_uv_system() returns false. But we still need some specific UV system initialization. See the uv_system_init() for change to a quick check if UV is applicable. This change should not increase overhead since is_uv_system() also called into this same area. Signed-off-by: Mike Travis <travis@sgi.com> Acked-by: Dimitri Sivanich <sivanich@hpe.com> Reviewed-by: Russ Anderson <rja@hpe.com> --- arch/x86/kernel/smpboot.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) --- linux.orig/arch/x86/kernel/smpboot.c +++ linux/arch/x86/kernel/smpboot.c @@ -1341,8 +1341,7 @@ void __init native_smp_prepare_cpus(unsi pr_info("CPU0: "); print_cpu_info(&cpu_data(0)); - if (is_uv_system()) - uv_system_init(); + uv_system_init(); set_mtrr_aps_delayed_init(); -- ^ permalink raw reply [flat|nested] 15+ messages in thread
end of thread, other threads:[~2017-02-01 10:10 UTC | newest] Thread overview: 15+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2017-01-13 15:21 [PATCH 0/8] x86/platform/UV: UV Update PatchSet 2 Mike Travis 2017-01-13 15:21 ` [PATCH 1/8] x86/platform/UV: Fix panic with missing UVsystab support Mike Travis 2017-01-14 9:57 ` [tip:x86/platform] " tip-bot for Mike Travis 2017-01-13 15:21 ` [PATCH 2/8] x86/platform/UV: Fix 2 socket config problem Mike Travis 2017-01-14 9:58 ` [tip:x86/platform] " tip-bot for Mike Travis 2017-01-13 15:21 ` [PATCH 3/8] x86/platform/UV: Add Support for UV4 Hubless systems Mike Travis 2017-01-13 15:21 ` [PATCH 4/8] x86/platform/UV: Add Support for UV4 Hubless NMIs Mike Travis 2017-01-14 7:52 ` Ingo Molnar 2017-01-14 8:26 ` [PATCH] x86/platform/UV: Clean up the UV APIC code Ingo Molnar 2017-01-14 13:11 ` Joe Perches 2017-02-01 10:09 ` [tip:x86/platform] " tip-bot for Ingo Molnar 2017-01-13 15:21 ` [PATCH 5/8] x86/platform/UV: Add basic CPU NMI health check Mike Travis 2017-01-13 15:21 ` [PATCH 6/8] x86/platform/UV: Verify NMI action is valid, default is standard Mike Travis 2017-01-13 15:21 ` [PATCH 7/8] x86/platform/UV: Initialize PCH GPP_D_0 NMI Pin to be NMI source Mike Travis 2017-01-13 15:21 ` [PATCH 8/8] x86/platform/UV: Insure uv_system_init is called when necessary Mike Travis
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.