x86: simpify direct_gbpages setting, add early_param_on_off()
cover x86: simpify direct_gbpages setting, add early_param_on_off()
From: "Luis R. Rodriguez" <[email protected]>
While reviewing DIRECT_GBPAGES I noticed some inconsistancies on when its cached "enable" variable is set and when in reality its full feature is enabled. There's also some code simplification possible by simply using Kconfig, so decided to take advantage of that and then generalize this a bit under a new early_param_on_off() which makes it easier to associate featureas with respective cached values and their own feature related Kconfig entries.
In this case in order to use early_param_on_off() given that init.o is shared with both 32-bit and 64-bit and before the early parameters were only exposed on 64-bit it does mean we're exposing this kernel parameter to 32-bit as well now but its not possible to enable it on 32-bit as its real setting is guarded by ENABLE_DIRECT_GBPAGES.
I was tempted to grep around for this sort of grammar use but couldn't easily find all uses with Coccinelle / grep due to the #ifder'y and the code being outside of routines. If the generalization is not desired we can drop the last two, but would ask that the first two patches to be considered and reviewed.
Luis R. Rodriguez (4):
x86: mm: use IS_ENABLED() for direct_gbpages
x86: mm: simplify enabling direct_gbpages init.h: add early_param_on_off()
x86: mm: use early_param_on_off() for direct_gbpages
arch/x86/Kconfig | 18 +++++++++++++----- arch/x86/mm/init.c | 24 +++++++++++------------- arch/x86/mm/init_64.c | 14 -------------- arch/x86/mm/pageattr.c | 2 -- include/linux/init.h | 15 +++++++++++++++ 5 files changed, 39 insertions(+), 34 deletions(-)
Signature · 1 line
2.2.2
RFC 1/4 x86: mm: use IS_ENABLED() for direct_gbpages
From: "Luis R. Rodriguez" <[email protected]>
Replace #ifdef eyesore with IS_ENABLED() use.
Cc Greg Kroah-Hartman <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Dave Hansen <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Xishi Qiu <[email protected]>
Cc Andrew Morton <[email protected]>
Cc David Vrabel <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Dexuan Cui <[email protected]>
Signed-off-by Luis R. Rodriguez <[email protected]>
--- arch/x86/mm/init.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-)
arch/x86/mm/init.c +1 -5
@@ -131,11 +131,7 @@ void __init early_alloc_pgt_buf(void)int after_bootmem; -int direct_gbpages -#ifdef CONFIG_DIRECT_GBPAGES - =1 -#endif -; +int direct_gbpages =IS_ENABLED (CONFIG_DIRECT_GBPAGES );static void __init init_gbpages (void ) {
Signature · 1 line
2.2.2
tip-bot for Luis R. Rodriguez
merge bot[tip:x86/mm] x86/mm: Use IS_ENABLED() for direct_gbpages
x86/mm: Use IS_ENABLED() for direct_gbpages
Commit-ID: d9fd579c218e22c897f0f1b9e132af9b436cf445 Gitweb: http://git.kernel.org/tip/d9fd579c218e22c897f0f1b9e132af9b436cf445 Author: Luis R. Rodriguez <[email protected]> AuthorDate: Wed, 4 Mar 2015 17:24:11 -0800 Committer: Ingo Molnar <[email protected]> CommitDate: Thu, 5 Mar 2015 08:02:11 +0100
x86/mm: Use IS_ENABLED() for direct_gbpages
Replace #ifdef eyesore with IS_ENABLED() use.
Signed-off-by Luis R. Rodriguez <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Dave Hansen <[email protected]>
Cc David Vrabel <[email protected]>
Cc Dexuan Cui <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc H. Peter Anvin <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Xishi Qiu <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
--- arch/x86/mm/init.c | 6 +----- 1 file changed, 1 insertion(+), 5 deletions(-)
arch/x86/mm/init.c +1 -5
@@ -131,11 +131,7 @@ void __init early_alloc_pgt_buf(void)int after_bootmem; -int direct_gbpages -#ifdef CONFIG_DIRECT_GBPAGES - =1 -#endif -; +int direct_gbpages =IS_ENABLED (CONFIG_DIRECT_GBPAGES );static void __init init_gbpages (void ) {
RFC 2/4 x86: mm: simplify enabling direct_gbpages
In reply to Luis R. RodriguezFrom: "Luis R. Rodriguez" <[email protected]>
direct_gbpages can be force enabled as an early parameter but not really have taken effect when DEBUG_PAGEALLOC or KMEMCHECK is enabled. You can also enable direct_gbpages right now if you have an x86_64 architecture but your CPU doesn't really have support for this feature. In both cases PG_LEVEL_1G won't actually be enabled but direct_gbpages is used in other areas under the assumptions that PG_LEVEL_1G was set. Fix this by putting together all requirements which make this feature sensible to enable under, and only enable both finally flipping on PG_LEVEL_1G and leaving PG_LEVEL_1G set when this is true.
We only enable this feature then to be possible on sensible builds defined by the new ENABLE_DIRECT_GBPAGES. If the CPU has support for it you can either enable this by using the DIRECT_GBPAGES option or using the early kernel parameter. If a platform had support for this you can always force disable it as well.
Cc Greg Kroah-Hartman <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Dave Hansen <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Xishi Qiu <[email protected]>
Cc Andrew Morton <[email protected]>
Cc David Vrabel <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Dexuan Cui <[email protected]>
Signed-off-by Luis R. Rodriguez <[email protected]>
--- arch/x86/Kconfig | 18 +++++++++++++----- arch/x86/mm/init.c | 17 +++++++++-------- arch/x86/mm/pageattr.c | 2 -- 3 files changed, 22 insertions(+), 15 deletions(-)
arch/x86/Kconfig +13 -5
@@ -1300,14 +1300,22 @@ config ARCH_DMA_ADDR_T_64BIT def_bool y depends on X86_64 || HIGHMEM64G +config ENABLE_DIRECT_GBPAGES + def_bool y + depends on X86_64 && !DEBUG_PAGEALLOC && !KMEMCHECK + config DIRECT_GBPAGES bool "Enable 1GB pages for kernel pagetables" if EXPERT default y - depends on X86_64 - ---help--- - Allow the kernel linear mapping to use 1GB pages on CPUs that - support it. This can improve the kernel's performance a tiny bit by - reducing TLB pressure. If in doubt, say "Y". + depends on ENABLE_DIRECT_GBPAGES + ---help--- + Enable by default the kernel linear mapping to use 1GB pages on CPUs + that support it. This can improve the kernel's performance a tiny bit + by reducing TLB pressure. If in doubt, say "Y". If you've disabled + option but your platform is capable of handling support for this + you can use the gbpages kernel parameter. Likewise if you've enabled + this but you'd like to force disable this option you can use the + nogbpages kernel parameter. # Common NUMA Features config NUMA
arch/x86/mm/init.c +9 -8
@@ -131,16 +131,21 @@ void __init early_alloc_pgt_buf(void)int after_bootmem; +static int page_size_mask; +int direct_gbpages =IS_ENABLED (CONFIG_DIRECT_GBPAGES );static void __init init_gbpages (void ) { -#ifdef CONFIG_X86_64 -if (direct_gbpages && cpu_has_gbpages) +if (!IS_ENABLED (CONFIG_ENABLE_DIRECT_GBPAGES )) { + direct_gbpages =0 ; +return ; + } +if (direct_gbpages && cpu_has_gbpages) {printk (KERN_INFO "Using GB pages for direct mapping\n" ); -else + page_size_mask |=1 <<PG_LEVEL_1G ; + }else direct_gbpages =0 ; -#endif }struct map_range {
@@ -149,8 +154,6 @@ struct map_range {unsigned page_size_mask; }; -static int page_size_mask; -static void __init probe_page_size_mask (void ) {init_gbpages ();
@@ -161,8 +164,6 @@ static void __init probe_page_size_mask(void)* This will simplify cpa(), which otherwise needs to support splitting * large pages into small in interrupt context, etc. */ -if (direct_gbpages) - page_size_mask |=1 <<PG_LEVEL_1G ;if (cpu_has_pse) page_size_mask |=1 <<PG_LEVEL_2M ;#endif
arch/x86/mm/pageattr.c +0 -2
@@ -81,11 +81,9 @@ void arch_report_meminfo(struct seq_file *m)seq_printf (m,"DirectMap4M: %8lu kB\n" , direct_pages_count[PG_LEVEL_2M ] <<12 );#endif -#ifdef CONFIG_X86_64 if (direct_gbpages)seq_printf (m,"DirectMap1G: %8lu kB\n" , direct_pages_count[PG_LEVEL_1G ] <<20 ); -#endif }#else static inline void split_page_count (int level) { }
Signature · 1 line
2.2.2
tip-bot for Luis R. Rodriguez
merge bot[tip:x86/mm] x86/mm: Simplify enabling direct_gbpages
x86/mm: Simplify enabling direct_gbpages
Commit-ID: e5008abe929c160d36e44b8c2b644d4330d2e389 Gitweb: http://git.kernel.org/tip/e5008abe929c160d36e44b8c2b644d4330d2e389 Author: Luis R. Rodriguez <[email protected]> AuthorDate: Wed, 4 Mar 2015 17:24:12 -0800 Committer: Ingo Molnar <[email protected]> CommitDate: Thu, 5 Mar 2015 08:02:12 +0100
x86/mm: Simplify enabling direct_gbpages
direct_gbpages can be force enabled as an early parameter but not really have taken effect when DEBUG_PAGEALLOC or KMEMCHECK is enabled. You can also enable direct_gbpages right now if you have an x86_64 architecture but your CPU doesn't really have support for this feature. In both cases PG_LEVEL_1G won't actually be enabled but direct_gbpages is used in other areas under the assumptions that PG_LEVEL_1G was set. Fix this by putting together all requirements which make this feature sensible to enable under, and only enable both finally flipping on PG_LEVEL_1G and leaving PG_LEVEL_1G set when this is true.
We only enable this feature then to be possible on sensible builds defined by the new ENABLE_DIRECT_GBPAGES. If the CPU has support for it you can either enable this by using the DIRECT_GBPAGES option or using the early kernel parameter. If a platform had support for this you can always force disable it as well.
Signed-off-by Luis R. Rodriguez <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Dave Hansen <[email protected]>
Cc David Vrabel <[email protected]>
Cc Dexuan Cui <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc H. Peter Anvin <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Xishi Qiu <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
--- arch/x86/Kconfig | 18 +++++++++++++----- arch/x86/mm/init.c | 17 +++++++++-------- arch/x86/mm/pageattr.c | 2 -- 3 files changed, 22 insertions(+), 15 deletions(-)
arch/x86/Kconfig +13 -5
@@ -1299,14 +1299,22 @@ config ARCH_DMA_ADDR_T_64BIT def_bool y depends on X86_64 || HIGHMEM64G +config ENABLE_DIRECT_GBPAGES + def_bool y + depends on X86_64 && !DEBUG_PAGEALLOC && !KMEMCHECK + config DIRECT_GBPAGES bool "Enable 1GB pages for kernel pagetables" if EXPERT default y - depends on X86_64 - ---help--- - Allow the kernel linear mapping to use 1GB pages on CPUs that - support it. This can improve the kernel's performance a tiny bit by - reducing TLB pressure. If in doubt, say "Y". + depends on ENABLE_DIRECT_GBPAGES + ---help--- + Enable by default the kernel linear mapping to use 1GB pages on CPUs + that support it. This can improve the kernel's performance a tiny bit + by reducing TLB pressure. If in doubt, say "Y". If you've disabled + option but your platform is capable of handling support for this + you can use the gbpages kernel parameter. Likewise if you've enabled + this but you'd like to force disable this option you can use the + nogbpages kernel parameter. # Common NUMA Features config NUMA
arch/x86/mm/init.c +9 -8
@@ -131,16 +131,21 @@ void __init early_alloc_pgt_buf(void)int after_bootmem; +static int page_size_mask; +int direct_gbpages =IS_ENABLED (CONFIG_DIRECT_GBPAGES );static void __init init_gbpages (void ) { -#ifdef CONFIG_X86_64 -if (direct_gbpages && cpu_has_gbpages) +if (!IS_ENABLED (CONFIG_ENABLE_DIRECT_GBPAGES )) { + direct_gbpages =0 ; +return ; + } +if (direct_gbpages && cpu_has_gbpages) {printk (KERN_INFO "Using GB pages for direct mapping\n" ); -else + page_size_mask |=1 <<PG_LEVEL_1G ; + }else direct_gbpages =0 ; -#endif }struct map_range {
@@ -149,8 +154,6 @@ struct map_range {unsigned page_size_mask; }; -static int page_size_mask; -static void __init probe_page_size_mask (void ) {init_gbpages ();
@@ -161,8 +164,6 @@ static void __init probe_page_size_mask(void)* This will simplify cpa(), which otherwise needs to support splitting * large pages into small in interrupt context, etc. */ -if (direct_gbpages) - page_size_mask |=1 <<PG_LEVEL_1G ;if (cpu_has_pse) page_size_mask |=1 <<PG_LEVEL_2M ;#endif
arch/x86/mm/pageattr.c +0 -2
@@ -81,11 +81,9 @@ void arch_report_meminfo(struct seq_file *m)seq_printf (m,"DirectMap4M: %8lu kB\n" , direct_pages_count[PG_LEVEL_2M ] <<12 );#endif -#ifdef CONFIG_X86_64 if (direct_gbpages)seq_printf (m,"DirectMap1G: %8lu kB\n" , direct_pages_count[PG_LEVEL_1G ] <<20 ); -#endif }#else static inline void split_page_count (int level) { }
RFC 3/4 init.h: add early_param_on_off()
In reply to Luis R. RodriguezFrom: "Luis R. Rodriguez" <[email protected]>
At times all you need is a kconfig variable to enable a feature, by default but you may want to also enable / disable it through a kernel parameter. In such cases the parameter routines to turn the thing on / off are really simple. Just use a wrapper for this, it lets us generalize the code and makes it easier to associate parameters with related kernel configuration options.
Cc Greg Kroah-Hartman <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Dave Hansen <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Xishi Qiu <[email protected]>
Cc Andrew Morton <[email protected]>
Cc David Vrabel <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Dexuan Cui <[email protected]>
Signed-off-by Luis R. Rodriguez <[email protected]>
--- include/linux/init.h | 15 +++++++++++++++ 1 file changed, 15 insertions(+)
include/linux/init.h +15 -0
@@ -268,6 +268,21 @@ struct obs_kernel_param {#define early_param (str, fn) \ __setup_param(str, fn, fn, 1) +#define early_param_on_off (str_on, str_off, var, config) \ + int var = IS_ENABLED(config); \ + static int __init parse_##var##_on(char *arg) \ + { \ + var = 1; \ + return 0; \ + } \ + static int __init parse_##var##_off(char *arg) \ + { \ + var = 0; \ + return 0; \ + } \ + __setup_param(str_on, parse_##var##_on, parse_##var##_on, 1); \ + __setup_param(str_off, parse_##var##_off, parse_##var##_off, 1) +/* Relies on boot_command_line being set */ void __init parse_early_param (void );void __init parse_early_options (char *cmdline);
Signature · 1 line
2.2.2
tip-bot for Luis R. Rodriguez
merge bot[tip:x86/mm] init.h: Add early_param_on_off()
init.h: Add early_param_on_off()
Commit-ID: bfb33bad83f650f265ed65cbfe8352b7c3ce8c76 Gitweb: http://git.kernel.org/tip/bfb33bad83f650f265ed65cbfe8352b7c3ce8c76 Author: Luis R. Rodriguez <[email protected]> AuthorDate: Wed, 4 Mar 2015 17:24:13 -0800 Committer: Ingo Molnar <[email protected]> CommitDate: Thu, 5 Mar 2015 08:02:12 +0100
init.h: Add early_param_on_off()
At times all you need is a kconfig variable to enable a feature, by default but you may want to also enable / disable it through a kernel parameter. In such cases the parameter routines to turn the thing on / off are really simple. Just use a wrapper for this, it lets us generalize the code and makes it easier to associate parameters with related kernel configuration options.
Signed-off-by Luis R. Rodriguez <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Dave Hansen <[email protected]>
Cc David Vrabel <[email protected]>
Cc Dexuan Cui <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc H. Peter Anvin <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Xishi Qiu <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
--- include/linux/init.h | 15 +++++++++++++++ 1 file changed, 15 insertions(+)
include/linux/init.h +15 -0
@@ -268,6 +268,21 @@ struct obs_kernel_param {#define early_param (str, fn) \ __setup_param(str, fn, fn, 1) +#define early_param_on_off (str_on, str_off, var, config) \ + int var = IS_ENABLED(config); \ + static int __init parse_##var##_on(char *arg) \ + { \ + var = 1; \ + return 0; \ + } \ + static int __init parse_##var##_off(char *arg) \ + { \ + var = 0; \ + return 0; \ + } \ + __setup_param(str_on, parse_##var##_on, parse_##var##_on, 1); \ + __setup_param(str_off, parse_##var##_off, parse_##var##_off, 1) +/* Relies on boot_command_line being set */ void __init parse_early_param (void );void __init parse_early_options (char *cmdline);
RFC 4/4 x86: mm: use early_param_on_off() for direct_gbpages
In reply to Luis R. RodriguezFrom: "Luis R. Rodriguez" <[email protected]>
The enabler / disabler is pretty simple, just use the provided wrappers, this lets us easily relate the variable to the associated Kconfig entry.
Cc Greg Kroah-Hartman <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc "H. Peter Anvin" <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Dave Hansen <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Xishi Qiu <[email protected]>
Cc Andrew Morton <[email protected]>
Cc David Vrabel <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Dexuan Cui <[email protected]>
Signed-off-by Luis R. Rodriguez <[email protected]>
--- arch/x86/mm/init.c | 3 ++- arch/x86/mm/init_64.c | 14 -------------- 2 files changed, 2 insertions(+), 15 deletions(-)
arch/x86/mm/init.c +2 -1
@@ -133,7 +133,8 @@ int after_bootmem;static int page_size_mask; -int direct_gbpages =IS_ENABLED (CONFIG_DIRECT_GBPAGES ); +early_param_on_off ("gbpages" ,"nogbpages" , + direct_gbpages,CONFIG_DIRECT_GBPAGES );static void __init init_gbpages (void ) {
arch/x86/mm/init_64.c +0 -14
@@ -130,20 +130,6 @@ int kernel_ident_mapping_init(struct x86_mapping_info *info, pgd_t *pgd_page,return 0 ; } -static int __init parse_direct_gbpages_off (char *arg) -{ - direct_gbpages =0 ; -return 0 ; -} -early_param ("nogbpages" , parse_direct_gbpages_off); - -static int __init parse_direct_gbpages_on (char *arg) -{ - direct_gbpages =1 ; -return 0 ; -} -early_param ("gbpages" , parse_direct_gbpages_on); -/* * NOTE: pagetable_init alloc all the fixmap pagetables contiguous on the * physical space so we can cache the place of the first one and move
Signature · 1 line
2.2.2
tip-bot for Luis R. Rodriguez
merge bot[tip:x86/mm] x86/mm: Use early_param_on_off() for direct_gbpages
x86/mm: Use early_param_on_off() for direct_gbpages
Commit-ID: 73c8c861dc5bddf1b24c6aeffee2292c96cf8db2 Gitweb: http://git.kernel.org/tip/73c8c861dc5bddf1b24c6aeffee2292c96cf8db2 Author: Luis R. Rodriguez <[email protected]> AuthorDate: Wed, 4 Mar 2015 17:24:14 -0800 Committer: Ingo Molnar <[email protected]> CommitDate: Thu, 5 Mar 2015 08:02:12 +0100
x86/mm: Use early_param_on_off() for direct_gbpages
The enabler / disabler is pretty simple, just use the provided wrappers, this lets us easily relate the variable to the associated Kconfig entry.
Signed-off-by Luis R. Rodriguez <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Dave Hansen <[email protected]>
Cc David Vrabel <[email protected]>
Cc Dexuan Cui <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc H. Peter Anvin <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Xishi Qiu <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
--- arch/x86/mm/init.c | 3 ++- arch/x86/mm/init_64.c | 14 -------------- 2 files changed, 2 insertions(+), 15 deletions(-)
arch/x86/mm/init.c +2 -1
@@ -133,7 +133,8 @@ int after_bootmem;static int page_size_mask; -int direct_gbpages =IS_ENABLED (CONFIG_DIRECT_GBPAGES ); +early_param_on_off ("gbpages" ,"nogbpages" , + direct_gbpages,CONFIG_DIRECT_GBPAGES );static void __init init_gbpages (void ) {
arch/x86/mm/init_64.c +0 -14
@@ -130,20 +130,6 @@ int kernel_ident_mapping_init(struct x86_mapping_info *info, pgd_t *pgd_page,return 0 ; } -static int __init parse_direct_gbpages_off (char *arg) -{ - direct_gbpages =0 ; -return 0 ; -} -early_param ("nogbpages" , parse_direct_gbpages_off); - -static int __init parse_direct_gbpages_on (char *arg) -{ - direct_gbpages =1 ; -return 0 ; -} -early_param ("gbpages" , parse_direct_gbpages_on); -/* * NOTE: pagetable_init alloc all the fixmap pagetables contiguous on the * physical space so we can cache the place of the first one and move
[PATCH 5/4] x86/mm: Further simplify 1 GB kernel linear mappings handling
In reply to Luis R. RodriguezIt's a bit pointless to allow Kconfig configuration for 1GB kernel mappings, it's already hidden behind a 'default y' and CONFIG_EXPERT.
Remove this complication and simplify the code by renaming CONFIG_ENABLE_DIRECT_GBPAGES to CONFIG_X86_DIRECT_GBPAGES and document the DEBUG_PAGE_ALLOC and KMEMCHECK quirks.
Cc Luis R. Rodriguez <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Dave Hansen <[email protected]>
Cc David Vrabel <[email protected]>
Cc Dexuan Cui <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc H. Peter Anvin <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Xishi Qiu <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
--- arch/x86/Kconfig | 23 +++++++---------------- arch/x86/mm/init.c | 7 +------ 2 files changed, 8 insertions(+), 22 deletions(-)
arch/x86/Kconfig +7 -16
@@ -1299,22 +1299,13 @@ config ARCH_DMA_ADDR_T_64BIT def_bool y depends on X86_64 || HIGHMEM64G -config ENABLE_DIRECT_GBPAGES - def_bool y - depends on X86_64 && !DEBUG_PAGEALLOC && !KMEMCHECK - -config DIRECT_GBPAGES - bool "Enable 1GB pages for kernel pagetables" if EXPERT - default y - depends on ENABLE_DIRECT_GBPAGES - ---help--- - Enable by default the kernel linear mapping to use 1GB pages on CPUs - that support it. This can improve the kernel's performance a tiny bit - by reducing TLB pressure. If in doubt, say "Y". If you've disabled - option but your platform is capable of handling support for this - you can use the gbpages kernel parameter. Likewise if you've enabled - this but you'd like to force disable this option you can use the - nogbpages kernel parameter. +config X86_DIRECT_GBPAGES + def_bool (X86_64 && !DEBUG_PAGEALLOC && !KMEMCHECK) + ---help--- + Certain kernel features effectively disable kernel + linear 1 GB mappings (even if the CPU otherwise + supports them), so don't confuse the user by printing + that we have them enabled. # Common NUMA Features config NUMA
arch/x86/mm/init.c +1 -6
@@ -133,15 +133,10 @@ int after_bootmem;static int page_size_mask; -early_param_on_off ("gbpages" ,"nogbpages" , - direct_gbpages,CONFIG_DIRECT_GBPAGES ); +early_param_on_off ("gbpages" ,"nogbpages" , direct_gbpages,CONFIG_X86_DIRECT_GBPAGES );static void __init init_gbpages (void ) { -if (!IS_ENABLED (CONFIG_ENABLE_DIRECT_GBPAGES )) { - direct_gbpages =0 ; -return ; - }if (direct_gbpages && cpu_has_gbpages) {printk (KERN_INFO "Using GB pages for direct mapping\n" ); page_size_mask |=1 <<PG_LEVEL_1G ;
Re: [PATCH 5/4] x86/mm: Further simplify 1 GB kernel linear mappings handling
Quoted 53 lines
Quoted 1 line
Quoted 1 line · On 05.03.15 at 08:23, <[email protected]> wrote:
It's a bit pointless to allow
Kconfigconfiguration for 1GB kernel mappings, it's already hidden behind a 'default y' andCONFIG_EXPERT.Remove this complication and simplify the code by renaming
CONFIG_ENABLE_DIRECT_GBPAGEStoCONFIG_X86_DIRECT_GBPAGESand document theDEBUG_PAGE_ALLOCand KMEMCHECK quirks.Cc Luis R. Rodriguez <[email protected]>Cc Andrew Morton <[email protected]>Cc Andy Lutomirski <[email protected]>Cc Borislav Petkov <[email protected]>Cc Borislav Petkov <[email protected]>Cc Dave Hansen <[email protected]>Cc David Vrabel <[email protected]>Cc Dexuan Cui <[email protected]>Cc Greg Kroah-Hartman <[email protected]>Cc H. Peter Anvin <[email protected]>Cc Jan Beulich <[email protected]>Cc Joonsoo Kim <[email protected]>Cc Juergen Gross <[email protected]>Cc Linus Torvalds <[email protected]>Cc Pavel Machek <[email protected]>Cc Thomas Gleixner <[email protected]>Cc Tony Lindgren <[email protected]>Cc Toshi Kani <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Xishi Qiu <[email protected]>Signed-off-by Ingo Molnar <[email protected]>--- arch/x86/Kconfig | 23 +++++++---------------- arch/x86/mm/init.c | 7 +------ 2 files changed, 8 insertions(+), 22 deletions(-)arch/x86/Kconfig +0
-8@@ -1299,22 +1299,13 @@ config ARCH_DMA_ADDR_T_64BIT def_bool y depends on X86_64 || HIGHMEM64G -config ENABLE_DIRECT_GBPAGES - def_bool y - depends on X86_64 && !DEBUG_PAGEALLOC && !KMEMCHECK - -config DIRECT_GBPAGES - bool "Enable 1GB pages for kernel pagetables" if EXPERT - default y - depends on ENABLE_DIRECT_GBPAGES
arch/x86/Kconfig config ARCH_DMA_ADDR_T_64BIT · line 1303
Quoted 10 lines
arch/x86/Kconfig +2
-8- ---help--- - Enable by default the kernel linear mapping to use 1GB pages on CPUs - that support it. This can improve the kernel's performance a tiny bit - by reducing TLB pressure. If in doubt, say "Y". If you've disabled - option but your platform is capable of handling support for this - you can use the gbpages kernel parameter. Likewise if you've enabled - this but you'd like to force disable this option you can use the - nogbpages kernel parameter. +config X86_DIRECT_GBPAGES + def_bool (X86_64 && !DEBUG_PAGEALLOC && !KMEMCHECK)
config X86_DIRECT_GBPAGES
def_bool y depends on X86_64 && !DEBUG_PAGEALLOC && !KMEMCHECK
or else it'll leave a pointless
# CONFIG_X86_DIRECT_GBPAGES is not set
in .config-s where the dependencies are not met.
Jan
Re: [PATCH 5/4] x86/mm: Further simplify 1 GB kernel linear mappings handling
Quoted 4 lines · Jan Beulich
arch/x86/Kconfig config ARCH_DMA_ADDR_T_64BIT · line 1303
Quoted 2 lines
arch/x86/Kconfig +2
-0+config X86_DIRECT_GBPAGES + def_bool (X86_64 && !DEBUG_PAGEALLOC && !KMEMCHECK)config
X86_DIRECT_GBPAGES
Quoted 8 lines
def_bool y depends on X86_64 && !DEBUG_PAGEALLOC && !KMEMCHECKor else it'll leave a pointless
#
CONFIG_X86_DIRECT_GBPAGESis not setin .config-s where the dependencies are not met.
Ok, fixed - thanks!
Ingo
[PATCH 6/4] x86/mm: Simplify probe_page_size_mask()
In reply to Luis R. RodriguezNow that we've simplified the gbpages config space, move the 'page_size_mask' initialization into probe_page_size_mask(), right next to the PSE and PGE enablement lines.
Cc Luis R. Rodriguez <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Dave Hansen <[email protected]>
Cc David Vrabel <[email protected]>
Cc Dexuan Cui <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc H. Peter Anvin <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Xishi Qiu <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
--- arch/x86/mm/init.c | 23 ++++++++++------------- 1 file changed, 10 insertions(+), 13 deletions(-)
arch/x86/mm/init.c +10 -13
@@ -131,29 +131,18 @@ void __init early_alloc_pgt_buf(void)int after_bootmem; -static int page_size_mask; -early_param_on_off ("gbpages" ,"nogbpages" , direct_gbpages,CONFIG_X86_DIRECT_GBPAGES ); -static void __init init_gbpages (void ) -{ -if (direct_gbpages && cpu_has_gbpages) { -printk (KERN_INFO "Using GB pages for direct mapping\n" ); - page_size_mask |=1 <<PG_LEVEL_1G ; - }else - direct_gbpages =0 ; -} -struct map_range {unsigned long start;unsigned long end;unsigned page_size_mask; }; +static int page_size_mask; +static void __init probe_page_size_mask (void ) { -init_gbpages (); -#if !defined(CONFIG_DEBUG_PAGEALLOC ) && !defined(CONFIG_KMEMCHECK )/* * For CONFIG_DEBUG_PAGEALLOC, identity mapping will use small pages.
@@ -173,6 +162,14 @@ static void __init probe_page_size_mask(void)cr4_set_bits_and_update_boot (X86_CR4_PGE ); __supported_pte_mask |= _PAGE_GLOBAL; } + +/* Enable 1 GB linear kernel mappings if available: */ +if (direct_gbpages && cpu_has_gbpages) { +printk (KERN_INFO "Using GB pages for direct mapping\n" ); + page_size_mask |=1 <<PG_LEVEL_1G ; + }else { + direct_gbpages =0 ; + } }#ifdef CONFIG_X86_32
Re: [PATCH 6/4] x86/mm: Simplify probe_page_size_mask()
Quoted 69 lines · Ingo Molnar
Now that we've simplified the gbpages config space, move the '
page_size_mask' initialization intoprobe_page_size_mask(), right next to the PSE and PGE enablement lines.Cc Luis R. Rodriguez <[email protected]>Cc Andrew Morton <[email protected]>Cc Andy Lutomirski <[email protected]>Cc Borislav Petkov <[email protected]>Cc Borislav Petkov <[email protected]>Cc Dave Hansen <[email protected]>Cc David Vrabel <[email protected]>Cc Dexuan Cui <[email protected]>Cc Greg Kroah-Hartman <[email protected]>Cc H. Peter Anvin <[email protected]>Cc Jan Beulich <[email protected]>Cc Joonsoo Kim <[email protected]>Cc Juergen Gross <[email protected]>Cc Linus Torvalds <[email protected]>Cc Pavel Machek <[email protected]>Cc Thomas Gleixner <[email protected]>Cc Tony Lindgren <[email protected]>Cc Toshi Kani <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Xishi Qiu <[email protected]>Signed-off-by Ingo Molnar <[email protected]>--- arch/x86/mm/init.c | 23 ++++++++++------------- 1 file changed, 10 insertions(+), 13 deletions(-)arch/x86/mm/init.c +2
-13@@ -131,29 +131,18 @@ void __init early_alloc_pgt_buf(void)int after_bootmem; -static int page_size_mask; -early_param_on_off ("gbpages" ,"nogbpages" , direct_gbpages,CONFIG_X86_DIRECT_GBPAGES ); -static void __init init_gbpages (void ) -{ -if (direct_gbpages && cpu_has_gbpages) { -printk (KERN_INFO "Using GB pages for direct mapping\n" ); - page_size_mask |=1 <<PG_LEVEL_1G ; - }else - direct_gbpages =0 ; -} -struct map_range {unsigned long start;unsigned long end;unsigned page_size_mask; }; +static int page_size_mask; +static void __init probe_page_size_mask (void ) { -init_gbpages (); -#if !defined(CONFIG_DEBUG_PAGEALLOC ) && !defined(CONFIG_KMEMCHECK )/* * For CONFIG_DEBUG_PAGEALLOC, identity mapping will use small pages.
arch/x86/mm/init.c static void __init probe_page_size_mask(void) · line 168
Quoted 7 lines
arch/x86/mm/init.c +4
-0@@ -173,6 +162,14 @@ static void __init probe_page_size_mask(void)cr4_set_bits_and_update_boot (X86_CR4_PGE ); __supported_pte_mask |= _PAGE_GLOBAL; } + +/* Enable 1 GB linear kernel mappings if available: */ +if (direct_gbpages && cpu_has_gbpages) { +printk (KERN_INFO "Using GB pages for direct mapping\n" );
pr_info()?
Quoted 13 lines
arch/x86/mm/init.c +4
-1+ page_size_mask |=1 <<PG_LEVEL_1G ; + }else { + direct_gbpages =0 ; + } }#ifdef CONFIG_X86_32 --To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Re: [PATCH 6/4] x86/mm: Simplify probe_page_size_mask()
Quoted 71 lines · Juergen Gross
Quoted 70 lines · Ingo Molnar
Now that we've simplified the gbpages config space, move the '
page_size_mask' initialization intoprobe_page_size_mask(), right next to the PSE and PGE enablement lines.Cc Luis R. Rodriguez <[email protected]>Cc Andrew Morton <[email protected]>Cc Andy Lutomirski <[email protected]>Cc Borislav Petkov <[email protected]>Cc Borislav Petkov <[email protected]>Cc Dave Hansen <[email protected]>Cc David Vrabel <[email protected]>Cc Dexuan Cui <[email protected]>Cc Greg Kroah-Hartman <[email protected]>Cc H. Peter Anvin <[email protected]>Cc Jan Beulich <[email protected]>Cc Joonsoo Kim <[email protected]>Cc Juergen Gross <[email protected]>Cc Linus Torvalds <[email protected]>Cc Pavel Machek <[email protected]>Cc Thomas Gleixner <[email protected]>Cc Tony Lindgren <[email protected]>Cc Toshi Kani <[email protected]>Cc Vlastimil Babka <[email protected]>Cc Xishi Qiu <[email protected]>Signed-off-by Ingo Molnar <[email protected]>--- arch/x86/mm/init.c | 23 ++++++++++------------- 1 file changed, 10 insertions(+), 13 deletions(-)arch/x86/mm/init.c +2
-13@@ -131,29 +131,18 @@ void __init early_alloc_pgt_buf(void)int after_bootmem; -static int page_size_mask; -early_param_on_off ("gbpages" ,"nogbpages" , direct_gbpages,CONFIG_X86_DIRECT_GBPAGES ); -static void __init init_gbpages (void ) -{ -if (direct_gbpages && cpu_has_gbpages) { -printk (KERN_INFO "Using GB pages for direct mapping\n" ); - page_size_mask |=1 <<PG_LEVEL_1G ; - }else - direct_gbpages =0 ; -} -struct map_range {unsigned long start;unsigned long end;unsigned page_size_mask; }; +static int page_size_mask; +static void __init probe_page_size_mask (void ) { -init_gbpages (); -#if !defined(CONFIG_DEBUG_PAGEALLOC ) && !defined(CONFIG_KMEMCHECK )/* * For CONFIG_DEBUG_PAGEALLOC, identity mapping will use small pages. @@ -173,6 +162,14 @@ static void __init probe_page_size_mask(void)cr4_set_bits_and_update_boot (X86_CR4_PGE );
Quoted 8 lines
arch/x86/mm/init.c static void __init probe_page_size_mask(void) · line 168
Quoted 6 lines
arch/x86/mm/init.c +4
-0__supported_pte_mask |= _PAGE_GLOBAL; } + +/* Enable 1 GB linear kernel mappings if available: */ +if (direct_gbpages && cpu_has_gbpages) { +printk (KERN_INFO "Using GB pages for direct mapping\n" );
pr_info()?
Yeah - probably in a separate patch, as there's three other KERN_ users in arch/x86/mm/init.c.
Thanks,
Ingo
[PATCH 7/4] init.h: Clean up the __setup()/early_param() macros
In reply to Luis R. RodriguezMake it all a bit easier on the eyes:
- Move the __setup_param() lines right after their init functions
- Use consistent vertical spacing
- Use more horizontal spacing to make it look like regular C code
- Use standard comment style
Cc Luis R. Rodriguez <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Borislav Petkov <[email protected]>
Cc Dave Hansen <[email protected]>
Cc David Vrabel <[email protected]>
Cc Dexuan Cui <[email protected]>
Cc Greg Kroah-Hartman <[email protected]>
Cc H. Peter Anvin <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Joonsoo Kim <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Pavel Machek <[email protected]>
Cc Thomas Gleixner <[email protected]>
Cc Tony Lindgren <[email protected]>
Cc Toshi Kani <[email protected]>
Cc Vlastimil Babka <[email protected]>
Cc Xishi Qiu <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
--- include/linux/init.h | 49 +++++++++++++++++++++++++++---------------------- 1 file changed, 27 insertions(+), 22 deletions(-)
include/linux/init.h +27 -22
@@ -253,34 +253,39 @@ struct obs_kernel_param {* obs_kernel_param "array" too far apart in .init.setup. */ #define __setup_param (str, unique_id, fn, early) \ - static const char __setup_str_##unique_id[] __initconst \ - __aligned(1) = str; \ - static struct obs_kernel_param __setup_##unique_id \ - __used __section(.init.setup) \ - __attribute__((aligned((sizeof(long))))) \ + static const char __setup_str_##unique_id[] __initconst \ + __aligned(1) = str; \ + static struct obs_kernel_param __setup_##unique_id \ + __used __section(.init.setup) \ + __attribute__((aligned((sizeof(long))))) \ = { __setup_str_##unique_id, fn, early } -#define __setup (str, fn) \ +#define __setup (str, fn) \ __setup_param(str, fn, fn, 0) -/* NOTE: fn is as per module_param, not __setup! Emits warning if fn -* returns non-zero. */ -#define early_param (str, fn) \ +/* +* NOTE: fn is as per module_param, not __setup! +* Emits warning if fn returns non-zero. +*/ +#define early_param (str, fn) \ __setup_param(str, fn, fn, 1) -#define early_param_on_off (str_on, str_off, var, config) \ - int var = IS_ENABLED(config); \ - static int __init parse_##var##_on(char *arg) \ - { \ - var = 1; \ - return 0; \ - } \ - static int __init parse_##var##_off(char *arg) \ - { \ - var = 0; \ - return 0; \ - } \ - __setup_param(str_on, parse_##var##_on, parse_##var##_on, 1); \ +#define early_param_on_off (str_on, str_off, var, config) \ + \ + int var = IS_ENABLED(config); \ + \ + static int __init parse_##var##_on(char *arg) \ + { \ + var = 1; \ + return 0; \ + } \ + __setup_param(str_on, parse_##var##_on, parse_##var##_on, 1); \ + \ + static int __init parse_##var##_off(char *arg) \ + { \ + var = 0; \ + return 0; \ + } \ __setup_param(str_off, parse_##var##_off, parse_##var##_off, 1)/* Relies on boot_command_line being set */
PATCH 1/1 x86/mm/pat: Initialize __cachemode2pte_tbl[] and __pte2cachemode_tbl[] in a bit more readable fashion
In reply to Luis R. Rodriguez( So this patch is not directly related to gbpages, but while reading
arch/x86/mm/init.c I could not resist ... )
=======================>
The initialization of these two arrays is a bit difficult to follow:
restructure it optically so that a 2D structure shows which bit in the PTE is set and which not.
Also improve on comments a bit.
No code or data changed:
# arch/x86/mm/init.o:
text data bss dec hex filename
4585 424 29776 34785 87e1 init.o.before 4585 424 29776 34785 87e1 init.o.after
md5:
a82e11ff58bcfd0af3a94662a701f65d init.o.before.asm a82e11ff58bcfd0af3a94662a701f65d init.o.after.asm
Cc Andy Lutomirski <[email protected]>
Cc Dave Hansen <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Juergen Gross <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Luis R. Rodriguez <[email protected]>
Cc Toshi Kani <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
--- arch/x86/mm/init.c | 40 ++++++++++++++++++++++------------------ 1 file changed, 22 insertions(+), 18 deletions(-)
arch/x86/mm/init.c +22 -18
@@ -29,29 +29,33 @@/* * Tables translating between page_cache_type_t and pte encoding. -* Minimal supported modes are defined statically, modified if more supported -* cache modes are available. -* Index into __cachemode2pte_tbl is the cachemode. -* Index into __pte2cachemode_tbl are the caching attribute bits of the pte -* (_PAGE_PWT, _PAGE_PCD, _PAGE_PAT) at index bit positions 0, 1, 2. +* +* Minimal supported modes are defined statically, they are modified +* during bootup if more supported cache modes are available. +* +* Index into __cachemode2pte_tbl[] is the cachemode. +* +* Index into __pte2cachemode_tbl[] are the caching attribute bits of the pte +* (_PAGE_PWT, _PAGE_PCD, _PAGE_PAT) at index bit positions 0, 1, 2. */ uint16_t __cachemode2pte_tbl[_PAGE_CACHE_MODE_NUM] = { - [_PAGE_CACHE_MODE_WB] =0 , - [_PAGE_CACHE_MODE_WC] = _PAGE_PWT, - [_PAGE_CACHE_MODE_UC_MINUS] = _PAGE_PCD, - [_PAGE_CACHE_MODE_UC] = _PAGE_PCD | _PAGE_PWT, - [_PAGE_CACHE_MODE_WT] = _PAGE_PCD, - [_PAGE_CACHE_MODE_WP] = _PAGE_PCD, + [_PAGE_CACHE_MODE_WB ] =0 |0 , + [_PAGE_CACHE_MODE_WC ] = _PAGE_PWT |0 , + [_PAGE_CACHE_MODE_UC_MINUS] =0 | _PAGE_PCD, + [_PAGE_CACHE_MODE_UC ] = _PAGE_PWT | _PAGE_PCD, + [_PAGE_CACHE_MODE_WT ] =0 | _PAGE_PCD, + [_PAGE_CACHE_MODE_WP ] =0 | _PAGE_PCD, };EXPORT_SYMBOL (__cachemode2pte_tbl); +uint8_t __pte2cachemode_tbl[8 ] = { - [__pte2cm_idx (0 )] = _PAGE_CACHE_MODE_WB, - [__pte2cm_idx (_PAGE_PWT)] = _PAGE_CACHE_MODE_WC, - [__pte2cm_idx (_PAGE_PCD)] = _PAGE_CACHE_MODE_UC_MINUS, - [__pte2cm_idx (_PAGE_PWT | _PAGE_PCD)] = _PAGE_CACHE_MODE_UC, - [__pte2cm_idx (_PAGE_PAT)] = _PAGE_CACHE_MODE_WB, - [__pte2cm_idx (_PAGE_PWT | _PAGE_PAT)] = _PAGE_CACHE_MODE_WC, - [__pte2cm_idx (_PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC_MINUS, + [__pte2cm_idx (0 |0 |0 )] = _PAGE_CACHE_MODE_WB, + [__pte2cm_idx (_PAGE_PWT |0 |0 )] = _PAGE_CACHE_MODE_WC, + [__pte2cm_idx (0 | _PAGE_PCD |0 )] = _PAGE_CACHE_MODE_UC_MINUS, + [__pte2cm_idx (_PAGE_PWT | _PAGE_PCD |0 )] = _PAGE_CACHE_MODE_UC, + [__pte2cm_idx (0 |0 | _PAGE_PAT)] = _PAGE_CACHE_MODE_WB, + [__pte2cm_idx (_PAGE_PWT |0 | _PAGE_PAT)] = _PAGE_CACHE_MODE_WC, + [__pte2cm_idx (0 | _PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC_MINUS, [__pte2cm_idx (_PAGE_PWT | _PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC, };EXPORT_SYMBOL (__pte2cachemode_tbl);
Re: [PATCH] x86/mm/pat: Initialize __cachemode2pte_tbl[] and __pte2cachemode_tbl[] in a bit more readable fashion
Quoted 24 lines · Ingo Molnar
( So this patch is not directly related to gbpages, but while reading
arch/x86/mm/init.cI could not resist ... )=======================>
The initialization of these two arrays is a bit difficult to follow:
restructure it optically so that a 2D structure shows which bit in the PTE is set and which not.Also improve on comments a bit.
No code or data changed:
# arch/x86/mm/init.o: text data bss dec hex filename 4585 424 29776 34785 87e1 init.o.before 4585 424 29776 34785 87e1 init.o.aftermd5:
a82e11ff58bcfd0af3a94662a701f65d init.o.before.asm a82e11ff58bcfd0af3a94662a701f65d init.o.after.asm
Quoted 9 lines
Cc Andy Lutomirski <[email protected]>Cc Dave Hansen <[email protected]>Cc Jan Beulich <[email protected]>Cc Juergen Gross <[email protected]>Cc Linus Torvalds <[email protected]>Cc Luis R. Rodriguez <[email protected]>Cc Toshi Kani <[email protected]>Signed-off-by Ingo Molnar <[email protected]>
Reviewed-by Juergen Gross <[email protected]>
Quoted 66 lines
--- arch/x86/mm/init.c | 40 ++++++++++++++++++++++------------------ 1 file changed, 22 insertions(+), 18 deletions(-)arch/x86/mm/init.c +22
-19@@ -29,29 +29,33 @@/* * Tables translating between page_cache_type_t and pte encoding. -* Minimal supported modes are defined statically, modified if more supported -* cache modes are available. -* Index into __cachemode2pte_tbl is the cachemode. -* Index into __pte2cachemode_tbl are the caching attribute bits of the pte -* (_PAGE_PWT, _PAGE_PCD, _PAGE_PAT) at index bit positions 0, 1, 2. +* +* Minimal supported modes are defined statically, they are modified +* during bootup if more supported cache modes are available. +* +* Index into __cachemode2pte_tbl[] is the cachemode. +* +* Index into __pte2cachemode_tbl[] are the caching attribute bits of the pte +* (_PAGE_PWT, _PAGE_PCD, _PAGE_PAT) at index bit positions 0, 1, 2. */ uint16_t __cachemode2pte_tbl[_PAGE_CACHE_MODE_NUM] = { - [_PAGE_CACHE_MODE_WB] =0 , - [_PAGE_CACHE_MODE_WC] = _PAGE_PWT, - [_PAGE_CACHE_MODE_UC_MINUS] = _PAGE_PCD, - [_PAGE_CACHE_MODE_UC] = _PAGE_PCD | _PAGE_PWT, - [_PAGE_CACHE_MODE_WT] = _PAGE_PCD, - [_PAGE_CACHE_MODE_WP] = _PAGE_PCD, + [_PAGE_CACHE_MODE_WB ] =0 |0 , + [_PAGE_CACHE_MODE_WC ] = _PAGE_PWT |0 , + [_PAGE_CACHE_MODE_UC_MINUS] =0 | _PAGE_PCD, + [_PAGE_CACHE_MODE_UC ] = _PAGE_PWT | _PAGE_PCD, + [_PAGE_CACHE_MODE_WT ] =0 | _PAGE_PCD, + [_PAGE_CACHE_MODE_WP ] =0 | _PAGE_PCD, };EXPORT_SYMBOL (__cachemode2pte_tbl); +uint8_t __pte2cachemode_tbl[8 ] = { - [__pte2cm_idx (0 )] = _PAGE_CACHE_MODE_WB, - [__pte2cm_idx (_PAGE_PWT)] = _PAGE_CACHE_MODE_WC, - [__pte2cm_idx (_PAGE_PCD)] = _PAGE_CACHE_MODE_UC_MINUS, - [__pte2cm_idx (_PAGE_PWT | _PAGE_PCD)] = _PAGE_CACHE_MODE_UC, - [__pte2cm_idx (_PAGE_PAT)] = _PAGE_CACHE_MODE_WB, - [__pte2cm_idx (_PAGE_PWT | _PAGE_PAT)] = _PAGE_CACHE_MODE_WC, - [__pte2cm_idx (_PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC_MINUS, + [__pte2cm_idx (0 |0 |0 )] = _PAGE_CACHE_MODE_WB, + [__pte2cm_idx (_PAGE_PWT |0 |0 )] = _PAGE_CACHE_MODE_WC, + [__pte2cm_idx (0 | _PAGE_PCD |0 )] = _PAGE_CACHE_MODE_UC_MINUS, + [__pte2cm_idx (_PAGE_PWT | _PAGE_PCD |0 )] = _PAGE_CACHE_MODE_UC, + [__pte2cm_idx (0 |0 | _PAGE_PAT)] = _PAGE_CACHE_MODE_WB, + [__pte2cm_idx (_PAGE_PWT |0 | _PAGE_PAT)] = _PAGE_CACHE_MODE_WC, + [__pte2cm_idx (0 | _PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC_MINUS, [__pte2cm_idx (_PAGE_PWT | _PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC, };EXPORT_SYMBOL (__pte2cachemode_tbl); --To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
tip-bot for Ingo Molnar
merge bot[tip:x86/mm] x86/mm/pat: Initialize __cachemode2pte_tbl[] and __pte2cachemode_tbl[] in a bit more readable fashion
In reply to Ingo Molnarx86/mm/pat: Initialize __cachemode2pte_tbl[] and __pte2cachemode_tbl[] in a bit more readable fashion
Commit-ID: c709feda56886c38af3116254f84cbe6a78b3a5d Gitweb: http://git.kernel.org/tip/c709feda56886c38af3116254f84cbe6a78b3a5d Author: Ingo Molnar <[email protected]> AuthorDate: Thu, 5 Mar 2015 08:58:44 +0100 Committer: Ingo Molnar <[email protected]> CommitDate: Thu, 5 Mar 2015 09:48:17 +0100
x86/mm/pat: Initialize __cachemode2pte_tbl[] and __pte2cachemode_tbl[] in a bit more readable fashion
The initialization of these two arrays is a bit difficult to follow:
restructure it optically so that a 2D structure shows which bit in the PTE is set and which not.
Also improve on comments a bit.
No code or data changed:
# arch/x86/mm/init.o:
text data bss dec hex filename
4585 424 29776 34785 87e1 init.o.before 4585 424 29776 34785 87e1 init.o.after
md5:
a82e11ff58bcfd0af3a94662a701f65d init.o.before.asm a82e11ff58bcfd0af3a94662a701f65d init.o.after.asm
Reviewed-by Juergen Gross <[email protected]>
Cc Andy Lutomirski <[email protected]>
Cc Dave Hansen <[email protected]>
Cc Jan Beulich <[email protected]>
Cc Linus Torvalds <[email protected]>
Cc Luis R. Rodriguez <[email protected]>
Cc Toshi Kani <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>
--- arch/x86/mm/init.c | 40 ++++++++++++++++++++++------------------ 1 file changed, 22 insertions(+), 18 deletions(-)
arch/x86/mm/init.c +22 -18
@@ -29,29 +29,33 @@/* * Tables translating between page_cache_type_t and pte encoding. -* Minimal supported modes are defined statically, modified if more supported -* cache modes are available. -* Index into __cachemode2pte_tbl is the cachemode. -* Index into __pte2cachemode_tbl are the caching attribute bits of the pte -* (_PAGE_PWT, _PAGE_PCD, _PAGE_PAT) at index bit positions 0, 1, 2. +* +* Minimal supported modes are defined statically, they are modified +* during bootup if more supported cache modes are available. +* +* Index into __cachemode2pte_tbl[] is the cachemode. +* +* Index into __pte2cachemode_tbl[] are the caching attribute bits of the pte +* (_PAGE_PWT, _PAGE_PCD, _PAGE_PAT) at index bit positions 0, 1, 2. */ uint16_t __cachemode2pte_tbl[_PAGE_CACHE_MODE_NUM] = { - [_PAGE_CACHE_MODE_WB] =0 , - [_PAGE_CACHE_MODE_WC] = _PAGE_PWT, - [_PAGE_CACHE_MODE_UC_MINUS] = _PAGE_PCD, - [_PAGE_CACHE_MODE_UC] = _PAGE_PCD | _PAGE_PWT, - [_PAGE_CACHE_MODE_WT] = _PAGE_PCD, - [_PAGE_CACHE_MODE_WP] = _PAGE_PCD, + [_PAGE_CACHE_MODE_WB ] =0 |0 , + [_PAGE_CACHE_MODE_WC ] = _PAGE_PWT |0 , + [_PAGE_CACHE_MODE_UC_MINUS] =0 | _PAGE_PCD, + [_PAGE_CACHE_MODE_UC ] = _PAGE_PWT | _PAGE_PCD, + [_PAGE_CACHE_MODE_WT ] =0 | _PAGE_PCD, + [_PAGE_CACHE_MODE_WP ] =0 | _PAGE_PCD, };EXPORT_SYMBOL (__cachemode2pte_tbl); +uint8_t __pte2cachemode_tbl[8 ] = { - [__pte2cm_idx (0 )] = _PAGE_CACHE_MODE_WB, - [__pte2cm_idx (_PAGE_PWT)] = _PAGE_CACHE_MODE_WC, - [__pte2cm_idx (_PAGE_PCD)] = _PAGE_CACHE_MODE_UC_MINUS, - [__pte2cm_idx (_PAGE_PWT | _PAGE_PCD)] = _PAGE_CACHE_MODE_UC, - [__pte2cm_idx (_PAGE_PAT)] = _PAGE_CACHE_MODE_WB, - [__pte2cm_idx (_PAGE_PWT | _PAGE_PAT)] = _PAGE_CACHE_MODE_WC, - [__pte2cm_idx (_PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC_MINUS, + [__pte2cm_idx (0 |0 |0 )] = _PAGE_CACHE_MODE_WB, + [__pte2cm_idx (_PAGE_PWT |0 |0 )] = _PAGE_CACHE_MODE_WC, + [__pte2cm_idx (0 | _PAGE_PCD |0 )] = _PAGE_CACHE_MODE_UC_MINUS, + [__pte2cm_idx (_PAGE_PWT | _PAGE_PCD |0 )] = _PAGE_CACHE_MODE_UC, + [__pte2cm_idx (0 |0 | _PAGE_PAT)] = _PAGE_CACHE_MODE_WB, + [__pte2cm_idx (_PAGE_PWT |0 | _PAGE_PAT)] = _PAGE_CACHE_MODE_WC, + [__pte2cm_idx (0 | _PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC_MINUS, [__pte2cm_idx (_PAGE_PWT | _PAGE_PCD | _PAGE_PAT)] = _PAGE_CACHE_MODE_UC, };EXPORT_SYMBOL (__pte2cachemode_tbl);