x86: cacheinfo: cache_disable fixes/cleanup
cover x86: cacheinfo: cache_disable fixes/cleanup
These patches primarily split up Mark's latest patch
[PATCH][retry 6] Conform L3 Cache Index Disable to Linux Standards (see http://marc.info/?l=linux-kernel&m=123922420925908&w=2)
into smaller pieces. I prefer to revert the last two patches on tip/x86/cpu to get a cleaner changelog for the following fixes.
Patches are against tip/x86/cpu as of today.
Please apply.
Thanks,
Andreas
Signature · 5 lines
Operating | Advanced Micro Devices GmbH System | Karl-Hammerschmidt-Str. 34, 85609 Dornach b. München, Germany Research | Geschäftsführer: Jochen Polster, Thomas M. McCoy, Giuliano Meroni Center | Sitz: Dornach, Gemeinde Aschheim, Landkreis München (OSRC) | Registergericht München, HRB Nr. 43632
PATCH 1/8 Revert "x86, cpu: intel_cacheinfo.c: use cpumask_first(to_cpumask())"
revert "x86, cpu: intel_cacheinfo.c: use cpumask_first(to_cpumask())" and "x86, cpu: conform L3 Cache Index Disable to Linux standards"
in order to incrementally fix the L3 Cache index disable feature
This reverts commit d42c33ef0a236c6874c748438500c78edfaa1187.
--- arch/x86/kernel/cpu/intel_cacheinfo.c | 6 ++---- 1 files changed, 2 insertions(+), 4 deletions(-)
arch/x86/kernel/cpu/intel_cacheinfo.c +2 -4
@@ -643,8 +643,7 @@ static ssize_t show_##file_name \static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf,unsigned int index) { -int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); -int node =cpu_to_node (cpu); +int node =cpu_to_node (first_cpu (this_leaf->shared_cpu_map));struct pci_dev *dev = k8_northbridges[node];unsigned int reg =0 ;
@@ -666,8 +665,7 @@ static ssize_t store_cache_disable(struct _cpuid4_info *this_leaf, const char *buf, size_t count, unsigned int index) { - int cpu = cpumask_first(to_cpumask(this_leaf->shared_cpu_map)); -int node =cpu_to_node (cpu); + int node = cpu_to_node(first_cpu(this_leaf->shared_cpu_map));struct pci_dev *dev = k8_northbridges[node];unsigned long val =0 ;unsigned int scrubber =0 ;
Signature · 1 line
1.6.2
PATCH 2/8 Revert "x86, cpu: conform L3 Cache Index Disable to Linux standards"
In reply to Andreas Herrmannrevert "x86, cpu: intel_cacheinfo.c: use cpumask_first(to_cpumask())" and "x86, cpu: conform L3 Cache Index Disable to Linux standards"
in order to incrementally fix the L3 Cache index disable feature
This reverts commit 45ca863a40306ccc99c68d13421b6577240760ca.
--- arch/x86/include/asm/k8.h | 4 - arch/x86/kernel/cpu/intel_cacheinfo.c | 180 +++++++++++++++++++-------------- 2 files changed, 102 insertions(+), 82 deletions(-)
arch/x86/include/asm/k8.h +0 -4
@@ -6,11 +6,7 @@extern struct pci_device_id k8_nb_ids[];extern int early_is_k8_nb (u32 value); -#ifdef CONFIG_K8_NB extern struct pci_dev **k8_northbridges; -#else -struct pci_dev **k8_northbridges; -#endif extern int num_k8_northbridges;extern int cache_k8_northbridges (void );extern void k8_flush_garts (void );
arch/x86/kernel/cpu/intel_cacheinfo.c +102 -78
222 lines not on this page · raw message
Signature · 1 line
1.6.2
PATCH 3/8 x86: cacheinfo: use L3 cache index disable feature only for CPUs that support it
In reply to Andreas HerrmannAMD family 0x11 CPU doesn't support the feature.
Some AMD family 0x10 CPUs do not support it or have an erratum, see erratum #382 in "Revision Guide for AMD Family 10h Processors, 41322 Rev. 3.40 February 2009".
--- arch/x86/kernel/cpu/intel_cacheinfo.c | 8 ++++++++ 1 files changed, 8 insertions(+), 0 deletions(-)
arch/x86/kernel/cpu/intel_cacheinfo.c +8 -0
@@ -291,6 +291,14 @@ amd_check_l3_disable(int index, struct _cpuid4_info_regs *this_leaf) {if (index <3 )return ; + +if (boot_cpu_data.x86 ==0x11 ) +return ; + +/* see erratum #382 */ +if ((boot_cpu_data.x86 ==0x10 ) && (boot_cpu_data.x86_model <0x8 )) +return ; + this_leaf->can_disable =1 ; }
Signature · 1 line
1.6.2
[tip:x86/cpu] x86: cacheinfo: use L3 cache index disable feature only for CPUs that support it
Commit-ID: bda869c614c937c318547c3ee1d65a316b693c21 Gitweb: http://git.kernel.org/tip/bda869c614c937c318547c3ee1d65a316b693c21 Author: Andreas Herrmann <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:05:10 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:21:40 +0200
x86: cacheinfo: use L3 cache index disable feature only for CPUs that support it
AMD family 0x11 CPU doesn't support the feature.
Some AMD family 0x10 CPUs do not support it or have an erratum, see erratum #382 in "Revision Guide for AMD Family 10h Processors, 41322 Rev. 3.40 February 2009".
LKML-Reference: <[email protected]>
--- arch/x86/kernel/cpu/intel_cacheinfo.c | 8 ++++++++ 1 files changed, 8 insertions(+), 0 deletions(-)
arch/x86/kernel/cpu/intel_cacheinfo.c +8 -0
@@ -291,6 +291,14 @@ amd_check_l3_disable(int index, struct _cpuid4_info_regs *this_leaf) {if (index <3 )return ; + +if (boot_cpu_data.x86 ==0x11 ) +return ; + +/* see erratum #382 */ +if ((boot_cpu_data.x86 ==0x10 ) && (boot_cpu_data.x86_model <0x8 )) +return ; + this_leaf->can_disable =1 ; }
PATCH 4/8 x86: cacheinfo: correct return value when cache_disable feature is not active
In reply to Andreas HerrmannImpact: bug fix
If user writes to "cache_disable" attribute on a CPU that does not support this feature, the process hangs due to an invalid return value in store_cache_disable().
--- arch/x86/kernel/cpu/intel_cacheinfo.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
arch/x86/kernel/cpu/intel_cacheinfo.c +1 -1
@@ -771,7 +771,7 @@ store_cache_disable(struct _cpuid4_info *this_leaf, const char *buf,unsigned int ret, index, val;if (!this_leaf->can_disable) -return 0 ; +return -EINVAL ;if (strlen (buf) >15 )return -EINVAL ;
Signature · 1 line
1.6.2
[tip:x86/cpu] x86: cacheinfo: correct return value when cache_disable feature is not active
Commit-ID: 845d8c761ec763871936c62b837c4a9ea6d0fbdb Gitweb: http://git.kernel.org/tip/845d8c761ec763871936c62b837c4a9ea6d0fbdb Author: Andreas Herrmann <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:07:29 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:21:42 +0200
x86: cacheinfo: correct return value when cache_disable feature is not active
Impact: bug fix
If user writes to "cache_disable" attribute on a CPU that does not support this feature, the process hangs due to an invalid return value in store_cache_disable().
LKML-Reference: <[email protected]>
--- arch/x86/kernel/cpu/intel_cacheinfo.c | 2 +- 1 files changed, 1 insertions(+), 1 deletions(-)
arch/x86/kernel/cpu/intel_cacheinfo.c +1 -1
@@ -771,7 +771,7 @@ store_cache_disable(struct _cpuid4_info *this_leaf, const char *buf,unsigned int ret, index, val;if (!this_leaf->can_disable) -return 0 ; +return -EINVAL ;if (strlen (buf) >15 )return -EINVAL ;
PATCH 5/8 x86: cacheinfo: use cached K8 NB_MISC devices instead of scanning for it
In reply to Andreas HerrmannImpact: avoid code duplication
--- arch/x86/include/asm/k8.h | 8 +++++++ arch/x86/kernel/cpu/intel_cacheinfo.c | 37 ++------------------------------ 2 files changed, 11 insertions(+), 34 deletions(-)
Ingo,
you have asked for a function to replace the exported k8_northbridges array.
Actually this is a good idea and I'll look at it.
But this also requires additional changes in pci-gart_64.c And I don't like to mix such changes with this cache_disable stuff.
For the time beeing I provide a small macro to access the array.
Regards,
Andreas
arch/x86/include/asm/k8.h +8 -0
@@ -12,4 +12,12 @@ extern int cache_k8_northbridges(void);extern void k8_flush_garts (void );extern int k8_scan_nodes (unsigned long start,unsigned long end); +#ifdef CONFIG_K8_NB +#define node_to_k8_nb_misc (node) \ + (node < num_k8_northbridges) ? k8_northbridges[node] : NULL +#else +#define node_to_k8_nb_misc (node) NULL +#endif + +#endif /* _ASM_X86_K8_H */
arch/x86/kernel/cpu/intel_cacheinfo.c +3 -34
@@ -17,6 +17,7 @@#include <asm/processor.h> #include <asm/smp.h> +#include <asm/k8.h> #define LVL_1_INST 1#define LVL_1_DATA 2
@@ -159,14 +160,6 @@ struct _cpuid4_info_regs {unsigned long can_disable; }; -#ifdef CONFIG_PCI -static struct pci_device_id k8_nb_id[] = { - {PCI_DEVICE (PCI_VENDOR_ID_AMD ,0x1103 ) }, - {PCI_DEVICE (PCI_VENDOR_ID_AMD ,0x1203 ) }, - {} -}; -#endif -unsigned short num_cache_leaves;/* AMD doesn't have CPUID4. Emulate it here to report the same
@@ -704,30 +697,6 @@ static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf)#define to_object (k) container_of(k, struct _index_kobject, kobj)#define to_attr (a) container_of(a, struct _cache_attr, attr) -#ifdef CONFIG_PCI -static struct pci_dev *get_k8_northbridge (int node) -{ -struct pci_dev *dev =NULL ; -int i; - -for (i =0 ; i <= node; i++) { -do { - dev =pci_get_device (PCI_ANY_ID ,PCI_ANY_ID , dev); -if (!dev) -break ; - }while (!pci_match_id (&k8_nb_id[0 ], dev)); -if (!dev) -break ; - } -return dev; -} -#else -static struct pci_dev *get_k8_northbridge (int node) -{ -return NULL ; -} -#endif -static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf) {const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map);
@@ -739,7 +708,7 @@ static ssize_t show_cache_disable(struct _cpuid4_info *this_leaf, char *buf)if (!this_leaf->can_disable)return sprintf (buf,"Feature not enabled\n" ); - dev =get_k8_northbridge (node); + dev =node_to_k8_nb_misc (node);if (!dev) {printk (KERN_ERR "Attempting AMD northbridge operation on a system with no northbridge\n" );return -EINVAL ;
@@ -783,7 +752,7 @@ store_cache_disable(struct _cpuid4_info *this_leaf, const char *buf,return -EINVAL ; val |=0xc0000000 ; - dev =get_k8_northbridge (node); + dev =node_to_k8_nb_misc (node);if (!dev) {printk (KERN_ERR "Attempting AMD northbridge operation on a system with no northbridge\n" );return -EINVAL ;
Signature · 1 line
1.6.2
[tip:x86/cpu] x86: cacheinfo: use cached K8 NB_MISC devices instead of scanning for it
Commit-ID: afd9fceec55225d33be878927056a548c2eef26c Gitweb: http://git.kernel.org/tip/afd9fceec55225d33be878927056a548c2eef26c Author: Andreas Herrmann <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:16:17 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:21:49 +0200
x86: cacheinfo: use cached K8 NB_MISC devices instead of scanning for it
Impact: avoid code duplication
LKML-Reference: <[email protected]>
--- arch/x86/include/asm/k8.h | 8 +++++++ arch/x86/kernel/cpu/intel_cacheinfo.c | 37 ++------------------------------ 2 files changed, 11 insertions(+), 34 deletions(-)
arch/x86/include/asm/k8.h +8 -0
@@ -12,4 +12,12 @@ extern int cache_k8_northbridges(void);extern void k8_flush_garts (void );extern int k8_scan_nodes (unsigned long start,unsigned long end); +#ifdef CONFIG_K8_NB +#define node_to_k8_nb_misc (node) \ + (node < num_k8_northbridges) ? k8_northbridges[node] : NULL +#else +#define node_to_k8_nb_misc (node) NULL +#endif + +#endif /* _ASM_X86_K8_H */
arch/x86/kernel/cpu/intel_cacheinfo.c +3 -34
@@ -17,6 +17,7 @@#include <asm/processor.h> #include <asm/smp.h> +#include <asm/k8.h> #define LVL_1_INST 1#define LVL_1_DATA 2
@@ -159,14 +160,6 @@ struct _cpuid4_info_regs {unsigned long can_disable; }; -#if defined(CONFIG_PCI ) && defined(CONFIG_SYSFS ) -static struct pci_device_id k8_nb_id[] = { - {PCI_DEVICE (PCI_VENDOR_ID_AMD ,0x1103 ) }, - {PCI_DEVICE (PCI_VENDOR_ID_AMD ,0x1203 ) }, - {} -}; -#endif -unsigned short num_cache_leaves;/* AMD doesn't have CPUID4. Emulate it here to report the same
@@ -704,30 +697,6 @@ static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf)#define to_object (k) container_of(k, struct _index_kobject, kobj)#define to_attr (a) container_of(a, struct _cache_attr, attr) -#ifdef CONFIG_PCI -static struct pci_dev *get_k8_northbridge (int node) -{ -struct pci_dev *dev =NULL ; -int i; - -for (i =0 ; i <= node; i++) { -do { - dev =pci_get_device (PCI_ANY_ID ,PCI_ANY_ID , dev); -if (!dev) -break ; - }while (!pci_match_id (&k8_nb_id[0 ], dev)); -if (!dev) -break ; - } -return dev; -} -#else -static struct pci_dev *get_k8_northbridge (int node) -{ -return NULL ; -} -#endif -static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf) {const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map);
@@ -739,7 +708,7 @@ static ssize_t show_cache_disable(struct _cpuid4_info *this_leaf, char *buf)if (!this_leaf->can_disable)return sprintf (buf,"Feature not enabled\n" ); - dev =get_k8_northbridge (node); + dev =node_to_k8_nb_misc (node);if (!dev) {printk (KERN_ERR "Attempting AMD northbridge operation on a system with no northbridge\n" );return -EINVAL ;
@@ -783,7 +752,7 @@ store_cache_disable(struct _cpuid4_info *this_leaf, const char *buf,return -EINVAL ; val |=0xc0000000 ; - dev =get_k8_northbridge (node); + dev =node_to_k8_nb_misc (node);if (!dev) {printk (KERN_ERR "Attempting AMD northbridge operation on a system with no northbridge\n" );return -EINVAL ;
PATCH 6/8 x86: cacheinfo: replace sysfs interface for cache_disable feature
In reply to Andreas HerrmannFrom: Mark Langsdorf <[email protected]>
Impact: replace sysfs attribute
Current interface violates against "one-value-per-sysfs-attribute rule". This patch replaces current attribute with two attributes -- one for each L3 Cache Index Disable register.
--- arch/x86/kernel/cpu/intel_cacheinfo.c | 90 ++++++++++++++++---------------- 1 files changed, 45 insertions(+), 45 deletions(-)
arch/x86/kernel/cpu/intel_cacheinfo.c +45 -45
@@ -697,73 +697,69 @@ static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf)#define to_object (k) container_of(k, struct _index_kobject, kobj)#define to_attr (a) container_of(a, struct _cache_attr, attr) -static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf) +static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf, +unsigned int index) { -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -ssize_t ret =0 ; -int i; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned int reg =0 ;if (!this_leaf->can_disable) -return sprintf (buf,"Feature not enabled\n" ); - - dev =node_to_k8_nb_misc (node); -if (!dev) { -printk (KERN_ERR "Attempting AMD northbridge operation on a system with no northbridge\n" );return -EINVAL ; - } -for (i =0 ; i <2 ; i++) { -unsigned int reg; +if (!dev) +return -EINVAL ; -pci_read_config_dword (dev,0x1BC + i *4 , ®); +pci_read_config_dword (dev,0x1BC + index *4 , ®); +return sprintf (buf,"%x\n" , reg); +} - ret +=sprintf (buf,"%sEntry: %d\n" , buf, i); - ret +=sprintf (buf,"%sReads: %s\tNew Entries: %s\n" , - buf, - reg &0x80000000 ?"Disabled" :"Allowed" , - reg &0x40000000 ?"Disabled" :"Allowed" ); - ret +=sprintf (buf,"%sSubCache: %x\tIndex: %x\n" , - buf, (reg &0x30000 ) >>16 , reg &0xfff ); - } -return ret; +#define SHOW_CACHE_DISABLE (index) \ +static ssize_t \ +show_cache_disable_##index(struct _cpuid4_info *this_leaf, char *buf) \ +{ \ + return show_cache_disable(this_leaf, buf, index); \ } +SHOW_CACHE_DISABLE(0) +SHOW_CACHE_DISABLE(1) -static ssize_t -store_cache_disable (struct _cpuid4_info *this_leaf,const char *buf, -size_t count) +static ssize_t store_cache_disable (struct _cpuid4_info *this_leaf, +const char *buf,size_t count,unsigned int index) { -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -unsigned int ret, index, val; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned long val =0 ;if (!this_leaf->can_disable)return -EINVAL ; -if (strlen (buf) >15 ) -return -EINVAL ; +if (!capable (CAP_SYS_ADMIN )) +return -EPERM ; - ret =sscanf (buf,"%x %x" , &index, &val); -if (ret !=2 ) -return -EINVAL ; -if (index >1 ) +if (!dev)return -EINVAL ; - val |=0xc0000000 ; - dev =node_to_k8_nb_misc (node); -if (!dev) { -printk (KERN_ERR "Attempting AMD northbridge operation on a system with no northbridge\n" ); +if (strict_strtoul (buf,10 , &val) <0 )return -EINVAL ; - } + val |=0xc0000000 ;pci_write_config_dword (dev,0x1BC + index *4 , val & ~0x40000000 );wbinvd ();pci_write_config_dword (dev,0x1BC + index *4 , val); +return count; +} -return 1 ; +#define STORE_CACHE_DISABLE (index) \ +static ssize_t \ +store_cache_disable_##index(struct _cpuid4_info *this_leaf, \ + const char *buf, size_t count) \ +{ \ + return store_cache_disable(this_leaf, buf, count, index); \ } +STORE_CACHE_DISABLE(0) +STORE_CACHE_DISABLE(1)struct _cache_attr {struct attribute attr;
@@ -785,7 +781,10 @@ define_one_ro(size);define_one_ro (shared_cpu_map);define_one_ro (shared_cpu_list); -static struct _cache_attr cache_disable =__ATTR (cache_disable,0644 , show_cache_disable, store_cache_disable); +static struct _cache_attr cache_disable_0 =__ATTR (cache_disable_0,0644 , + show_cache_disable_0, store_cache_disable_0); +static struct _cache_attr cache_disable_1 =__ATTR (cache_disable_1,0644 , + show_cache_disable_1, store_cache_disable_1);static struct attribute * default_attrs[] = { &type.attr,
@@ -797,7 +796,8 @@ static struct attribute * default_attrs[] = { &size.attr, &shared_cpu_map.attr, &shared_cpu_list.attr, - &cache_disable.attr, + &cache_disable_0.attr, + &cache_disable_1.attr, NULL };
Signature · 1 line
1.6.2
[tip:x86/cpu] x86: cacheinfo: replace sysfs interface for cache_disable feature
Commit-ID: f8b201fc7110c3673437254e8ba02451461ece0b Gitweb: http://git.kernel.org/tip/f8b201fc7110c3673437254e8ba02451461ece0b Author: Mark Langsdorf <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:18:49 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:21:53 +0200
x86: cacheinfo: replace sysfs interface for cache_disable feature
Impact: replace sysfs attribute
Current interface violates against "one-value-per-sysfs-attribute rule". This patch replaces current attribute with two attributes -- one for each L3 Cache Index Disable register.
LKML-Reference: <[email protected]>
--- arch/x86/kernel/cpu/intel_cacheinfo.c | 90 ++++++++++++++++---------------- 1 files changed, 45 insertions(+), 45 deletions(-)
arch/x86/kernel/cpu/intel_cacheinfo.c +45 -45
@@ -697,73 +697,69 @@ static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf)#define to_object (k) container_of(k, struct _index_kobject, kobj)#define to_attr (a) container_of(a, struct _cache_attr, attr) -static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf) +static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf, +unsigned int index) { -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -ssize_t ret =0 ; -int i; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned int reg =0 ;if (!this_leaf->can_disable) -return sprintf (buf,"Feature not enabled\n" ); - - dev =node_to_k8_nb_misc (node); -if (!dev) { -printk (KERN_ERR "Attempting AMD northbridge operation on a system with no northbridge\n" );return -EINVAL ; - } -for (i =0 ; i <2 ; i++) { -unsigned int reg; +if (!dev) +return -EINVAL ; -pci_read_config_dword (dev,0x1BC + i *4 , ®); +pci_read_config_dword (dev,0x1BC + index *4 , ®); +return sprintf (buf,"%x\n" , reg); +} - ret +=sprintf (buf,"%sEntry: %d\n" , buf, i); - ret +=sprintf (buf,"%sReads: %s\tNew Entries: %s\n" , - buf, - reg &0x80000000 ?"Disabled" :"Allowed" , - reg &0x40000000 ?"Disabled" :"Allowed" ); - ret +=sprintf (buf,"%sSubCache: %x\tIndex: %x\n" , - buf, (reg &0x30000 ) >>16 , reg &0xfff ); - } -return ret; +#define SHOW_CACHE_DISABLE (index) \ +static ssize_t \ +show_cache_disable_##index(struct _cpuid4_info *this_leaf, char *buf) \ +{ \ + return show_cache_disable(this_leaf, buf, index); \ } +SHOW_CACHE_DISABLE(0) +SHOW_CACHE_DISABLE(1) -static ssize_t -store_cache_disable (struct _cpuid4_info *this_leaf,const char *buf, -size_t count) +static ssize_t store_cache_disable (struct _cpuid4_info *this_leaf, +const char *buf,size_t count,unsigned int index) { -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -unsigned int ret, index, val; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned long val =0 ;if (!this_leaf->can_disable)return -EINVAL ; -if (strlen (buf) >15 ) -return -EINVAL ; +if (!capable (CAP_SYS_ADMIN )) +return -EPERM ; - ret =sscanf (buf,"%x %x" , &index, &val); -if (ret !=2 ) -return -EINVAL ; -if (index >1 ) +if (!dev)return -EINVAL ; - val |=0xc0000000 ; - dev =node_to_k8_nb_misc (node); -if (!dev) { -printk (KERN_ERR "Attempting AMD northbridge operation on a system with no northbridge\n" ); +if (strict_strtoul (buf,10 , &val) <0 )return -EINVAL ; - } + val |=0xc0000000 ;pci_write_config_dword (dev,0x1BC + index *4 , val & ~0x40000000 );wbinvd ();pci_write_config_dword (dev,0x1BC + index *4 , val); +return count; +} -return 1 ; +#define STORE_CACHE_DISABLE (index) \ +static ssize_t \ +store_cache_disable_##index(struct _cpuid4_info *this_leaf, \ + const char *buf, size_t count) \ +{ \ + return store_cache_disable(this_leaf, buf, count, index); \ } +STORE_CACHE_DISABLE(0) +STORE_CACHE_DISABLE(1)struct _cache_attr {struct attribute attr;
@@ -785,7 +781,10 @@ define_one_ro(size);define_one_ro (shared_cpu_map);define_one_ro (shared_cpu_list); -static struct _cache_attr cache_disable =__ATTR (cache_disable,0644 , show_cache_disable, store_cache_disable); +static struct _cache_attr cache_disable_0 =__ATTR (cache_disable_0,0644 , + show_cache_disable_0, store_cache_disable_0); +static struct _cache_attr cache_disable_1 =__ATTR (cache_disable_1,0644 , + show_cache_disable_1, store_cache_disable_1);static struct attribute * default_attrs[] = { &type.attr,
@@ -797,7 +796,8 @@ static struct attribute * default_attrs[] = { &size.attr, &shared_cpu_map.attr, &shared_cpu_list.attr, - &cache_disable.attr, + &cache_disable_0.attr, + &cache_disable_1.attr, NULL };
Re: [tip:x86/cpu] x86: cacheinfo: replace sysfs interface for cache_disable feature
Hello Mark,
Quoted 37 lines · Mark Langsdorf
Commit-ID: f8b201fc7110c3673437254e8ba02451461ece0b Gitweb: http://git.kernel.org/tip/f8b201fc7110c3673437254e8ba02451461ece0b Author: Mark Langsdorf <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:18:49 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:21:53 +0200x86: cacheinfo: replace sysfs interface for
cache_disablefeatureImpact: replace sysfs attribute
Current interface violates against "one-value-per-sysfs-attribute rule". This patch replaces current attribute with two attributes -- one for each L3 Cache Index Disable register.
Signed-off-by Mark Langsdorf <[email protected]>Signed-off-by Andreas Herrmann <[email protected]>Cc Andrew Morton <[email protected]>LKML-Reference: <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>--- arch/x86/kernel/cpu/intel_cacheinfo.c | 90 ++++++++++++++++---------------- 1 files changed, 45 insertions(+), 45 deletions(-)arch/x86/kernel/cpu/intel_cacheinfo.c +2
-1@@ -697,73 +697,69 @@ static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf)#define to_object (k) container_of(k, struct _index_kobject, kobj)#define to_attr (a) container_of(a, struct _cache_attr, attr) -static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf) +static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf, +unsigned int index)
arch/x86/kernel/cpu/intel_cacheinfo.c static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) · line 706
Quoted 10 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +4
-5{ -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -ssize_t ret =0 ; -int i; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned int reg =0 ;
This leads to compilation warnings:
CC arch/x86/kernel/cpu/intel_cacheinfo.oarch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘show_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:712: warning: unused variable ‘node’
Quoted 6 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +2
-3-static ssize_t -store_cache_disable (struct _cpuid4_info *this_leaf,const char *buf, -size_t count) +static ssize_t store_cache_disable (struct _cpuid4_info *this_leaf, +const char *buf,size_t count,unsigned int index)
arch/x86/kernel/cpu/intel_cacheinfo.c static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) · line 734
Quoted 10 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +4
-4{ -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -unsigned int ret, index, val; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned long val =0 ;
CC arch/x86/kernel/cpu/intel_cacheinfo.oarch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘store_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:737: warning: unused variable ‘cpu’
Thanks,
--
JSR
Re: [tip:x86/cpu] x86: cacheinfo: replace sysfs interface for cache_disable feature
Quoted 68 lines · Jaswinder Singh Rajput
Hello Mark,
Quoted 37 lines · Mark Langsdorf
Commit-ID: f8b201fc7110c3673437254e8ba02451461ece0b Gitweb: http://git.kernel.org/tip/f8b201fc7110c3673437254e8ba02451461ece0b Author: Mark Langsdorf <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:18:49 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:21:53 +0200x86: cacheinfo: replace sysfs interface for
cache_disablefeatureImpact: replace sysfs attribute
Current interface violates against "one-value-per-sysfs-attribute rule". This patch replaces current attribute with two attributes -- one for each L3 Cache Index Disable register.
Signed-off-by Mark Langsdorf <[email protected]>Signed-off-by Andreas Herrmann <[email protected]>Cc Andrew Morton <[email protected]>LKML-Reference: <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>--- arch/x86/kernel/cpu/intel_cacheinfo.c | 90 ++++++++++++++++---------------- 1 files changed, 45 insertions(+), 45 deletions(-)arch/x86/kernel/cpu/intel_cacheinfo.c +2
-1@@ -697,73 +697,69 @@ static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf)#define to_object (k) container_of(k, struct _index_kobject, kobj)#define to_attr (a) container_of(a, struct _cache_attr, attr) -static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf) +static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf, +unsigned int index)arch/x86/kernel/cpu/intel_cacheinfo.c static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) · line 706
Quoted 10 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +4
-5{ -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -ssize_t ret =0 ; -int i; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned int reg =0 ;This leads to compilation warnings:
CCarch/x86/kernel/cpu/intel_cacheinfo.oarch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘show_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:712: warning: unused variable ‘node’Quoted 12 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +3
-7-static ssize_t -store_cache_disable (struct _cpuid4_info *this_leaf,const char *buf, -size_t count) +static ssize_t store_cache_disable (struct _cpuid4_info *this_leaf, +const char *buf,size_t count,unsigned int index) { -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -unsigned int ret, index, val; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map));
Quoted 8 lines
arch/x86/kernel/cpu/intel_cacheinfo.c static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) · line 734
Quoted 4 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +3
-0+int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned long val =0 ;CC
arch/x86/kernel/cpu/intel_cacheinfo.oarch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘store_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:737: warning: unused variable ‘cpu’
Oops it is also pointing to node:
arch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘store_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:739: warning: unused variable ‘node’
--
JSR
Re: [tip:x86/cpu] x86: cacheinfo: replace sysfs interface for cache_disable feature
Quoted 78 lines · Jaswinder Singh Rajput
Quoted 68 lines · Jaswinder Singh Rajput
Hello Mark,
Quoted 37 lines · Mark Langsdorf
Commit-ID: f8b201fc7110c3673437254e8ba02451461ece0b Gitweb: http://git.kernel.org/tip/f8b201fc7110c3673437254e8ba02451461ece0b Author: Mark Langsdorf <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:18:49 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:21:53 +0200x86: cacheinfo: replace sysfs interface for
cache_disablefeatureImpact: replace sysfs attribute
Current interface violates against "one-value-per-sysfs-attribute rule". This patch replaces current attribute with two attributes -- one for each L3 Cache Index Disable register.
Signed-off-by Mark Langsdorf <[email protected]>Signed-off-by Andreas Herrmann <[email protected]>Cc Andrew Morton <[email protected]>LKML-Reference: <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>--- arch/x86/kernel/cpu/intel_cacheinfo.c | 90 ++++++++++++++++---------------- 1 files changed, 45 insertions(+), 45 deletions(-)arch/x86/kernel/cpu/intel_cacheinfo.c +2
-1@@ -697,73 +697,69 @@ static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf)#define to_object (k) container_of(k, struct _index_kobject, kobj)#define to_attr (a) container_of(a, struct _cache_attr, attr) -static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf) +static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf, +unsigned int index)arch/x86/kernel/cpu/intel_cacheinfo.c static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) · line 706
Quoted 10 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +4
-5{ -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -ssize_t ret =0 ; -int i; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned int reg =0 ;This leads to compilation warnings:
CCarch/x86/kernel/cpu/intel_cacheinfo.oarch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘show_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:712: warning: unused variable ‘node’Quoted 12 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +3
-7-static ssize_t -store_cache_disable (struct _cpuid4_info *this_leaf,const char *buf, -size_t count) +static ssize_t store_cache_disable (struct _cpuid4_info *this_leaf, +const char *buf,size_t count,unsigned int index) { -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -unsigned int ret, index, val; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map));Quoted 8 lines
arch/x86/kernel/cpu/intel_cacheinfo.c static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) · line 734
Quoted 4 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +3
-0+int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned long val =0 ;CC
arch/x86/kernel/cpu/intel_cacheinfo.oarch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘store_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:737: warning: unused variable ‘cpu’
Quoted 4 lines
Oops it is also pointing to node:
arch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘store_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:739: warning: unused variable ‘node’
Mind sending a fix? We really dont want warnings noise in the x86 tree.
I was also thinking about adopting the Sparc trick: adding -Werror to arch/x86/ files only, fixing all the fallout and ensuring no new warnings from a flag day on. Anyone interested in doing that?
Ingo
Re: [tip:x86/cpu] x86: cacheinfo: replace sysfs interface for cache_disable feature
Quoted 80 lines · Ingo Molnar
Quoted 78 lines · Jaswinder Singh Rajput
Quoted 76 lines · Jaswinder Singh Rajput
Hello Mark,
Quoted 37 lines · Mark Langsdorf
Commit-ID: f8b201fc7110c3673437254e8ba02451461ece0b Gitweb: http://git.kernel.org/tip/f8b201fc7110c3673437254e8ba02451461ece0b Author: Mark Langsdorf <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:18:49 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:21:53 +0200x86: cacheinfo: replace sysfs interface for
cache_disablefeatureImpact: replace sysfs attribute
Current interface violates against "one-value-per-sysfs-attribute rule". This patch replaces current attribute with two attributes -- one for each L3 Cache Index Disable register.
Signed-off-by Mark Langsdorf <[email protected]>Signed-off-by Andreas Herrmann <[email protected]>Cc Andrew Morton <[email protected]>LKML-Reference: <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>--- arch/x86/kernel/cpu/intel_cacheinfo.c | 90 ++++++++++++++++---------------- 1 files changed, 45 insertions(+), 45 deletions(-)arch/x86/kernel/cpu/intel_cacheinfo.c +2
-1@@ -697,73 +697,69 @@ static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf)#define to_object (k) container_of(k, struct _index_kobject, kobj)#define to_attr (a) container_of(a, struct _cache_attr, attr) -static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf) +static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf, +unsigned int index)arch/x86/kernel/cpu/intel_cacheinfo.c static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) · line 706
Quoted 10 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +4
-5{ -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -ssize_t ret =0 ; -int i; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned int reg =0 ;This leads to compilation warnings:
CCarch/x86/kernel/cpu/intel_cacheinfo.oarch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘show_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:712: warning: unused variable ‘node’Quoted 6 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +2
-3-static ssize_t -store_cache_disable (struct _cpuid4_info *this_leaf,const char *buf, -size_t count) +static ssize_t store_cache_disable (struct _cpuid4_info *this_leaf, +const char *buf,size_t count,unsigned int index)arch/x86/kernel/cpu/intel_cacheinfo.c static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) · line 734
Quoted 10 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +4
-4{ -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -unsigned int ret, index, val; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned long val =0 ;CC
arch/x86/kernel/cpu/intel_cacheinfo.oarch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘store_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:737: warning: unused variable ‘cpu’
Quoted 8 lines
Quoted 4 lines
Oops it is also pointing to node:
arch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘store_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:739: warning: unused variable ‘node’Mind sending a fix? We really dont want warnings noise in the x86 tree.
Ok I send it.
when I fixed node then it gives warnings for :
arch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘show_cache_disable’:
arch/x86/kernel/cpu/intel_cacheinfo.c:713: warning: unused variable ‘cpu’
So I also fixed it.
Quoted 4 lines
I was also thinking about adopting the Sparc trick: adding -Werror to
arch/x86/files only, fixing all the fallout and ensuring no new warnings from a flag day on. Anyone interested in doing that?
Yes, this will be nice idea. Should I send patch for:
"adding -Werror to arch/x86/ files only"
Thanks,
--
JSR
Re: [tip:x86/cpu] x86: cacheinfo: replace sysfs interface for cache_disable feature
Quoted 99 lines · Jaswinder Singh Rajput
Quoted 80 lines · Ingo Molnar
Quoted 78 lines · Jaswinder Singh Rajput
Quoted 76 lines · Jaswinder Singh Rajput
Hello Mark,
Quoted 37 lines · Mark Langsdorf
Commit-ID: f8b201fc7110c3673437254e8ba02451461ece0b Gitweb: http://git.kernel.org/tip/f8b201fc7110c3673437254e8ba02451461ece0b Author: Mark Langsdorf <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:18:49 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:21:53 +0200x86: cacheinfo: replace sysfs interface for
cache_disablefeatureImpact: replace sysfs attribute
Current interface violates against "one-value-per-sysfs-attribute rule". This patch replaces current attribute with two attributes -- one for each L3 Cache Index Disable register.
Signed-off-by Mark Langsdorf <[email protected]>Signed-off-by Andreas Herrmann <[email protected]>Cc Andrew Morton <[email protected]>LKML-Reference: <[email protected]>
Signed-off-by Ingo Molnar <[email protected]>--- arch/x86/kernel/cpu/intel_cacheinfo.c | 90 ++++++++++++++++---------------- 1 files changed, 45 insertions(+), 45 deletions(-)arch/x86/kernel/cpu/intel_cacheinfo.c +2
-1@@ -697,73 +697,69 @@ static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf)#define to_object (k) container_of(k, struct _index_kobject, kobj)#define to_attr (a) container_of(a, struct _cache_attr, attr) -static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf) +static ssize_t show_cache_disable (struct _cpuid4_info *this_leaf,char *buf, +unsigned int index)arch/x86/kernel/cpu/intel_cacheinfo.c static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) · line 706
Quoted 10 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +4
-5{ -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -ssize_t ret =0 ; -int i; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned int reg =0 ;This leads to compilation warnings:
CCarch/x86/kernel/cpu/intel_cacheinfo.oarch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘show_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:712: warning: unused variable ‘node’Quoted 6 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +2
-3-static ssize_t -store_cache_disable (struct _cpuid4_info *this_leaf,const char *buf, -size_t count) +static ssize_t store_cache_disable (struct _cpuid4_info *this_leaf, +const char *buf,size_t count,unsigned int index)arch/x86/kernel/cpu/intel_cacheinfo.c static ssize_t show_type(struct _cpuid4_info *this_leaf, char *buf) · line 734
Quoted 10 lines
arch/x86/kernel/cpu/intel_cacheinfo.c +4
-4{ -const struct cpumask *mask =to_cpumask (this_leaf->shared_cpu_map); -int node =cpu_to_node (cpumask_first (mask)); -struct pci_dev *dev =NULL ; -unsigned int ret, index, val; +int cpu =cpumask_first (to_cpumask (this_leaf->shared_cpu_map)); +int node =cpu_to_node (cpu); +struct pci_dev *dev =node_to_k8_nb_misc (node); +unsigned long val =0 ;CC
arch/x86/kernel/cpu/intel_cacheinfo.oarch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘store_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:737: warning: unused variable ‘cpu’Quoted 8 lines
Quoted 4 lines
Oops it is also pointing to node:
arch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘store_cache_disable’:arch/x86/kernel/cpu/intel_cacheinfo.c:739: warning: unused variable ‘node’Mind sending a fix? We really dont want warnings noise in the x86 tree.
Ok I send it.
when I fixed node then it gives warnings for :
arch/x86/kernel/cpu/intel_cacheinfo.c: In function ‘show_cache_disable’:
arch/x86/kernel/cpu/intel_cacheinfo.c:713: warning: unused variable ‘cpu’So I also fixed it.
Quoted 7 lines
Quoted 4 lines
I was also thinking about adopting the Sparc trick: adding -Werror to
arch/x86/files only, fixing all the fallout and ensuring no new warnings from a flag day on. Anyone interested in doing that?Yes, this will be nice idea. Should I send patch for:
"adding -Werror toarch/x86/files only"
Yes, that would be nice. Perhaps also first fix all remaining warnings on 32-bit and 64-bit defconfig and allnoconfig. (i'll check allyesconfig - that takes a lot of time to build)
Ingo
PATCH 7/8 x86: cacheinfo: disable L3 ECC scrubbing when L3 cache index is disabled
In reply to Andreas HerrmannFrom: Mark Langsdorf <[email protected]>
x86: cacheinfo: disable L3 ECC scrubbing when L3 cache index is disabled
(Use correct mask to zero out bits 24-28 by Andreas)
--- arch/x86/kernel/cpu/intel_cacheinfo.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-)
arch/x86/kernel/cpu/intel_cacheinfo.c +6 -0
@@ -731,6 +731,7 @@ static ssize_t store_cache_disable(struct _cpuid4_info *this_leaf,int node =cpu_to_node (cpu);struct pci_dev *dev =node_to_k8_nb_misc (node);unsigned long val =0 ; +unsigned int scrubber =0 ;if (!this_leaf->can_disable)return -EINVAL ;
@@ -745,6 +746,11 @@ static ssize_t store_cache_disable(struct _cpuid4_info *this_leaf,return -EINVAL ; val |=0xc0000000 ; + +pci_read_config_dword (dev,0x58 , &scrubber); + scrubber &= ~0x1f000000 ; +pci_write_config_dword (dev,0x58 , scrubber); +pci_write_config_dword (dev,0x1BC + index *4 , val & ~0x40000000 );wbinvd ();pci_write_config_dword (dev,0x1BC + index *4 , val);
Signature · 1 line
1.6.2
[tip:x86/cpu] x86: cacheinfo: disable L3 ECC scrubbing when L3 cache index is disabled
Commit-ID: ba518bea2db21c72d44a6cbfd825b026ef9cdcb6 Gitweb: http://git.kernel.org/tip/ba518bea2db21c72d44a6cbfd825b026ef9cdcb6 Author: Mark Langsdorf <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:24:06 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:22:34 +0200
x86: cacheinfo: disable L3 ECC scrubbing when L3 cache index is disabled
(Use correct mask to zero out bits 24-28 by Andreas)
LKML-Reference: <[email protected]>
--- arch/x86/kernel/cpu/intel_cacheinfo.c | 6 ++++++ 1 files changed, 6 insertions(+), 0 deletions(-)
arch/x86/kernel/cpu/intel_cacheinfo.c +6 -0
@@ -731,6 +731,7 @@ static ssize_t store_cache_disable(struct _cpuid4_info *this_leaf,int node =cpu_to_node (cpu);struct pci_dev *dev =node_to_k8_nb_misc (node);unsigned long val =0 ; +unsigned int scrubber =0 ;if (!this_leaf->can_disable)return -EINVAL ;
@@ -745,6 +746,11 @@ static ssize_t store_cache_disable(struct _cpuid4_info *this_leaf,return -EINVAL ; val |=0xc0000000 ; + +pci_read_config_dword (dev,0x58 , &scrubber); + scrubber &= ~0x1f000000 ; +pci_write_config_dword (dev,0x58 , scrubber); +pci_write_config_dword (dev,0x1BC + index *4 , val & ~0x40000000 );wbinvd ();pci_write_config_dword (dev,0x1BC + index *4 , val);
PATCH 8/8 x86/docs: add description for cache_disable sysfs interface
In reply to Andreas HerrmannFrom: Mark Langsdorf <[email protected]>
--- .../ABI/testing/sysfs-devices-cache_disable | 18 ++++++++++++++++++ 1 files changed, 18 insertions(+), 0 deletions(-) create mode 100644 Documentation/ABI/testing/sysfs-devices-cache_disable
Oops, shouldn't Date and KernelVersion be adapted to reflect 2.6.30 and say June 2009.
I don't know.
Regards,
Andreas
Documentation/ABI/testing/sysfs-devices-cache_disable new file +18 -0
@@ -0,0 +1,18 @@ +What: /sys/devices/system/cpu/cpu*/cache/index*/cache_disable_X +Date: August 2008 +KernelVersion: 2.6.27 +Contact: [email protected] +Description: These files exist in every cpu's cache index directories. + There are currently 2 cache_disable_# files in each + directory. Reading from these files on a supported + processor will return that cache disable index value + for that processor and node. Writing to one of these + files will cause the specificed cache index to be disabled. + + Currently, only AMD Family 10h Processors support cache index + disable, and only for their L3 caches. See the BIOS and + Kernel Developer's Guide at + http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/31116-Public-GH-BKDG_3.20_2-4-09.pdf + for formatting information and other details on the + cache index disable. +Users: [email protected]
Signature · 1 line
1.6.2
[tip:x86/cpu] x86/docs: add description for cache_disable sysfs interface
Commit-ID: 2fad2d9bb8310889f3261035b594b4e068b6eb8b Gitweb: http://git.kernel.org/tip/2fad2d9bb8310889f3261035b594b4e068b6eb8b Author: Mark Langsdorf <[email protected]> AuthorDate: Thu, 9 Apr 2009 15:31:53 +0200 Committer: Ingo Molnar <[email protected]> CommitDate: Fri, 10 Apr 2009 14:22:40 +0200
x86/docs: add description for cache_disable sysfs interface
LKML-Reference: <[email protected]>
--- .../ABI/testing/sysfs-devices-cache_disable | 18 ++++++++++++++++++ 1 files changed, 18 insertions(+), 0 deletions(-)
Documentation/ABI/testing/sysfs-devices-cache_disable new file +18 -0
@@ -0,0 +1,18 @@ +What: /sys/devices/system/cpu/cpu*/cache/index*/cache_disable_X +Date: August 2008 +KernelVersion: 2.6.27 +Contact: [email protected] +Description: These files exist in every cpu's cache index directories. + There are currently 2 cache_disable_# files in each + directory. Reading from these files on a supported + processor will return that cache disable index value + for that processor and node. Writing to one of these + files will cause the specificed cache index to be disabled. + + Currently, only AMD Family 10h Processors support cache index + disable, and only for their L3 caches. See the BIOS and + Kernel Developer's Guide at + http://www.amd.com/us-en/assets/content_type/white_papers_and_tech_docs/31116-Public-GH-BKDG_3.20_2-4-09.pdf + for formatting information and other details on the + cache index disable. +Users: [email protected]
Ingo Molnar
AppliedQuoted 11 lines · Andreas Herrmann
These patches primarily split up Mark's latest patch
[PATCH][retry 6] Conform L3 Cache Index Disable to Linux Standards (see http://marc.info/?l=linux-kernel&m=123922420925908&w=2)
into smaller pieces. I prefer to revert the last two patches on tip/x86/cpu to get a cleaner changelog for the following fixes.
Patches are against tip/x86/cpu as of today.
Please apply.
thanks, applied - this series looks much nicer!
I rebased the topic to get rid of the double revert sequence, and have started testing it.
Thanks,
Ingo