login
Search
Hi, Lately a question has been bugging me: Why do we keep complicated(specific ioctls to set up and set free) LOOP driver to emulate a data file as block-device, for it to be mounted/formated. Wouldn't it be nice if we didnt have to specify '-o loop' or use losetup -d etc? With inspiration from UMS-Gadget and LOOP driver itself, I implemented a proof of concept: A kernel module which creats a node in /sys/devices/ and un/loads files to be emulated as block devices. The interface doesn't implement any new ioctls or a syscall. To load a file all we need to do is:- $ echo +filename > /sys/devices/virblk/manage //NOTE the - sign To unload a file:- $ echo -filename > /sys/devices/virblk/manage //NOTE the - sign The module creates one thread per node and alloc/free the data structures in the runtime thereby neither limiting the max number of files that could be emulated nor hogging space when not necessary. /*************************/ [root@localhost jmod]# dd if=/dev/zero of=image bs=1024 count=102400 [root@localhost jmod]# echo +image > /sys/devices/virblk/manage [root@localhost jmod]# cat /sys/devices/virblk/manage /home/jassi/jmod/image' as 'v0blk' [root@localhost jmod]# dmesg -c File(image) is being emulated as block device v0blk v0blk: unknown partition table [root@localhost jmod]# fdisk /dev/v0blk Device contains neither a valid DOS partition table, nor Sun, SGI or OSF disklabel Building a new DOS disklabel with disk identifier 0x360df22f. Changes will remain in memory only, until you decide to write them. After that, of course, the previous content won't be recoverable. Warning: invalid flag 0x0000 of partition table 4 will be corrected by w(rite) Command (m for help): n Command action e extended p primary partition (1-4) p Partition number (1-4): 1 First cylinder (1-100, default 1): Using default value 1 Last cylinder or +size or +sizeM or +sizeK (1-100, default 100): 50 Command (m for help): n Command action e ...
calibrated_APIC_clock still have two copies. it seems we should use 32 bit, because some system may not has tsc but it is some long than 64 bits... Thanks Yinghai Lu --
From: Cyrill Gorcunov <gorcunov@gmail.com>
There is no need to hold this code if CPU_HOTPLUG is not
defined
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_64.c | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/arch/x86/kernel/apic_64.c b/arch/x86/kernel/apic_64.c
index cd86085..8f55127 100644
--- a/arch/x86/kernel/apic_64.c
+++ b/arch/x86/kernel/apic_64.c
@@ -400,7 +400,7 @@ static void lapic_timer_broadcast(cpumask_t mask)
* Setup the local APIC timer for this CPU. Copy the initilized values
* of the boot CPU and register the clock event in the framework.
*/
-static void setup_APIC_timer(void)
+static void __cpuinit setup_APIC_timer(void)
{
struct clock_event_device *levt = &__get_cpu_var(lapic_events);
--
1.5.4.5
--
From: Cyrill Gorcunov <gorcunov@gmail.com>
We don't really use it now on 64bit mode but
could reserve it for future.
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_64.c | 10 ++++++++++
1 files changed, 10 insertions(+), 0 deletions(-)
diff --git a/arch/x86/kernel/apic_64.c b/arch/x86/kernel/apic_64.c
index 8f55127..bb46711 100644
--- a/arch/x86/kernel/apic_64.c
+++ b/arch/x86/kernel/apic_64.c
@@ -244,6 +244,16 @@ void __cpuinit enable_NMI_through_LVT0(void)
apic_write(APIC_LVT0, v);
}
+#ifdef CONFIG_X86_32
+/**
+ * get_physical_broadcast - Get number of physical broadcast IDs
+ */
+int get_physical_broadcast(void)
+{
+ return modern_apic() ? 0xff : 0xf;
+}
+#endif
+
/**
* lapic_get_maxlvt - get the maximum number of local vector table entries
*/
--
1.5.4.5
--
From: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_32.c | 10 ++++++++++
arch/x86/kernel/apic_64.c | 2 ++
2 files changed, 12 insertions(+), 0 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index c168cec..cca72da 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -1774,6 +1774,16 @@ static int __init parse_nolapic_timer(char *arg)
}
early_param("nolapic_timer", parse_nolapic_timer);
+#ifdef CONFIG_X86_64
+static __init int setup_apicpmtimer(char *s)
+{
+ apic_calibrate_pmtmr = 1;
+ notsc_setup(NULL);
+ return 0;
+}
+__setup("apicpmtimer", setup_apicpmtimer);
+#endif
+
static int __init apic_set_verbosity(char *arg)
{
if (!arg) {
diff --git a/arch/x86/kernel/apic_64.c b/arch/x86/kernel/apic_64.c
index bb46711..ddc5b24 100644
--- a/arch/x86/kernel/apic_64.c
+++ b/arch/x86/kernel/apic_64.c
@@ -1810,6 +1810,7 @@ static int __init parse_nolapic_timer(char *arg)
}
early_param("nolapic_timer", parse_nolapic_timer);
+#ifdef CONFIG_X86_64
static __init int setup_apicpmtimer(char *s)
{
apic_calibrate_pmtmr = 1;
@@ -1817,6 +1818,7 @@ static __init int setup_apicpmtimer(char *s)
return 0;
}
__setup("apicpmtimer", setup_apicpmtimer);
+#endif
static int __init apic_set_verbosity(char *arg)
{
--
1.5.4.5
--
From: Cyrill Gorcunov <gorcunov@gmail.com> Do not check for SPUTIOUS_APIC_VECTOR definition twice. Check it once - is what we need. Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com> Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> --- arch/x86/kernel/apic_64.c | 9 +++++++-- 1 files changed, 7 insertions(+), 2 deletions(-) diff --git a/arch/x86/kernel/apic_64.c b/arch/x86/kernel/apic_64.c index ddc5b24..4587e16 100644 --- a/arch/x86/kernel/apic_64.c +++ b/arch/x86/kernel/apic_64.c @@ -46,6 +46,13 @@ #include <mach_ipi.h> #include <mach_apic.h> +/* + * Sanity check + */ +#if ((SPURIOUS_APIC_VECTOR & 0x0F) != 0x0F) +# error SPURIOUS_APIC_VECTOR definition error +#endif + /* Disable local APIC timer from the kernel commandline or via dmi quirk */ static int disable_apic_timer __cpuinitdata; static int apic_calibrate_pmtmr __initdata; @@ -939,8 +946,6 @@ void __cpuinit setup_local_APIC(void) preempt_disable(); value = apic_read(APIC_LVR); - BUILD_BUG_ON((SPURIOUS_APIC_VECTOR & 0x0f) != 0x0f); - /* * Double-check whether this APIC is really registered. * This is meaningless in clustered apic mode, so we skip it. -- 1.5.4.5 --
From: Cyrill Gorcunov <gorcunov@gmail.com>
- remove useless read of APIC_LVR
- wrap with preempt_disable/enable
- check for integrated APIC just in place
v2: fix by Yinghai Lu.
fix lapic_is_intergrated using
let 64bit to have pic_mode
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_32.c | 21 ++++++++++++++----
arch/x86/kernel/apic_64.c | 51 ++++++++++++++++++++++++++++++++++++++++----
2 files changed, 62 insertions(+), 10 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index cca72da..42acb3a 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -1033,9 +1033,10 @@ static void __cpuinit lapic_setup_esr(void)
*/
void __cpuinit setup_local_APIC(void)
{
- unsigned long value, integrated;
+ unsigned int value;
int i, j;
+#ifdef CONFIG_X86_32
/* Pound the ESR really hard over the head with a big hammer - mbligh */
if (esr_disable) {
apic_write(APIC_ESR, 0);
@@ -1043,14 +1044,16 @@ void __cpuinit setup_local_APIC(void)
apic_write(APIC_ESR, 0);
apic_write(APIC_ESR, 0);
}
+#endif
- integrated = lapic_is_integrated();
+ preempt_disable();
/*
* Double-check whether this APIC is really registered.
+ * This is meaningless in clustered apic mode, so we skip it.
*/
if (!apic_id_registered())
- WARN_ON_ONCE(1);
+ BUG();
/*
* Intel recommends to set DFR, LDR and TPR before enabling
@@ -1096,6 +1099,7 @@ void __cpuinit setup_local_APIC(void)
*/
value |= APIC_SPIV_APIC_ENABLED;
+#ifdef CONFIG_X86_32
/*
* Some unknown Intel IO/APIC (or APIC) errata is biting us with
* certain networking cards. If high frequency interrupts are
@@ -1116,8 +1120,13 @@ void __cpuinit setup_local_APIC(void)
* See also the comment in end_level_ioapic_irq(). --macro
*/
- /* Enable focus processor (bit==0) */
+ /*
+ * - enable focus processor (bit==0)
+ * - 64bit mode always ...From: Cyrill Gorcunov <gorcunov@gmail.com>
All callers are __init or __cpuinit so there is no need
to hold this code without CPU_HOTPLUG being set.
Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com>
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_32.c | 6 +++---
1 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index 42acb3a..63bebb3 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -383,7 +383,7 @@ static void lapic_timer_broadcast(cpumask_t mask)
* Setup the local APIC timer for this CPU. Copy the initilized values
* of the boot CPU and register the clock event in the framework.
*/
-static void __devinit setup_APIC_timer(void)
+static void __cpuinit setup_APIC_timer(void)
{
struct clock_event_device *levt = &__get_cpu_var(lapic_events);
@@ -652,7 +652,7 @@ void __init setup_boot_APIC_clock(void)
setup_APIC_timer();
}
-void __devinit setup_secondary_APIC_clock(void)
+void __cpuinit setup_secondary_APIC_clock(void)
{
setup_APIC_timer();
}
@@ -1713,7 +1713,7 @@ static struct sys_device device_lapic = {
.cls = &lapic_sysclass,
};
-static void __devinit apic_pm_activate(void)
+static void __cpuinit apic_pm_activate(void)
{
apic_pm_state.active = 1;
}
--
1.5.4.5
--
From: Cyrill Gorcunov <gorcunov@gmail.com> Signed-off-by: Cyrill Gorcunov <gorcunov@gmail.com> Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> --- arch/x86/kernel/apic_32.c | 3 +++ arch/x86/kernel/apic_64.c | 2 ++ 2 files changed, 5 insertions(+), 0 deletions(-) diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c index 63bebb3..0f52bdd 100644 --- a/arch/x86/kernel/apic_32.c +++ b/arch/x86/kernel/apic_32.c @@ -718,6 +718,9 @@ void smp_apic_timer_interrupt(struct pt_regs *regs) * Besides, if we don't timer interrupts ignore the global * interrupt lock, which is the WrongThing (tm) to do. */ +#ifdef CONFIG_X86_64 + exit_idle(); +#endif irq_enter(); local_apic_timer_interrupt(); irq_exit(); diff --git a/arch/x86/kernel/apic_64.c b/arch/x86/kernel/apic_64.c index 283968d..2e10911 100644 --- a/arch/x86/kernel/apic_64.c +++ b/arch/x86/kernel/apic_64.c @@ -625,7 +625,9 @@ void smp_apic_timer_interrupt(struct pt_regs *regs) * Besides, if we don't timer interrupts ignore the global * interrupt lock, which is the WrongThing (tm) to do. */ +#ifdef CONFIG_X86_64 exit_idle(); +#endif irq_enter(); local_apic_timer_interrupt(); irq_exit(); -- 1.5.4.5 --
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_32.c | 46 ++++++++++++++++++++++--------------------
arch/x86/kernel/apic_64.c | 48 ++++++++++++++++++++++++++++++---------------
2 files changed, 56 insertions(+), 38 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index 0f52bdd..cbac5ff 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -50,16 +50,37 @@
# error SPURIOUS_APIC_VECTOR definition error
#endif
-unsigned long mp_lapic_addr;
-
+#ifdef CONFIG_X86_32
/*
* Knob to control our willingness to enable the local APIC.
*
* +1=force-enable
*/
static int force_enable_local_apic;
-int disable_apic;
+/*
+ * APIC command line parameters
+ */
+static int __init parse_lapic(char *arg)
+{
+ force_enable_local_apic = 1;
+ return 0;
+}
+early_param("lapic", parse_lapic);
+#endif
+#ifdef CONFIG_X86_64
+static int apic_calibrate_pmtmr __initdata;
+static __init int setup_apicpmtimer(char *s)
+{
+ apic_calibrate_pmtmr = 1;
+ notsc_setup(NULL);
+ return 0;
+}
+__setup("apicpmtimer", setup_apicpmtimer);
+#endif
+
+unsigned long mp_lapic_addr;
+int disable_apic;
/* Disable local APIC timer from the kernel commandline or via dmi quirk */
static int disable_apic_timer __cpuinitdata;
/* Local APIC timer works in C2 */
@@ -1742,15 +1763,6 @@ static void apic_pm_activate(void) { }
#endif /* CONFIG_PM */
-/*
- * APIC command line parameters
- */
-static int __init parse_lapic(char *arg)
-{
- force_enable_local_apic = 1;
- return 0;
-}
-early_param("lapic", parse_lapic);
static int __init setup_disableapic(char *arg)
{
@@ -1788,16 +1800,6 @@ static int __init parse_nolapic_timer(char *arg)
}
early_param("nolapic_timer", parse_nolapic_timer);
-#ifdef CONFIG_X86_64
-static __init int setup_apicpmtimer(char *s)
-{
- apic_calibrate_pmtmr = 1;
- notsc_setup(NULL);
- return 0;
-}
-__setup("apicpmtimer", setup_apicpmtimer);
-#endif
-
...Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_64.c | 35 +++++++++++++++++++++++------------
1 files changed, 23 insertions(+), 12 deletions(-)
diff --git a/arch/x86/kernel/apic_64.c b/arch/x86/kernel/apic_64.c
index c40a900..d3ec746 100644
--- a/arch/x86/kernel/apic_64.c
+++ b/arch/x86/kernel/apic_64.c
@@ -82,10 +82,23 @@ static __init int setup_apicpmtimer(char *s)
__setup("apicpmtimer", setup_apicpmtimer);
#endif
-int disable_x2apic;
+#ifdef CONFIG_X86_64
+#define HAVE_X2APIC
+#endif
+
+#ifdef HAVE_X2APIC
int x2apic;
/* x2apic enabled before OS handover */
int x2apic_preenabled;
+int disable_x2apic;
+static __init int setup_nox2apic(char *str)
+{
+ disable_x2apic = 1;
+ setup_clear_cpu_cap(X86_FEATURE_X2APIC);
+ return 0;
+}
+early_param("nox2apic", setup_nox2apic);
+#endif
unsigned long mp_lapic_addr;
int disable_apic;
@@ -228,6 +241,7 @@ static struct apic_ops xapic_ops = {
struct apic_ops __read_mostly *apic_ops = &xapic_ops;
EXPORT_SYMBOL_GPL(apic_ops);
+#ifdef HAVE_X2APIC
static void x2apic_wait_icr_idle(void)
{
/* no need to wait for icr idle in x2apic */
@@ -261,6 +275,7 @@ static struct apic_ops x2apic_ops = {
.wait_icr_idle = x2apic_wait_icr_idle,
.safe_wait_icr_idle = safe_x2apic_wait_icr_idle,
};
+#endif
/**
* enable_NMI_through_LVT0 - enable NMI through local vector table 0
@@ -1125,6 +1140,7 @@ void __cpuinit end_local_APIC_setup(void)
apic_pm_activate();
}
+#ifdef HAVE_X2APIC
void check_x2apic(void)
{
int msr, msr2;
@@ -1243,6 +1259,7 @@ end:
return;
}
+#endif /* HAVE_X2APIC */
/*
* Detect and enable local APICs on non-SMP boards.
@@ -1291,10 +1308,12 @@ void __init early_init_lapic_mapping(void)
*/
void __init init_apic_mappings(void)
{
+#ifdef HAVE_X2APIC
if (x2apic) {
boot_cpu_physical_apicid = read_apic_id();
return;
}
+#endif
/*
* If no local APIC can be found then set up a fake all
@@ -1335,8 +1354,9 @@ int ...Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_32.c | 7 +++++++
arch/x86/kernel/apic_64.c | 2 ++
2 files changed, 9 insertions(+), 0 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index cbac5ff..6fb2b45 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -1600,6 +1600,13 @@ void __cpuinit generic_processor_info(int apicid, int version)
cpu_set(cpu, cpu_present_map);
}
+#ifdef CONFIG_X86_64
+int hard_smp_processor_id(void)
+{
+ return read_apic_id();
+}
+#endif
+
/*
* Power management
*/
diff --git a/arch/x86/kernel/apic_64.c b/arch/x86/kernel/apic_64.c
index d3ec746..eeb6983 100644
--- a/arch/x86/kernel/apic_64.c
+++ b/arch/x86/kernel/apic_64.c
@@ -1612,10 +1612,12 @@ void __cpuinit generic_processor_info(int apicid, int version)
cpu_set(cpu, cpu_present_map);
}
+#ifdef CONFIG_X86_64
int hard_smp_processor_id(void)
{
return read_apic_id();
}
+#endif
/*
* Power management
--
1.5.4.5
--
except x2apic, detec_init_APIC, and calibrating_APIC_clock
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_32.c | 125 ++++++++++++++++++++++++++++++++++++++++----
arch/x86/kernel/apic_64.c | 10 +++-
2 files changed, 122 insertions(+), 13 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index 6fb2b45..294256f 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -66,6 +66,9 @@ static int __init parse_lapic(char *arg)
return 0;
}
early_param("lapic", parse_lapic);
+/* Local APIC was disabled by the BIOS and enabled by the kernel */
+static int enabled_via_apicbase;
+
#endif
#ifdef CONFIG_X86_64
@@ -131,9 +134,6 @@ static struct clock_event_device lapic_clockevent = {
};
static DEFINE_PER_CPU(struct clock_event_device, lapic_events);
-/* Local APIC was disabled by the BIOS and enabled by the kernel */
-static int enabled_via_apicbase;
-
static unsigned long apic_phys;
/*
@@ -240,6 +240,7 @@ void __cpuinit enable_NMI_through_LVT0(void)
apic_write(APIC_LVT0, v);
}
+#ifdef CONFIG_X86_32
/**
* get_physical_broadcast - Get number of physical broadcast IDs
*/
@@ -247,6 +248,7 @@ int get_physical_broadcast(void)
{
return modern_apic() ? 0xff : 0xf;
}
+#endif
/**
* lapic_get_maxlvt - get the maximum number of local vector table entries
@@ -1291,6 +1293,32 @@ no_apic:
return -1;
}
+#ifdef CONFIG_X86_64
+void __init early_init_lapic_mapping(void)
+{
+ unsigned long phys_addr;
+
+ /*
+ * If no local APIC can be found then go out
+ * : it means there is no mpatable and MADT
+ */
+ if (!smp_found_config)
+ return;
+
+ phys_addr = mp_lapic_addr;
+
+ set_fixmap_nocache(FIX_APIC_BASE, phys_addr);
+ apic_printk(APIC_VERBOSE, "mapped APIC to %16lx (%16lx)\n",
+ APIC_BASE, phys_addr);
+
+ /*
+ * Fetch the APIC ID of the BSP in case we have a
+ * default configuration (or the MP table is broken).
+ ...Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_32.c | 51 +++++++++++++++++++++++++++++++++++++++-----
arch/x86/kernel/apic_64.c | 48 +++++++++++++++++++++++++++++++++++++++--
2 files changed, 90 insertions(+), 9 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index 294256f..0d9c000 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -1355,6 +1355,17 @@ int apic_version[MAX_APICS];
int __init APIC_init_uniprocessor(void)
{
+#ifdef CONFIG_X86_64
+ if (disable_apic) {
+ printk(KERN_INFO "Apic disabled\n");
+ return -1;
+ }
+ if (!cpu_has_apic) {
+ disable_apic = 1;
+ printk(KERN_INFO "Apic disabled by BIOS\n");
+ return -1;
+ }
+#else
if (!smp_found_config && !cpu_has_apic)
return -1;
@@ -1368,34 +1379,62 @@ int __init APIC_init_uniprocessor(void)
clear_cpu_cap(&boot_cpu_data, X86_FEATURE_APIC);
return -1;
}
+#endif
+#ifdef HAVE_X2APIC
+ enable_IR_x2apic();
+#endif
+#ifdef CONFIG_X86_64
+ setup_apic_routing();
+#endif
verify_local_APIC();
-
connect_bsp_APIC();
+#ifdef CONFIG_X86_64
+ apic_write(APIC_ID, SET_APIC_ID(boot_cpu_physical_apicid));
+#else
/*
* Hack: In case of kdump, after a crash, kernel might be booting
* on a cpu with non-zero lapic id. But boot_cpu_physical_apicid
* might be zero if read from MP tables. Get it from LAPIC.
*/
-#ifdef CONFIG_CRASH_DUMP
+# ifdef CONFIG_CRASH_DUMP
boot_cpu_physical_apicid = read_apic_id();
+# endif
#endif
physid_set_mask_of_physid(boot_cpu_physical_apicid, &phys_cpu_present_map);
-
setup_local_APIC();
+#ifdef CONFIG_X86_64
+ /*
+ * Now enable IO-APICs, actually call clear_IO_APIC
+ * We need clear_IO_APIC before enabling vector on BP
+ */
+ if (!skip_ioapic_setup && nr_ioapics)
+ enable_IO_APIC();
+#endif
+
#ifdef CONFIG_X86_IO_APIC
if (!smp_found_config || skip_ioapic_setup || !nr_ioapics)
#endif
localise_nmi_watchdog();
...Signed-off-by: Yinghai Lu <yhlu.kernel@mgail.com>
---
arch/x86/kernel/apic_32.c | 20 +++++++++++
arch/x86/kernel/apic_64.c | 82 ++++++++++++++++++++++++++++++++++++++++++++-
2 files changed, 101 insertions(+), 1 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index 0d9c000..87566a3 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -1214,6 +1214,25 @@ void __cpuinit end_local_APIC_setup(void)
apic_pm_activate();
}
+#ifdef CONFIG_X86_64
+/*
+ * Detect and enable local APICs on non-SMP boards.
+ * Original code written by Keir Fraser.
+ * On AMD64 we trust the BIOS - if it says no APIC it is likely
+ * not correctly set up (usually the APIC timer won't work etc.)
+ */
+static int __init detect_init_APIC(void)
+{
+ if (!cpu_has_apic) {
+ printk(KERN_INFO "No local APIC present\n");
+ return -1;
+ }
+
+ mp_lapic_addr = APIC_DEFAULT_PHYS_BASE;
+ boot_cpu_physical_apicid = 0;
+ return 0;
+}
+#else
/*
* Detect and initialize APIC
*/
@@ -1292,6 +1311,7 @@ no_apic:
printk(KERN_INFO "No local APIC present or hardware disabled\n");
return -1;
}
+#endif
#ifdef CONFIG_X86_64
void __init early_init_lapic_mapping(void)
diff --git a/arch/x86/kernel/apic_64.c b/arch/x86/kernel/apic_64.c
index 16a30c2..b7268f5 100644
--- a/arch/x86/kernel/apic_64.c
+++ b/arch/x86/kernel/apic_64.c
@@ -684,7 +684,6 @@ int setup_profiling_timer(unsigned int multiplier)
return -EINVAL;
}
-
/*
* Local APIC start and shutdown
*/
@@ -1264,6 +1263,7 @@ end:
}
#endif /* HAVE_X2APIC */
+#ifdef CONFIG_X86_64
/*
* Detect and enable local APICs on non-SMP boards.
* Original code written by Keir Fraser.
@@ -1281,6 +1281,86 @@ static int __init detect_init_APIC(void)
boot_cpu_physical_apicid = 0;
return 0;
}
+#else
+/*
+ * Detect and initialize APIC
+ */
+static int __init detect_init_APIC(void)
+{
+ u32 h, l, features;
+
+ /* Disabled by kernel option? */
+ if ...Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> --- arch/x86/kernel/apic_32.c | 8 ++++++++ arch/x86/kernel/apic_64.c | 7 ++++++- 2 files changed, 14 insertions(+), 1 deletions(-) diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c index 87566a3..749adb5 100644 --- a/arch/x86/kernel/apic_32.c +++ b/arch/x86/kernel/apic_32.c @@ -23,11 +23,13 @@ #include <linux/mc146818rtc.h> #include <linux/kernel_stat.h> #include <linux/sysdev.h> +#include <linux/ioport.h> #include <linux/cpu.h> #include <linux/clockchips.h> #include <linux/acpi_pmtmr.h> #include <linux/module.h> #include <linux/dmi.h> +#include <linux/dmar.h> #include <asm/atomic.h> #include <asm/smp.h> @@ -36,8 +38,14 @@ #include <asm/desc.h> #include <asm/arch_hooks.h> #include <asm/hpet.h> +#include <asm/pgalloc.h> #include <asm/i8253.h> #include <asm/nmi.h> +#include <asm/idle.h> +#include <asm/proto.h> +#include <asm/timex.h> +#include <asm/apic.h> +#include <asm/i8259.h> #include <mach_apic.h> #include <mach_apicdef.h> diff --git a/arch/x86/kernel/apic_64.c b/arch/x86/kernel/apic_64.c index b7268f5..ebe417b 100644 --- a/arch/x86/kernel/apic_64.c +++ b/arch/x86/kernel/apic_64.c @@ -24,9 +24,11 @@ #include <linux/kernel_stat.h> #include <linux/sysdev.h> #include <linux/ioport.h> +#include <linux/cpu.h> #include <linux/clockchips.h> #include <linux/acpi_pmtmr.h> #include <linux/module.h> +#include <linux/dmi.h> #include <linux/dmar.h> #include <asm/atomic.h> @@ -34,8 +36,10 @@ #include <asm/mtrr.h> #include <asm/mpspec.h> #include <asm/desc.h> +#include <asm/arch_hooks.h> #include <asm/hpet.h> #include <asm/pgalloc.h> +#include <asm/i8253.h> #include <asm/nmi.h> #include <asm/idle.h> #include <asm/proto.h> @@ -43,8 +47,9 @@ #include <asm/apic.h> #include <asm/i8259.h> -#include <mach_ipi.h> #include <mach_apic.h> +#include <mach_apicdef.h> +#include <mach_ipi.h> /* * Sanity check -- 1.5.4.5 --
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_32.c | 24 +++++++++++++++++++++---
arch/x86/kernel/apic_64.c | 24 ++++++++++++++++++++++--
2 files changed, 43 insertions(+), 5 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index 749adb5..d357038 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -1474,10 +1474,17 @@ int __init APIC_init_uniprocessor(void)
/*
* This interrupt should _never_ happen with our APIC/SMP architecture
*/
+#ifdef CONFIG_X86_64
+asmlinkage void smp_spurious_interrupt(void)
+#else
void smp_spurious_interrupt(struct pt_regs *regs)
+#endif
{
- unsigned long v;
+ u32 v;
+#ifdef CONFIG_X86_64
+ exit_idle();
+#endif
irq_enter();
/*
* Check if this really is a spurious interrupt and ACK it
@@ -1488,20 +1495,31 @@ void smp_spurious_interrupt(struct pt_regs *regs)
if (v & (1 << (SPURIOUS_APIC_VECTOR & 0x1f)))
ack_APIC_irq();
+#ifdef CONFIG_X86_64
+ add_pda(irq_spurious_count, 1);
+#else
/* see sw-dev-man vol 3, chapter 7.4.13.5 */
printk(KERN_INFO "spurious APIC interrupt on CPU#%d, "
"should never happen.\n", smp_processor_id());
__get_cpu_var(irq_stat).irq_spurious_count++;
+#endif
irq_exit();
}
/*
* This interrupt should never happen with our APIC/SMP architecture
*/
+#ifdef CONFIG_X86_64
+asmlinkage void smp_error_interrupt(void)
+#else
void smp_error_interrupt(struct pt_regs *regs)
+#endif
{
- unsigned long v, v1;
+ u32 v, v1;
+#ifdef CONFIG_X86_64
+ exit_idle();
+#endif
irq_enter();
/* First tickle the hardware, only then report what went on. -- REW */
v = apic_read(APIC_ESR);
@@ -1520,7 +1538,7 @@ void smp_error_interrupt(struct pt_regs *regs)
6: Received illegal vector
7: Illegal register address
*/
- printk(KERN_DEBUG "APIC error on CPU%d: %02lx(%02lx)\n",
+ printk(KERN_DEBUG "APIC error on CPU%d: %02x(%02x)\n",
smp_processor_id(), v , v1);
irq_exit();
...Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_32.c | 86 ++++++++++++++++++
arch/x86/kernel/apic_64.c | 215 +++++++++++++++++++++++++++++++++++++++++++++
2 files changed, 301 insertions(+), 0 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index d357038..b21b772 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -424,6 +424,90 @@ static void __cpuinit setup_APIC_timer(void)
clockevents_register_device(levt);
}
+#ifdef CONFIG_X86_64
+/*
+ * In this function we calibrate APIC bus clocks to the external
+ * timer. Unfortunately we cannot use jiffies and the timer irq
+ * to calibrate, since some later bootup code depends on getting
+ * the first irq? Ugh.
+ *
+ * We want to do the calibration only once since we
+ * want to have local timer irqs syncron. CPUs connected
+ * by the same APIC bus have the very same bus frequency.
+ * And we want to have irqs off anyways, no accidental
+ * APIC irq that way.
+ */
+
+#define TICK_COUNT 100000000
+
+static int __init calibrate_APIC_clock(void)
+{
+ unsigned apic, apic_start;
+ unsigned long tsc, tsc_start;
+ int result;
+
+ local_irq_disable();
+
+ /*
+ * Put whatever arbitrary (but long enough) timeout
+ * value into the APIC clock, we just want to get the
+ * counter running for calibration.
+ *
+ * No interrupt enable !
+ */
+ __setup_APIC_LVTT(250000000, 0, 0);
+
+ apic_start = apic_read(APIC_TMCCT);
+#ifdef CONFIG_X86_PM_TIMER
+ if (apic_calibrate_pmtmr && pmtmr_ioport) {
+ pmtimer_wait(5000); /* 5ms wait */
+ apic = apic_read(APIC_TMCCT);
+ result = (apic_start - apic) * 1000L / 5;
+ } else
+#endif
+ {
+ rdtscll(tsc_start);
+
+ do {
+ apic = apic_read(APIC_TMCCT);
+ rdtscll(tsc);
+ } while ((tsc - tsc_start) < TICK_COUNT &&
+ (apic_start - apic) < TICK_COUNT);
+
+ result = (apic_start - apic) * 1000L * tsc_khz /
+ (tsc - ...Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/apic_32.c | 188 +++++++++++++++++++++++++++++++++++++++++++++
1 files changed, 188 insertions(+), 0 deletions(-)
diff --git a/arch/x86/kernel/apic_32.c b/arch/x86/kernel/apic_32.c
index b21b772..6e99af5 100644
--- a/arch/x86/kernel/apic_32.c
+++ b/arch/x86/kernel/apic_32.c
@@ -90,6 +90,24 @@ static __init int setup_apicpmtimer(char *s)
__setup("apicpmtimer", setup_apicpmtimer);
#endif
+#ifdef CONFIG_X86_64
+#define HAVE_X2APIC
+#endif
+
+#ifdef HAVE_X2APIC
+int x2apic;
+/* x2apic enabled before OS handover */
+int x2apic_preenabled;
+int disable_x2apic;
+static __init int setup_nox2apic(char *str)
+{
+ disable_x2apic = 1;
+ setup_clear_cpu_cap(X86_FEATURE_X2APIC);
+ return 0;
+}
+early_param("nox2apic", setup_nox2apic);
+#endif
+
unsigned long mp_lapic_addr;
int disable_apic;
/* Disable local APIC timer from the kernel commandline or via dmi quirk */
@@ -231,6 +249,42 @@ static struct apic_ops xapic_ops = {
struct apic_ops __read_mostly *apic_ops = &xapic_ops;
EXPORT_SYMBOL_GPL(apic_ops);
+#ifdef HAVE_X2APIC
+static void x2apic_wait_icr_idle(void)
+{
+ /* no need to wait for icr idle in x2apic */
+ return;
+}
+
+static u32 safe_x2apic_wait_icr_idle(void)
+{
+ /* no need to wait for icr idle in x2apic */
+ return 0;
+}
+
+void x2apic_icr_write(u32 low, u32 id)
+{
+ wrmsrl(APIC_BASE_MSR + (APIC_ICR >> 4), ((__u64) id) << 32 | low);
+}
+
+u64 x2apic_icr_read(void)
+{
+ unsigned long val;
+
+ rdmsrl(APIC_BASE_MSR + (APIC_ICR >> 4), val);
+ return val;
+}
+
+static struct apic_ops x2apic_ops = {
+ .read = native_apic_msr_read,
+ .write = native_apic_msr_write,
+ .icr_read = x2apic_icr_read,
+ .icr_write = x2apic_icr_write,
+ .wait_icr_idle = x2apic_wait_icr_idle,
+ .safe_wait_icr_idle = safe_x2apic_wait_icr_idle,
+};
+#endif
+
/**
* enable_NMI_through_LVT0 - enable NMI through local vector table 0
*/
@@ -1308,6 +1362,127 @@ void __cpuinit ...Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
---
arch/x86/kernel/Makefile | 2 +-
arch/x86/kernel/{apic_32.c => apic.c} | 0
arch/x86/kernel/Makefile | 2
arch/x86/kernel/apic.c | 2311 ++++++++++++++++++++++++++++++++++++++++++++++
arch/x86/kernel/apic_32.c | 2311 ----------------------------------------------
arch/x86/kernel/apic_64.c | 2311 ----------------------------------------------
4 files changed, 2312 insertions(+), 4623 deletions(-)
rename arch/x86/kernel/{apic_32.c => apic.c} (100%)
delete mode 100644 arch/x86/kernel/apic_64.c
Index: linux-2.6/arch/x86/kernel/Makefile
===================================================================
--- linux-2.6.orig/arch/x86/kernel/Makefile
+++ linux-2.6/arch/x86/kernel/Makefile
@@ -64,7 +64,7 @@ obj-$(CONFIG_X86_32_SMP) += smpcommon.o
obj-$(CONFIG_X86_64_SMP) += tsc_sync.o smpcommon.o
obj-$(CONFIG_X86_TRAMPOLINE) += trampoline_$(BITS).o
obj-$(CONFIG_X86_MPPARSE) += mpparse.o
-obj-$(CONFIG_X86_LOCAL_APIC) += apic_$(BITS).o nmi.o
+obj-$(CONFIG_X86_LOCAL_APIC) += apic.o nmi.o
obj-$(CONFIG_X86_IO_APIC) += io_apic.o
obj-$(CONFIG_X86_REBOOTFIXUPS) += reboot_fixups_32.o
obj-$(CONFIG_DYNAMIC_FTRACE) += ftrace.o
Index: linux-2.6/arch/x86/kernel/apic_64.c
===================================================================
--- linux-2.6.orig/arch/x86/kernel/apic_64.c
+++ /dev/null
@@ -1,2311 +0,0 @@
-/*
- * Local APIC handling, local APIC timers
- *
- * (c) 1999, 2000 Ingo Molnar <mingo@redhat.com>
- *
- * Fixes
- * Maciej W. Rozycki : Bits for genuine 82489DX APICs;
- * thanks to Eric Gilmore
- * and Rolf G. Tews
- * for testing these extensively.
- * Maciej W. Rozycki : Various updates and fixes.
- * Mikael Pettersson : Power Management for UP-APIC.
- * Pavel Machek and
- * Mikael Pettersson : PM converted to driver model.
- */
-
-#include <linux/init.h>
-
-#include <linux/mm.h>
-#include <linux/delay.h>
-#include <linux/bootmem.h>
-#include ...correct, the 32-bit calibration code looks more mature, so we should pick that. (after making sure it's properly 64-bit clean) Ingo --
very nice series! I've applied it to tip/irq/sparseirq and did a test-integration with tip/master - lets see how it behaves in -tip testing. Ingo --
Hi,
I like this approach: it is nice that you don't require a special tool
like losetup with it, and that virblk can handle partitioned disks is
a nice bonus.
Some minor remarks below.
[ disclaimer: I am not an expert in this and might talk nonsense ]
Shouldn't this output be replaced by some format which is easier to
parse? This would make it easier for tools like mount to
Kernel coding style seems to prefer spaces after "if" and before "{"
Can the list change while we are iterating over it? If so, do wee
Maybe return more informative error codes here, e.g. -EBUSY if
Can the list change while we are iterating over it? If so, do we need
The example code in "Linux Device Drivers", third edition, p.469 uses
KERN_WARNING and returns -EBUSY in this case. No idea whether this is
I hope this helps,
Jochen
--
http://seehuhn.de/
--
It's better to check this patch with scripts/checkpatch.pl - it will tell you You seem to use this semaphore as a mutex - why don't you convert it to mutex then? Semaphores should be used only when they are absolutely necessary. Marcin --
This seems like a really interesting idea, but will it interface with crypto applications like LUKS? Crypto seems to be a vital feature for anything related to devices and/or filesystems if an approach is going to be more than a niche slution. -- Bill Davidsen <davidsen@tmr.com> "We have more to fear from the bungling of the incompetent than from the machinations of the wicked." - from Slashdot --
Can you please expand a bit why you think losetup is that complicated and what the problem is with it? AFAIK you're essentially just moving a minimal version of losetup (with missing features like no offsets etc.) into the kernel and frankly I fail to see the beauty in that. Or rather if you start with losetup, why stop at mount, modprobe, ifconfig, mkfs, fsck, ls[1], ...? For me it seems more that most of the file system based command interfaces (/proc/mtrr comes to mind) are quite hard to use and I prefer a proper command line tool with a manpage and --help and a real parser any day. -Andi [1] I'm sure someone could come up with some scheme to do ls using sysfs and you could find someone on this list who said "cool" :) --
Hi Andi, Before starting i would like to point out the 'philosophy' behind the idea. Exceptions(special cases) make for entropy and hence complexity. Generalization keeps uniformity and hences Order and Ease. We can't escape 'entropy' when emulating something that it is actually not, nevertheless we can minimize it by introducing as least and as localized changes as possible. My idea congregates only determining changes at the lowest level(driver), unlike the respectable LOOP driver which hardcodes limits and adds to the entropy(new ioctls, special utilities etc). I am sure you, being a vetran, understand the point well. Sir, i don't object to losetup as such. It is a utility made(specifically?) for LOOP devices: which implements unnecessary gears and switches to operate. Its simpler than that :) All i do is implement non-ioctl method to load/unload a file as a (emulated)block device. And alloc/free resources for them on need basis rather than limiting them at driver-module load-time. Further about some features of LOOP drivers like encryption: i think they are better done at filesystem level. I believe any operation that could be done on a real block device shud be possible on an emulated one. And exactly that. If there is any need to do some stuff(offsets, encryption etc) then that cud be done on a real block and we would surely already have utilities for doing that on real devices(and hences on a transparent emulated device). I am unable to think of something practical that can _only_ be done using the 'offset feature' of losetup on /dev/loop AND which can not be accomplished for (emulated or not)block devices. The point is to only make least intrusive and most generic code, for making easy the inheritance of standard system-wide features, utilities and usages. I am not starting with any application. It would indeed be cool if the one doesn't make use of 'cat' 'grep' et al and the method is more efficient. Otherwise it would just be four motobikes tugging a car ...
I fail to see what your patch generalizes? AFAIK it just adds a new Your goal is to replace all ioctls with sysfs files? If it's that then I'm going on the book as saying that's a bad idea, especially for this case. While ioctls have their problems they work quite well for many things. I don't see any particular reason why ioctls should not be used to configure loop devices. -Andi --
Btw, my code is not a patch to the loop driver, its an altogether new module. May i daresay, it aims squarely at drivers/block/loop.c and mean to replace it altogether.
My module generalizes in the way that it doesn't add or make use of any ioctl. It doesn't even export a variable and makes uses only of what other subsystems provide for(block, sysfs, vfs).
Please do have a look at the code.
I add only one sysfs interface (manage), which when read returns the status of all the files being emulated and when written updates(add/remove) emulation of a file.
The interface could be made more useful by echo'ing in other parameters along with filename for example:
echo +[r/w]+[sects]+[cyls]+[heads]+[filename] > /sys/devices/virblk/manage
I don't intend to revolt against the concept of ioctls. Being an embedded linux engineer I do understand the importance of ioctls: i declare one every other day for configuring custom h/w: Graphics and Multimedia esp so.
As for loop devices, i think we can make do without'em.
Regards,
-Jassi
--
> As far as features are concerned, please suggest me what could be done with /dev/loop0 and not for /dev/vblk? Serious question? cryptoloop, offsets are the big ticket items, i'm sure there are more. -Andi --
As far as cryptoloop goes, isnt that obsoleted anyway? is there anyone that doesent use dm-crypt by now? as for offsets, isnt this a logical thing to do with devicemapper aswell? ofcourse loop/losetup cannot just be removed, theres probably LOTS of --
Cryptoloop? No, my code isn't the 'same' as loop.c and hence can't provide for cryptoloop. I lose, you win :)
Cryptoloop was designed so b'coz 'loop' feature was available, not the other way round. Had we not had '-o loop', developers would have devised mechanism for encrypting block devices rather than exploiting the -o loop hack. In retrospect, it seems Cryptoloop was a bad design based upon a hack.
Encryption? Yes, my code does provide a way for it. It supports better dm-crypt/LUKS off the shelf.
I see one mail from Andrew Morton suggesting removal of cryptoloop altogether in order to ramp up its obsoletion! I am only providing a more-transparent-way of using dm-crypt.
I assume the concede Kasper's point on 'offset' feature of loop.c.
Anyways, seems we are moving from discussion to debate. You've got the whole picture and now its purely your call to approve of it or not.
Better still, you could suggest how to make it better: we are for sure gonna drop cryptoloop for dm-crypt. What wud we do with the then-unnecessary crypt support of loop.c?
Or maybe, for the time being, it would be better to modify Loop.c to allow for runtime alloc/free of as many resources as needed and provide for sysfs interface to add/remove 'loop' rather than ioctls?
I had an idea and i cooked up a code. I wanted to put it somewhere for someone to stumble upon the idea sometime. I am done.
Best Regards,
-Jassi
--
I disagree, as i recall, i stopped using cryptoloop and went directly to In the case that i am correct above, it would appear that i did not loose, given that the newer stuff has backwards compatibility (which i am fairly certain it does) also do not forget that i did say its unlikely that loop be removed due --
I, for one, have never been able to figure out how to use dm-crypt with cryptoloop files. If I had, for instance, a file and knew that if I started at a certain offset using a certain encryption with a certain key, and I wanted to mount my part of it, I have never persuaded dm-crypt to do that. All proposed solutions have started with copy or convert, neither of -- Bill Davidsen <davidsen@tmr.com> "We have more to fear from the bungling of the incompetent than from the machinations of the wicked." - from Slashdot --
