kernel_init
- The first function called by
kernel_initis wait_for_completion, it checks completionkthreadd_done.kthreadd_doneis updated after kernel daemon thread created, it's used to ignore race conditon between root thread and init thread. Ifkthead_doneisn't updated, the eventkthreadd_done.done is zero,kernel_initwon't continue its execution until timeout orkthreadd_done.doneis non-zero.
Breakpoint 2, kernel_init (unused=0x0) at init/main.c:859
859 wait_for_completion(&kthreadd_done);
(gdb) s
wait_for_completion (x=x@entry=0xc1730480 <kthreadd_done>)
at kernel/sched.c:6091
6091 wait_for_common(x, MAX_SCHEDULE_TIMEOUT, TASK_UNINTERRUPTIBLE);
(gdb) s
wait_for_common (x=x@entry=0xc1730480 <kthreadd_done>,
timeout=timeout@entry=2147483647, state=state@entry=2)
at kernel/sched.c:6070
6070 {
(gdb) n
6071 might_sleep(); # might be preemted
(gdb)
6073 spin_lock_irq(&x->wait.lock);
(gdb)
6074 timeout = do_wait_for_common(x, timeout, state);
(gdb) s
do_wait_for_common (state=2, timeout=2147483647, x=0xc1730480 <kthreadd_done>)
at kernel/sched.c:6045
6045 if (!x->done) {
(gdb) p *x
$1 = {done = 1, wait = {lock = {raw_lock = {slock = 513}, magic = 3735899821,
owner_cpu = 0, owner = 0xc7070000, dep_map = {
key = 0xc1730494 <kthreadd_done+20>,
class_cache = 0xc1b5f8d0 <lock_classes+151696>,
name = 0xc15c53c4 "(kthreadd_done).wait.lock", cpu = 0,
ip = 3242755061}}, task_list = {next = 0xc17304a8 <kthreadd_done+40>,
prev = 0xc17304a8 <kthreadd_done+40>}}}
(gdb) n
6064 x->done--;
(gdb)
6065 return timeout ?: 1;
(gdb)
wait_for_common (x=x@entry=0xc1730480 <kthreadd_done>,
timeout=<optimized out>, timeout@entry=2147483647, state=state@entry=2)
at kernel/sched.c:6075
6075 spin_unlock_irq(&x->wait.lock);
(gdb)
6077 }
(gdb)
wait_for_completion (x=x@entry=0xc1730480 <kthreadd_done>)
at kernel/sched.c:6092
6092 }
- Call
lock_kernelto get BKL
kernel_init (unused=<optimized out>) at init/main.c:860
860 lock_kernel();
(gdb) s
lock_kernel () at lib/kernel_lock.c:117
117 {
(gdb) n
118 int depth = current->lock_depth+1;
(gdb)
119 if (likely(!depth))
(gdb) p depth
$2 = 0
(gdb) n
120 __lock_kernel();
(gdb) s
__lock_kernel () at lib/kernel_lock.c:96
96 _raw_spin_lock(&kernel_flag);
(gdb) p kernel_flag
$3 = {raw_lock = {slock = 257}, magic = 3735899821, owner_cpu = 4294967295,
owner = 0xffffffff, dep_map = {key = 0x0, class_cache = 0x0,
name = 0xc1613adc "kernel_flag", cpu = 0, ip = 0}}
(gdb) s
_raw_spin_lock (lock=lock@entry=0xc16922a0 <kernel_flag>)
at lib/spinlock_debug.c:130
130 debug_spin_lock_before(lock);
(gdb) n
131 if (unlikely(!__raw_spin_trylock(&lock->raw_lock)))
(gdb) s
__raw_spin_trylock (lock=0xc16922a0 <kernel_flag>) at lib/spinlock_debug.c:131
131 if (unlikely(!__raw_spin_trylock(&lock->raw_lock)))
(gdb) s
__ticket_spin_trylock (lock=0xc16922a0 <kernel_flag>)
at /home/start-kernel/work_space/github/linux_startup/linux-2.6.32.69/arch/x86/include/asm/spinlock.h:84
84 asm volatile("movzwl %2, %0\n\t"
(gdb) n
_raw_spin_lock (lock=lock@entry=0xc16922a0 <kernel_flag>)
at lib/spinlock_debug.c:131
131 if (unlikely(!__raw_spin_trylock(&lock->raw_lock)))
(gdb)
133 debug_spin_lock_after(lock);
(gdb)
134 }
(gdb)
lock_kernel () at lib/kernel_lock.c:121
121 current->lock_depth = depth;
(gdb) p depth
$4 = 0
(gdb) n
122 }
(gdb)
- Sets zone of memory where allow init to allocate
kernel_init (unused=<optimized out>) at init/main.c:865
865 set_mems_allowed(node_states[N_HIGH_MEMORY]);
(gdb) s
set_mems_allowed (nodemask=...) at include/linux/cpuset.h:91
91 current->mems_allowed = nodemask;
(gdb) p /d N_HIGH_MEMORY
$8 = 3
(gdb) n
kernel_init (unused=<optimized out>) at init/main.c:869
869 set_cpus_allowed_ptr(current, cpu_all_mask);
- Change a given task's CPU affinity. Migrate the thread to a proper CPU and schedule it away if the CPU it's executing on is removed from the allowed bitmask.
kernel_init (unused=<optimized out>) at init/main.c:869
869 set_cpus_allowed_ptr(current, cpu_all_mask);
(gdb) s
set_cpus_allowed_ptr (p=p@entry=0xc7070000, new_mask=0xc1497fa4 <cpu_all_bits>)
at kernel/sched.c:7372
7372 while (task_is_waking(p)) # check state of the task, if it's waking, wait here
(gdb) p p->state
$9 = 0
(gdb) n
7374 rq = task_rq_lock(p, &flags);
(gdb)
7375 if (task_is_waking(p)) {
(gdb)
7380 if (!cpumask_intersects(new_mask, cpu_active_mask)) {
(gdb)
7385 if (unlikely((p->flags & PF_THREAD_BOUND) && p != current &&
(gdb) p p->flags
$10 = 2105408
(gdb) n
7391 if (p->sched_class->set_cpus_allowed)
(gdb) p p->sched_class->set_cpus_allowed
$11 = (void (*)(struct task_struct *, const struct cpumask *)) 0x0
(gdb) n
7394 cpumask_copy(&p->cpus_allowed, new_mask);
(gdb)
7395 p->rt.nr_cpus_allowed = cpumask_weight(new_mask);
(gdb)
7399 if (cpumask_test_cpu(task_cpu(p), new_mask))
# test of task is allowed to executing with new mask
# if yes, complete the function execution
(gdb) p p->rt.nr_cpus_allowed
$12 = 8
(gdb) p *new_mask
$13 = {bits = {255}}
(gdb) n
7365 int ret = 0;
(gdb)
7415 task_rq_unlock(rq, &flags);
(gdb)
7417 return ret;
(gdb)
7418 }
- Prepare for SMP bootup. The MP table or ACPI has been read earlier. Just do some sanity checking here and enable APIC mode.
882 smp_prepare_cpus(setup_max_cpus);
(gdb) s
smp_prepare_cpus (max_cpus=8)
at /home/start-kernel/work_space/github/linux_startup/linux-2.6.32.69/arch/x86/include/asm/smp.h:91
91 smp_ops.smp_prepare_cpus(max_cpus);
(gdb) s
native_smp_prepare_cpus (max_cpus=8) at arch/x86/kernel/smpboot.c:1069
1069 smp_cpu_index_default();
(gdb) s
smp_cpu_index_default () at arch/x86/kernel/smpboot.c:1053
1053 for_each_possible_cpu(i) {
(gdb) n
1054 c = &cpu_data(i);
(gdb)
1056 c->cpu_index = nr_cpu_ids;
(gdb) p nr_cpu_ids
$25 = 1
(gdb) n
172 return native_smp_prepare_cpus (max_cpus=8) at arch/x86/kernel/smpboot.c:1070
1070 current_cpu_data = boot_cpu_data;
(gdb)
1071 cpumask_copy(cpu_callin_mask, cpumask_of(0));
(gdb) p boot_cpu_data
$26 = {x86 = 6 '\006', x86_vendor = 0 '\000', x86_model = 6 '\006',
x86_mask = 3 '\003', wp_works_ok = 1 '\001', hlt_works_ok = 1 '\001',
hard_math = 1 '\001', rfu = 0 '\000', fdiv_bug = 0 '\000',
f00f_bug = 0 '\000', coma_bug = 0 '\000', pad0 = 0 '\000',
x86_virt_bits = 32 ' ', x86_phys_bits = 36 '$', x86_coreid_bits = 0 '\000',
extended_cpuid_level = 2147483652, cpuid_level = 4, x86_capability = {
125873145, 0, 0, 262720, 2155872257, 0, 0, 0, 0},
x86_vendor_id = "GenuineIntel\000\000\000",
x86_model_id = "QEMU Virtual CPU version 2.0.0", '\000' <repeats 33 times>,
x86_cache_size = 4096, x86_cache_alignment = 32, x86_power = 0,
loops_per_jiffy = 13568660, llc_shared_map = {{bits = {0}}},
x86_max_cores = 1, apicid = 0, initial_apicid = 0, x86_clflush_size = 32,
booted_cores = 0, phys_proc_id = 0, cpu_core_id = 0, cpu_index = 0,
x86_hyper_vendor = 0}
(gdb) p cpu_callin_mask
$27 = {{bits = {0}}}
(gdb) n
1072 mb();
(gdb) p cpu_callin_mask
$28 = {{bits = {1}}}
(gdb)
1076 smp_store_cpu_info(0); /* Final full version of the data */
(gdb) s
smp_store_cpu_info (id=id@entry=0) at arch/x86/kernel/smpboot.c:388
388 struct cpuinfo_x86 *c = &cpu_data(id);
(gdb) n
390 copy_cpuinfo_x86(c, &boot_cpu_data);
(gdb) s
copy_cpuinfo_x86 (src=<optimized out>, dst=0xc234a200)
at arch/x86/kernel/smpboot.c:377
377 *dst = *src;
(gdb) n
smp_store_cpu_info (id=id@entry=0) at arch/x86/kernel/smpboot.c:391
391 c->cpu_index = id;
(gdb)
392 if (id != 0)
(gdb)
394 }
(gdb)
native_smp_prepare_cpus (max_cpus=8) at arch/x86/kernel/smpboot.c:1078
1078 boot_cpu_logical_apicid = logical_smp_processor_id();
(gdb) s
logical_smp_processor_id () at arch/x86/kernel/smpboot.c:1078
1078 boot_cpu_logical_apicid = logical_smp_processor_id();
(gdb) s
apic_read (reg=208)
at /home/start-kernel/work_space/github/linux_startup/linux-2.6.32.69/arch/x86/include/asm/apic.h:378
378 return apic->read(reg);
(gdb) s
native_apic_mem_read (reg=208)
at /home/start-kernel/work_space/github/linux_startup/linux-2.6.32.69/arch/x86/include/asm/apic.h:112
112 return *((volatile u32 *)(APIC_BASE + reg));
(gdb) n
113 }
(gdb)
logical_smp_processor_id ()
at /home/start-kernel/work_space/github/linux_startup/linux-2.6.32.69/arch/x86/include/asm/smp.h:196
196 return GET_APIC_LOGICAL_ID(apic_read(APIC_LDR));
(gdb) n
native_smp_prepare_cpus (max_cpus=8) at arch/x86/kernel/smpboot.c:1080
1080 current_thread_info()->cpu = 0; /* needed? */
(gdb)
1081 for_each_possible_cpu(i) { # allocate memory for per cpu variables
(gdb)
1082 zalloc_cpumask_var(&per_cpu(cpu_sibling_map, i), GFP_KERNEL);
(gdb)
1083 zalloc_cpumask_var(&per_cpu(cpu_core_map, i), GFP_KERNEL);
(gdb)
1084 zalloc_cpumask_var(&cpu_data(i).llc_shared_map, GFP_KERNEL);
(gdb)
1086 set_cpu_sibling_map(0);
(gdb)
1088 enable_IR_x2apic();
(gdb)
1089 default_setup_apic_routing();
(gdb)
1091 if (smp_sanity_check(max_cpus) < 0) {
(gdb) p max_cpus
$29 = 8
(gdb) n
1098 if (read_apic_id() != boot_cpu_physical_apicid) {
(gdb) p boot_cpu_physical_apicid
$30 = 0
(gdb) n
1105 connect_bsp_APIC();
(gdb)
1110 setup_local_APIC();
(gdb)
1115 if (!skip_ioapic_setup && nr_ioapics)
(gdb)
1116 enable_IO_APIC();
(gdb)
1118 end_local_APIC_setup();
(gdb)
1120 map_cpu_to_logical_apicid();
(gdb) s
map_cpu_to_logical_apicid () at arch/x86/kernel/smpboot.c:172
172 int cpu = smp_processor_id();
(gdb) n
173 int apicid = logical_smp_processor_id();
(gdb)
174 int node = apic->apicid_to_node(apicid);
(gdb)
179 cpu_2_logical_apicid[cpu] = apicid;
(gdb)
181 }
(gdb)
native_smp_prepare_cpus (max_cpus=<optimized out>)
at arch/x86/kernel/smpboot.c:1122
1122 if (apic->setup_portio_remap)
(gdb)
1125 smpboot_setup_io_apic();
(gdb) s
smpboot_setup_io_apic ()
at /home/start-kernel/work_space/github/linux_startup/linux-2.6.32.69/arch/x86/include/asm/smpboot_hooks.h:47
47 if (!skip_ioapic_setup && nr_ioapics)
(gdb) n
48 setup_IO_APIC();
(gdb)
native_smp_prepare_cpus (max_cpus=<optimized out>)
at arch/x86/kernel/smpboot.c:1130
1130 printk(KERN_INFO "CPU%d: ", 0);
(gdb)
1131 print_cpu_info(&cpu_data(0));
(gdb)
1132 x86_init.timers.setup_percpu_clockev();
(gdb) s
setup_boot_APIC_clock () at arch/x86/kernel/apic/apic.c:749
749 if (disable_apic_timer) {
(gdb) n
759 apic_printk(APIC_VERBOSE, "Using local APIC timer interrupts.\n"
(gdb)
762 if (calibrate_APIC_clock()) {
(gdb)
774 if (nmi_watchdog != NMI_IO_APIC)
(gdb)
775 lapic_clockevent.features &= ~CLOCK_EVT_FEAT_DUMMY;
(gdb)
765 setup_APIC_timer();
(gdb)
782 }
(gdb)
native_smp_prepare_cpus (max_cpus=<optimized out>)
at arch/x86/kernel/smpboot.c:1137
1137 set_mtrr_aps_delayed_init();
(gdb) s
set_mtrr_aps_delayed_init () at arch/x86/kernel/cpu/mtrr/main.c:768
768 if (!use_intel())
(gdb) n
772 }
(gdb)
native_smp_prepare_cpus (max_cpus=<optimized out>)
at arch/x86/kernel/smpboot.c:1140
1140 }
(gdb)
- Invoke SMP init routines
migration_init: create migration thread for boot cpu and wake up the created thread, registers notifier to response new added CPU.
spawn_ksoftirqd: create ksoftirqd thread and wake up the created thread, register notifier to response added CPU.
init_call_single_data: initialize per CPU variable call_single_queue and allocate CPU mask for up CPU, register notifier to response new added CPU.
spawn_softlockup_task: create watch dog thread for CPU up and down, register notifier to response new added CPU
relay_init: register CPU hotplug callback
tracer_alloc_buffers: prepare trace for all possible CPU and register response routine
init_trace_printk: register module notifier for treace printk
Breakpoint 3, kernel_init (unused=<optimized out>) at init/main.c:884
884 do_pre_smp_initcalls();
(gdb) s
do_pre_smp_initcalls () at init/main.c:797
797 for (call = __initcall_start; call < __early_initcall_end; call++)
(gdb) p __initcall_start
$1 = 0xc177d998
(gdb) p *(initcall_t*)__initcall_start
$3 = (initcall_t) 0xc170d5e3 <migration_init>
(gdb) p __early_initcall_end
$4 = 0xc177d9b4 <__initcall_init_mmap_min_addr0>
(gdb) p __initcall_start+1
$6 = (initcall_t *) 0xc177d99c <__initcall_spawn_ksoftirqdearly>
(gdb) p __initcall_start+2
$7 = (initcall_t *) 0xc177d9a0 <__initcall_init_call_single_dataearly>
(gdb) p __initcall_start+3
$8 = (initcall_t *) 0xc177d9a4 <__initcall_spawn_softlockup_taskearly>
(gdb) p __initcall_start+4
$9 = (initcall_t *) 0xc177d9a8 <__initcall_relay_initearly>
(gdb) p __initcall_start+5
$10 = (initcall_t *) 0xc177d9ac <__initcall_tracer_alloc_buffersearly>
(gdb) p __initcall_start+6
$11 = (initcall_t *) 0xc177d9b0 <__initcall_init_trace_printkearly>
(gdb)
- rest of initialization of SMP in boot CPU
887 smp_init();
(gdb) s
smp_init () at init/main.c:373
373 for_each_present_cpu(cpu) {
(gdb)
374 if (num_online_cpus() >= setup_max_cpus)
(gdb)
376 if (!cpu_online(cpu))
(gdb) p setup_max_cpus
$12 = 8
(gdb) n
381 printk(KERN_INFO "Brought up %ld CPUs\n", (long)num_online_cpus());
(gdb)
382 smp_cpus_done(setup_max_cpus);
(gdb) s
smp_cpus_done (max_cpus=8)
at /home/start-kernel/work_space/github/linux_startup/linux-2.6.32.69/arch/x86/include/asm/smp.h:96
96 smp_ops.smp_cpus_done(max_cpus);
(gdb) s
native_smp_cpus_done (max_cpus=8) at arch/x86/kernel/smpboot.c:1166
1166 pr_debug("Boot done.\n");
(gdb) n
1168 impress_friends();
(gdb) s
impress_friends () at arch/x86/kernel/smpboot.c:483
483 pr_debug("Before bogomips.\n");
(gdb) n
484 for_each_possible_cpu(cpu)
(gdb)
485 if (cpumask_test_cpu(cpu, cpu_callout_mask))
(gdb)
486 bogosum += cpu_data(cpu).loops_per_jiffy;
(gdb)
484 for_each_possible_cpu(cpu)
(gdb)
487 printk(KERN_INFO
(gdb) p bogosum
$14 = 13568592
(gdb) n
491 (bogosum/(5000/HZ))%100);
(gdb)
493 pr_debug("Before bogocount - setting activated=1.\n");
(gdb)
native_smp_cpus_done (max_cpus=<optimized out>)
at arch/x86/kernel/smpboot.c:1170
1170 setup_ioapic_dest();
(gdb) s
setup_ioapic_dest () at arch/x86/kernel/apic/io_apic.c:4096
4096 if (skip_ioapic_setup == 1)
(gdb) p skip_ioapic_setup
$15 = 0
(gdb) n
4099 for (ioapic = 0; ioapic < nr_ioapics; ioapic++)
(gdb) p nr_ioapics
$16 = 1
(gdb) n
4100 for (pin = 0; pin < nr_ioapic_registers[ioapic]; pin++) {
(gdb) p nr_ioapic_registers[0]
$17 = 24
(gdb) n
4101 irq_entry = find_irq_entry(ioapic, pin, mp_INT);
(gdb)
4102 if (irq_entry == -1)
(gdb)
4104 irq = pin_2_irq(irq_entry, ioapic, pin);
(gdb)
4106 if ((ioapic > 0) && (irq > 16))
(gdb) p irq
$19 = 1
(gdb) n
4109 desc = irq_to_desc(irq);
(gdb)
4114 if (desc->status &
(gdb) p desc
$21 = (struct irq_desc *) 0xc1690c00 <irq_desc_legacy+160>
(gdb) p *desc
$22 = {irq = 1, kstat_irqs = 0xc7001004,
handle_irq = 0xc10ab4a0 <handle_edge_irq>, chip = 0xc16f7140 <ioapic_chip>,
msi_desc = 0x0, handler_data = 0x0, chip_data = 0xc1697194 <irq_cfgx+20>,
action = 0x0, status = 512, depth = 1, wake_depth = 0, irq_count = 0,
last_unhandled = 0, irqs_unhandled = 0, lock = {raw_lock = {slock = 1028},
magic = 3735899821, owner_cpu = 4294967295, owner = 0xffffffff, dep_map = {
key = 0xc1e3376c <irq_desc_lock_class>, class_cache = 0x0,
name = 0xc15dae6f "&irq_desc_lock_class", cpu = 0, ip = 3238703088}},
affinity = {{bits = {255}}}, node = 0, pending_mask = {{bits = {0}}},
threads_active = {counter = 0}, wait_for_threads = {lock = {raw_lock = {
slock = 0}, magic = 0, owner_cpu = 0, owner = 0x0, dep_map = {
key = 0x0, class_cache = 0x0, name = 0x0, cpu = 0, ip = 0}},
task_list = {next = 0x0, prev = 0x0}}, dir = 0x0, name = 0xc15d11ec "edge"}
(gdb) n
4118 mask = apic->target_cpus();
(gdb) s
default_target_cpus ()
at /home/start-kernel/work_space/github/linux_startup/linux-2.6.32.69/arch/x86/include/asm/apic.h:475
475 }
(gdb) n
setup_ioapic_dest () at arch/x86/kernel/apic/io_apic.c:4123
4123 set_ioapic_affinity_irq_desc(desc, mask);
(gdb) p mask
$23 = (const struct cpumask *) 0xc16f7260 <cpu_online_bits>
(gdb) p *mask
$24 = {bits = {1}}
(gdb) s
set_ioapic_affinity_irq_desc (
desc=desc@entry=0xc1690c00 <irq_desc_legacy+160>,
mask=0xc16f7260 <cpu_online_bits>) at arch/x86/kernel/apic/io_apic.c:2366
2366 cfg = desc->chip_data;
(gdb) n
2368 spin_lock_irqsave(&ioapic_lock, flags);
(gdb)
2369 dest = set_desc_affinity(desc, mask);
(gdb)
2370 if (dest != BAD_APICID) {
(gdb)
2372 dest = SET_APIC_LOGICAL_ID(dest);
(gdb)
2373 __target_IO_APIC_irq(irq, dest, cfg);
(gdb)
2374 ret = 0;
(gdb)
2376 spin_unlock_irqrestore(&ioapic_lock, flags);
(gdb)
2379 }
(gdb) break check_nmi_watchdog
Breakpoint 9 at 0xc1707f40: file arch/x86/kernel/apic/nmi.c, line 134.
(gdb) c
Continuing.
Error in testing breakpoint condition:
value has been optimized out
Breakpoint 9, check_nmi_watchdog () at arch/x86/kernel/apic/nmi.c:134
134 if (!nmi_watchdog_active() || !atomic_read(&nmi_active))
(gdb) n
native_smp_cpus_done (max_cpus=<optimized out>)
at arch/x86/kernel/smpboot.c:1173
1173 mtrr_aps_init();
(gdb) s
mtrr_aps_init () at arch/x86/kernel/cpu/mtrr/main.c:779
779 if (!use_intel())
(gdb) n
792 }
(gdb)
native_smp_cpus_done (max_cpus=<optimized out>)
at arch/x86/kernel/smpboot.c:1174
1174 }
(gdb)
- Initialize schedule variables and doamins, register callback function to response CPU come and go
Breakpoint 3, sched_init_smp () at kernel/sched.c:9509
9509 get_online_cpus(); # increase reference
(gdb) s
get_online_cpus () at kernel/cpu.c:48
48 might_sleep();
(gdb) n
49 if (cpu_hotplug.active_writer == current)
(gdb) p cpu_hotplug
$3 = {active_writer = 0x0, lock = {count = {counter = 1}, wait_lock = {
raw_lock = {slock = 2056}, magic = 3735899821, owner_cpu = 4294967295,
owner = 0xffffffff, dep_map = {key = 0x0, class_cache = 0x0,
name = 0xc15d41ad "cpu_hotplug.lock.wait_lock", cpu = 0, ip = 0}},
wait_list = {next = 0xc1698eec <cpu_hotplug+44>,
prev = 0xc1698eec <cpu_hotplug+44>}, owner = 0x0, name = 0x0,
magic = 0xc1698ec4 <cpu_hotplug+4>, dep_map = {
key = 0xc1698f00 <cpu_hotplug+64>,
class_cache = 0xc1b605e0 <lock_classes+155040>,
name = 0xc15d41c8 "cpu_hotplug.lock", cpu = 0, ip = 3238298482}},
refcount = 0}
(gdb) n
51 mutex_lock(&cpu_hotplug.lock);
(gdb)
52 cpu_hotplug.refcount++;
(gdb)
53 mutex_unlock(&cpu_hotplug.lock);
(gdb)
55 }
(gdb)
sched_init_smp () at kernel/sched.c:9510
9510 mutex_lock(&sched_domains_mutex);
(gdb)
9511 arch_init_sched_domains(cpu_active_mask);
(gdb) p *cpu_active_mask
$5 = {bits = {1}}
(gdb) s
arch_init_sched_domains (cpu_map=<optimized out>) at kernel/sched.c:9220
9220 arch_update_cpu_topology();
(gdb) s
arch_update_cpu_topology () at kernel/sched.c:9209
9209 }
(gdb) n
arch_init_sched_domains (cpu_map=<optimized out>) at kernel/sched.c:9221
9221 ndoms_cur = 1;
(gdb)
9222 doms_cur = kmalloc(cpumask_size(), GFP_KERNEL);
(gdb)
9225 cpumask_andnot(doms_cur, cpu_map, cpu_isolated_map);
(gdb) s
cpumask_andnot (src2p=0xc18eb14c <cpu_isolated_map>, src1p=<optimized out>,
dstp=<optimized out>) at kernel/sched.c:9225
9225 cpumask_andnot(doms_cur, cpu_map, cpu_isolated_map);
(gdb) s
bitmap_andnot (src2=0xc18eb14c <cpu_isolated_map>, nbits=8,
src1=<optimized out>, dst=<optimized out>) at include/linux/bitmap.h:204
204 return (*dst = *src1 & ~(*src2)) != 0;
(gdb) n
arch_init_sched_domains (cpu_map=<optimized out>) at kernel/sched.c:9226
9226 dattr_cur = NULL;
(gdb)
9227 err = build_sched_domains(doms_cur);
(gdb) s
build_sched_domains (cpu_map=0xc7020188) at kernel/sched.c:9186
9186 return __build_sched_domains(cpu_map, NULL);
(gdb) n
arch_init_sched_domains (cpu_map=<optimized out>) at kernel/sched.c:9228
9228 register_sched_domain_sysctl();
(gdb) s
register_sched_domain_sysctl () at kernel/sched.c:7825
7825 struct ctl_table *entry = sd_alloc_ctl_entry(cpu_num + 1);
(gdb) n
7828 WARN_ON(sd_ctl_dir[0].child);
(gdb)
7829 sd_ctl_dir[0].child = entry;
(gdb)
7831 if (entry == NULL)
(gdb)
7834 for_each_possible_cpu(i) {
(gdb)
7835 snprintf(buf, 32, "cpu%d", i);
(gdb)
7836 entry->procname = kstrdup(buf, GFP_KERNEL);
(gdb)
7837 entry->mode = 0555;
(gdb)
7838 entry->child = sd_alloc_ctl_cpu_table(i);
(gdb)
7839 entry++;
(gdb)
7834 for_each_possible_cpu(i) {
(gdb)
7842 WARN_ON(sd_sysctl_header);
(gdb)
7843 sd_sysctl_header = register_sysctl_table(sd_ctl_root);
(gdb)
7844 }
(gdb)
sched_init_smp () at kernel/sched.c:9512
9512 cpumask_andnot(non_isolated_cpus, cpu_possible_mask, cpu_isolated_map);
(gdb)
9513 if (cpumask_empty(non_isolated_cpus))
(gdb) p non_isolated_cpus
$14 = {{bits = {3245306464}}}
(gdb) n
9515 mutex_unlock(&sched_domains_mutex);
(gdb)
9516 put_online_cpus(); # decrease reference
(gdb) s
put_online_cpus () at kernel/cpu.c:60
60 if (cpu_hotplug.active_writer == current)
(gdb) n
62 mutex_lock(&cpu_hotplug.lock);
(gdb)
63 if (!--cpu_hotplug.refcount && unlikely(cpu_hotplug.active_writer))
(gdb)
65 mutex_unlock(&cpu_hotplug.lock);
(gdb)
67 }
(gdb)
sched_init_smp () at kernel/sched.c:9524
9524 hotcpu_notifier(update_runtime, 0); # RT runtime code needs to handle some hotplug events
(gdb) s
register_cpu_notifier (nb=nb@entry=0xc16de1f0 <update_runtime_nb>)
at kernel/cpu.c:131
131 cpu_maps_update_begin();
(gdb) n
132 ret = raw_notifier_chain_register(&cpu_chain, nb);
(gdb)
133 cpu_maps_update_done();
(gdb)
135 }
(gdb)
sched_init_smp () at kernel/sched.c:9526
9526 init_hrtick(); # high resolutuib time to handle hotplug events
(gdb) s
init_hrtick () at kernel/sched.c:1102
1102 hotcpu_notifier(hotplug_hrtick, 0);
(gdb) n
sched_init_smp () at kernel/sched.c:9529
9529 if (set_cpus_allowed_ptr(current, non_isolated_cpus) < 0)
(gdb)
sched_init_smp () at kernel/sched.c:9531
9531 sched_init_granularity(); # increase the granularity value when there are more CPUs
(gdb)
9534 init_sched_rt_class(); # allocate cpu mask
(gdb)
9535 }
(gdb)
kernel_init (unused=<optimized out>) at init/main.c:890
890 do_basic_setup();
(gdb)
- Initialize some basic initialization
init_workqueues: initialize work queue for all up CPUs
cpuset_init_smp: intialize cpu and memory allowed, register notifier to response CPU hotplug, create work queue and work queue thread for cpuset
usermodehelper_init: create work queue and work queue thread for khelper
kernel_init (unused=<optimized out>) at init/main.c:890
890 do_basic_setup();
(gdb) s
do_basic_setup () at init/main.c:783
783 init_workqueues();
(gdb) s
init_workqueues () at kernel/workqueue.c:1049
1049 cpumask_copy(cpu_populated_map, cpu_online_mask);
(gdb) p *cpu_online_mask
$5 = {bits = {1}}
(gdb) n
1050 singlethread_cpu = cpumask_first(cpu_possible_mask);
(gdb) p *cpu_possible_mask
$6 = {bits = {1}}
(gdb) n
1051 cpu_singlethread_map = cpumask_of(singlethread_cpu);
(gdb)
1052 hotcpu_notifier(workqueue_cpu_callback, 0);
(gdb)
1053 keventd_wq = create_workqueue("events");
(gdb) p *cpu_singlethread_map
$10 = {bits = {1}}
(gdb) s
__create_workqueue_key (name=name@entry=0xc1619d0a "events",
singlethread=singlethread@entry=0, freezeable=freezeable@entry=0,
rt=rt@entry=0, key=key@entry=0xc190e014 <__key.22871>,
lock_name=lock_name@entry=0xc1619d0a "events") at kernel/workqueue.c:839
839 wq = kzalloc(sizeof(*wq), GFP_KERNEL);
(gdb) n
840 if (!wq)
(gdb)
843 wq->cpu_wq = alloc_percpu(struct cpu_workqueue_struct);
(gdb)
844 if (!wq->cpu_wq) {
(gdb)
849 wq->name = name;
(gdb)
850 lockdep_init_map(&wq->lockdep_map, lock_name, key, 0);
(gdb) s
lockdep_init_map (lock=lock@entry=0xc7001e1c,
name=name@entry=0xc1619d0a "events",
key=key@entry=0xc190e014 <__key.22871>, subclass=subclass@entry=0)
at kernel/lockdep.c:2683
2683 if (DEBUG_LOCKS_WARN_ON(!name)) {
(gdb) n
2678 lock->class_cache = NULL;
(gdb)
2680 lock->cpu = raw_smp_processor_id();
(gdb)
2683 if (DEBUG_LOCKS_WARN_ON(!name)) {
(gdb)
2688 lock->name = name;
(gdb)
2690 if (DEBUG_LOCKS_WARN_ON(!key))
(gdb)
2695 if (!static_obj(key)) {
(gdb)
2700 lock->key = key;
(gdb)
2702 if (unlikely(!debug_locks))
(gdb)
2705 if (subclass)
(gdb)
2707 }
(gdb)
__create_workqueue_key (name=name@entry=0xc1619d0a "events",
singlethread=singlethread@entry=0, freezeable=freezeable@entry=0,
rt=rt@entry=0, key=key@entry=0xc190e014 <__key.22871>,
lock_name=lock_name@entry=0xc1619d0a "events") at kernel/workqueue.c:851
851 wq->singlethread = singlethread;
(gdb)
852 wq->freezeable = freezeable;
(gdb)
853 wq->rt = rt;
(gdb)
854 INIT_LIST_HEAD(&wq->list);
(gdb)
856 if (singlethread) {
(gdb)
861 cpu_maps_update_begin();
(gdb) s
cpu_maps_update_begin () at kernel/cpu.c:78
78 mutex_lock(&cpu_add_remove_lock);
(gdb) n
79 }
(gdb)
__create_workqueue_key (name=name@entry=0xc1619d0a "events",
singlethread=singlethread@entry=0, freezeable=freezeable@entry=0,
rt=rt@entry=0, key=key@entry=0xc190e014 <__key.22871>,
lock_name=lock_name@entry=0xc1619d0a "events") at kernel/workqueue.c:868
868 spin_lock(&workqueue_lock);
(gdb)
869 list_add(&wq->list, &workqueues);
(gdb)
870 spin_unlock(&workqueue_lock);
(gdb)
877 for_each_possible_cpu(cpu) {
(gdb)
878 cwq = init_cpu_workqueue(wq, cpu);
(gdb)
879 if (err || !cpu_online(cpu))
(gdb)
881 err = create_workqueue_thread(cwq, cpu);
(gdb) s
create_workqueue_thread (cwq=cwq@entry=0xc234a8c0, cpu=cpu@entry=0)
at kernel/workqueue.c:792
792 struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 };
(gdb)
793 struct workqueue_struct *wq = cwq->wq;
(gdb) n
794 const char *fmt = is_wq_single_threaded(wq) ? "%s" : "%s/%d";
(gdb)
797 p = kthread_create(worker_thread, cwq, fmt, wq->name, cpu);
(gdb)
806 if (IS_ERR(p))
(gdb)
808 if (cwq->wq->rt)
(gdb)
812 trace_workqueue_creation(cwq->thread, cpu);
(gdb)
810 cwq->thread = p;
(gdb)
812 trace_workqueue_creation(cwq->thread, cpu);
(gdb)
814 return 0;
(gdb)
815 }
(gdb)
__create_workqueue_key (name=name@entry=0xc1619d0a "events",
singlethread=singlethread@entry=0, freezeable=freezeable@entry=0,
rt=rt@entry=0, key=key@entry=0xc190e014 <__key.22871>,
lock_name=lock_name@entry=0xc1619d0a "events") at kernel/workqueue.c:882
882 start_workqueue_thread(cwq, cpu);
(gdb) s
start_workqueue_thread (cpu=cpu@entry=0, cwq=0xc234a8c0)
at kernel/workqueue.c:821
821 if (p != NULL) {
(gdb) n
822 if (cpu >= 0)
(gdb)
823 kthread_bind(p, cpu);
(gdb)
824 wake_up_process(p);
(gdb)
826 }
(gdb)
__create_workqueue_key (name=name@entry=0xc1619d0a "events",
singlethread=singlethread@entry=0, freezeable=freezeable@entry=0,
rt=rt@entry=0, key=key@entry=0xc190e014 <__key.22871>,
lock_name=lock_name@entry=0xc1619d0a "events") at kernel/workqueue.c:877
877 for_each_possible_cpu(cpu) {
(gdb) n
884 cpu_maps_update_done();
(gdb)
887 if (err) {
(gdb)
892 }
(gdb)
init_workqueues () at kernel/workqueue.c:1054
1054 BUG_ON(!keventd_wq);
(gdb)
1055 }
(gdb)
do_basic_setup () at init/main.c:784
784 cpuset_init_smp();
(gdb) s
cpuset_init_smp () at kernel/cpuset.c:2127
2127 cpumask_copy(top_cpuset.cpus_allowed, cpu_active_mask);
(gdb) n
2128 top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
(gdb)
2130 hotcpu_notifier(cpuset_track_online_cpus, 0);
(gdb)
2133 cpuset_wq = create_singlethread_workqueue("cpuset");
(gdb)
2134 BUG_ON(!cpuset_wq);
2135 }
(gdb)
do_basic_setup () at init/main.c:785
785 usermodehelper_init();
(gdb) s
usermodehelper_init () at kernel/kmod.c:640
640 khelper_wq = create_singlethread_workqueue("khelper");
(gdb) n
641 BUG_ON(!khelper_wq);
(gdb)
642 }
init_tmpfs: initialize tmp file system
super block -> inode -> dentry
786 init_tmpfs();
(gdb) s
init_tmpfs () at mm/shmem.c:2533
2533 error = bdi_init(&shmem_backing_dev_info); # initialize block device information
(gdb) n
2534 if (error)
(gdb)
2537 error = init_inodecache(); # create shmem_inode_cache
(gdb) s
init_inodecache () at mm/shmem.c:2410
2410 shmem_inode_cachep = kmem_cache_create("shmem_inode_cache",
(gdb) n
init_tmpfs () at mm/shmem.c:2541
2541 error = register_filesystem(&tmpfs_fs_type); # register tmpfs_fs_type
(gdb)
2542 if (error) {
(gdb)
2547 shm_mnt = vfs_kern_mount(&tmpfs_fs_type, MS_NOUSER,
(gdb) s
vfs_kern_mount (type=type@entry=0xc16a4560 <tmpfs_fs_type>,
flags=flags@entry=-2147483648, name=0xc1620c4e "tmpfs",
data=data@entry=0x0) at fs/super.c:920
920 if (!type)
(gdb) n
924 mnt = alloc_vfsmnt(name);
(gdb) s
alloc_vfsmnt (name=name@entry=0xc1620c4e "tmpfs") at fs/namespace.c:130
130 struct vfsmount *mnt = kmem_cache_zalloc(mnt_cache, GFP_KERNEL);
(gdb) n
131 if (mnt) {
(gdb)
134 err = mnt_alloc_id(mnt);
(gdb) s
mnt_alloc_id (mnt=<optimized out>) at fs/namespace.c:73
73 ida_pre_get(&mnt_id_ida, GFP_KERNEL);
(gdb) s
ida_pre_get (ida=ida@entry=0xc16a98c0 <mnt_id_ida>,
gfp_mask=gfp_mask@entry=208) at lib/idr.c:737
737 if (!idr_pre_get(&ida->idr, gfp_mask))
(gdb) s
idr_pre_get (idp=idp@entry=0xc16a98c0 <mnt_id_ida>,
gfp_mask=gfp_mask@entry=208) at lib/idr.c:122
122 while (idp->id_free_cnt < IDR_FREE_MAX) {
(gdb)
124 new = kmem_cache_zalloc(idr_layer_cache, gfp_mask);
(gdb)
125 if (new == NULL)
(gdb)
127 move_to_free_list(idp, new);
(gdb) s
move_to_free_list (p=0xc7024300, idp=0xc16a98c0 <mnt_id_ida>) at lib/idr.c:83
83 spin_lock_irqsave(&idp->lock, flags);
(gdb) n
84 __move_to_free_list(idp, p);
(gdb) s
__move_to_free_list (idp=<optimized out>, idp=<optimized out>, p=0xc7024300)
at lib/idr.c:71
71 p->ary[0] = idp->id_free;
(gdb) n
72 idp->id_free = p;
(gdb)
73 idp->id_free_cnt++;
(gdb)
move_to_free_list (p=0xc7024300, idp=0xc16a98c0 <mnt_id_ida>) at lib/idr.c:85
85 spin_unlock_irqrestore(&idp->lock, flags);
(gdb)
idr_pre_get (idp=idp@entry=0xc16a98c0 <mnt_id_ida>,
gfp_mask=gfp_mask@entry=208) at lib/idr.c:122
122 while (idp->id_free_cnt < IDR_FREE_MAX) {
(gdb) p idp->id_free_cnt
$1 = 14
(gdb) n
129 return 1;
(gdb)
130 }
(gdb)
ida_pre_get (ida=ida@entry=0xc16a98c0 <mnt_id_ida>,
gfp_mask=gfp_mask@entry=208) at lib/idr.c:741
741 if (!ida->free_bitmap) {
(gdb)
751 return 1;
(gdb)
752 }
(gdb)
mnt_alloc_id (mnt=<optimized out>) at fs/namespace.c:74
74 spin_lock(&vfsmount_lock);
(gdb)
75 res = ida_get_new_above(&mnt_id_ida, mnt_id_start, &mnt->mnt_id);
(gdb) s
ida_get_new_above (ida=ida@entry=0xc16a98c0 <mnt_id_ida>, starting_id=4,
p_id=p_id@entry=0xc701d360) at lib/idr.c:775
775 int idr_id = starting_id / IDA_BITMAP_BITS;
(gdb) n
776 int offset = starting_id % IDA_BITMAP_BITS;
(gdb)
781 t = idr_get_empty_slot(&ida->idr, idr_id, pa);
(gdb) s
idr_get_empty_slot (idp=idp@entry=0xc16a98c0 <mnt_id_ida>,
starting_id=starting_id@entry=0, pa=pa@entry=0xc706bf38) at lib/idr.c:204
204 {
(gdb) n
211 p = idp->top;
(gdb)
213 if (unlikely(!p)) {
(gdb)
223 while ((layers < (MAX_LEVEL - 1)) && (id >= (1 << (layers*IDR_BITS)))) {
(gdb) p *idp
$3 = {top = 0xc701e9c0, id_free = 0xc7024300, layers = 1, id_free_cnt = 14,
lock = {raw_lock = {slock = 6939}, magic = 3735899821,
owner_cpu = 4294967295, owner = 0xffffffff, dep_map = {
key = 0xc16a98e0 <mnt_id_ida+32>,
class_cache = 0xc1b5d070 <lock_classes+141360>,
name = 0xc15ef581 "mnt_id_ida.lock", cpu = 0, ip = 3240328551}}}
(gdb) n
255 rcu_assign_pointer(idp->top, p);
(gdb)
256 idp->layers = layers;
(gdb)
257 v = sub_alloc(idp, &id, pa);
(gdb) s
sub_alloc (pa=0xc706bf38, starting_id=<synthetic pointer>,
idp=0xc16a98c0 <mnt_id_ida>) at lib/idr.c:144
144 pa[l--] = NULL;
(gdb) n
149 n = (id >> (IDR_BITS*l)) & IDR_MASK;
(gdb) p p
$4 = (struct idr_layer *) 0xc701e9c0
(gdb) p l
$5 = 0
(gdb) n
149 n = (id >> (IDR_BITS*l)) & IDR_MASK;
(gdb)
150 bm = ~p->bitmap;
(gdb)
151 m = find_next_bit(&bm, IDR_SIZE, n);
(gdb) p /t p->bitmap
$6 = 0
(gdb) n
152 if (m == IDR_SIZE) {
(gdb) p m
$7 = 0
(gdb) n
175 if (m != n) {
(gdb)
179 if ((id >= MAX_ID_BIT) || (id < 0))
(gdb)
181 if (l == 0)
(gdb)
idr_get_empty_slot (idp=idp@entry=0xc16a98c0 <mnt_id_ida>,
starting_id=starting_id@entry=0, pa=pa@entry=0xc706bf38) at lib/idr.c:257
257 v = sub_alloc(idp, &id, pa);
(gdb)
261 }
(gdb) p pa[0]
$8 = (struct idr_layer *) 0xc701e9c0
(gdb) n
ida_get_new_above (ida=ida@entry=0xc16a98c0 <mnt_id_ida>,
starting_id=<optimized out>, p_id=p_id@entry=0xc701d360) at lib/idr.c:782
782 if (t < 0)
(gdb)
785 if (t * IDA_BITMAP_BITS >= MAX_ID_BIT)
(gdb)
789 offset = 0;
(gdb)
793 bitmap = (void *)pa[0]->ary[idr_id & IDR_MASK];
(gdb)
794 if (!bitmap) {
(gdb)
810 t = find_next_zero_bit(bitmap->bitmap, IDA_BITMAP_BITS, offset);
(gdb) p *bitmap
$10 = {nr_busy = 4, bitmap = {15, 0 <repeats 30 times>}}
(gdb) n
811 if (t == IDA_BITMAP_BITS) {
(gdb) p t
$11 = 4
(gdb) n
819 if (id >= MAX_ID_BIT)
(gdb)
822 __set_bit(t, bitmap->bitmap);
(gdb) p bitmap->bitmap
$12 = {15, 0 <repeats 30 times>}
(gdb) n
823 if (++bitmap->nr_busy == IDA_BITMAP_BITS)
(gdb) p bitmap->bitmap
$13 = {31, 0 <repeats 30 times>}
(gdb) n
826 *p_id = id;
(gdb)
833 if (ida->idr.id_free_cnt || ida->free_bitmap) {
(gdb)
834 struct idr_layer *p = get_from_free_list(&ida->idr);
(gdb)
835 if (p)
(gdb)
836 kmem_cache_free(idr_layer_cache, p);
(gdb)
840 }
(gdb) p p_id
$14 = (int *) 0xc701d360
(gdb) p *p_id
$15 = 4
(gdb) n
839 return 0;
(gdb)
840 }
(gdb)
mnt_alloc_id (mnt=<optimized out>) at fs/namespace.c:76
76 if (!res)
(gdb)
77 mnt_id_start = mnt->mnt_id + 1;
(gdb)
78 spin_unlock(&vfsmount_lock);
(gdb)
alloc_vfsmnt (name=name@entry=0xc1620c4e "tmpfs") at fs/namespace.c:138
138 if (name) {
(gdb)
139 mnt->mnt_devname = kstrdup(name, GFP_KERNEL);
(gdb)
140 if (!mnt->mnt_devname)
(gdb)
144 atomic_set(&mnt->mnt_count, 1);
(gdb)
145 INIT_LIST_HEAD(&mnt->mnt_hash);
(gdb)
146 INIT_LIST_HEAD(&mnt->mnt_child);
(gdb)
147 INIT_LIST_HEAD(&mnt->mnt_mounts);
(gdb)
148 INIT_LIST_HEAD(&mnt->mnt_list);
(gdb)
149 INIT_LIST_HEAD(&mnt->mnt_expire);
(gdb)
150 INIT_LIST_HEAD(&mnt->mnt_share);
(gdb)
151 INIT_LIST_HEAD(&mnt->mnt_slave_list);
(gdb)
152 INIT_LIST_HEAD(&mnt->mnt_slave);
(gdb)
154 mnt->mnt_writers = alloc_percpu(int);
(gdb)
155 if (!mnt->mnt_writers)
(gdb)
172 }
(gdb)
vfs_kern_mount (type=type@entry=0xc16a4560 <tmpfs_fs_type>,
flags=flags@entry=-2147483648, name=0xc1620c4e "tmpfs",
data=data@entry=0x0) at fs/super.c:925
925 if (!mnt)
(gdb)
928 if (data && !(type->fs_flags & FS_BINARY_MOUNTDATA)) {
(gdb)
938 error = type->get_sb(type, flags, name, data, mnt);
(gdb) s
shmem_get_sb (fs_type=0xc16a4560 <tmpfs_fs_type>, flags=-2147483648,
dev_name=0xc1620c4e "tmpfs", data=0x0, mnt=0xc701d300) at mm/shmem.c:2519
2519 return get_sb_nodev(fs_type, flags, data, shmem_fill_super, mnt);
(gdb) s
get_sb_nodev (fs_type=0xc16a4560 <tmpfs_fs_type>, flags=-2147483648,
data=data@entry=0x0, fill_super=0xc10e49b0 <shmem_fill_super>,
mnt=mnt@entry=0xc701d300) at fs/super.c:859
859 {
(gdb) n
861 struct super_block *s = sget(fs_type, NULL, set_anon_super, NULL);
(gdb)
863 if (IS_ERR(s))
(gdb)
866 s->s_flags = flags;
(gdb)
868 error = fill_super(s, data, flags & MS_SILENT ? 1 : 0);
(gdb) s
shmem_fill_super (sb=0xc7009920, data=0x0, silent=0) at mm/shmem.c:2319
2319 sbinfo = kzalloc(max((int)sizeof(struct shmem_sb_info),
(gdb)
2321 if (!sbinfo)
(gdb)
2324 sbinfo->mode = S_IRWXUGO | S_ISVTX;
(gdb)
2325 sbinfo->uid = current_fsuid();
(gdb)
2326 sbinfo->gid = current_fsgid();
(gdb)
2327 sb->s_fs_info = sbinfo;
(gdb)
2335 if (!(sb->s_flags & MS_NOUSER)) {
(gdb)
2343 sb->s_export_op = &shmem_export_ops;
(gdb)
2348 spin_lock_init(&sbinfo->stat_lock);
(gdb)
2349 sbinfo->free_blocks = sbinfo->max_blocks;
(gdb)
2350 sbinfo->free_inodes = sbinfo->max_inodes;
(gdb)
2352 sb->s_maxbytes = SHMEM_MAX_BYTES;
(gdb)
2353 sb->s_blocksize = PAGE_CACHE_SIZE;
(gdb)
2354 sb->s_blocksize_bits = PAGE_CACHE_SHIFT;
(gdb)
2355 sb->s_magic = TMPFS_MAGIC;
(gdb)
2356 sb->s_op = &shmem_ops;
(gdb)
2357 sb->s_time_gran = 1;
(gdb)
2359 sb->s_xattr = shmem_xattr_handlers;
(gdb)
2360 sb->s_flags |= MS_POSIXACL;
(gdb)
2363 inode = shmem_get_inode(sb, S_IFDIR | sbinfo->mode, 0, VM_NORESERVE);
(gdb) s
shmem_get_inode (sb=sb@entry=0xc7009920, mode=17407, dev=dev@entry=0,
flags=flags@entry=2097152) at mm/shmem.c:1548
1548 if (shmem_reserve_inode(sb))
(gdb) s
shmem_reserve_inode (sb=sb@entry=0xc7009920) at mm/shmem.c:245
245 struct shmem_sb_info *sbinfo = SHMEM_SB(sb);
(gdb) n
246 if (sbinfo->max_inodes) {
(gdb)
255 return 0;
(gdb)
256 }
(gdb)
shmem_get_inode (sb=sb@entry=0xc7009920, mode=17407, dev=dev@entry=0,
flags=flags@entry=2097152) at mm/shmem.c:1551
1551 inode = new_inode(sb);
(gdb) s
new_inode (sb=sb@entry=0xc7009920) at fs/inode.c:677
677 spin_lock_prefetch(&inode_lock);
(gdb) n
679 inode = alloc_inode(sb);
(gdb) s
alloc_inode (sb=sb@entry=0xc7009920) at fs/inode.c:213
213 {
(gdb) n
216 if (sb->s_op->alloc_inode)
(gdb)
217 inode = sb->s_op->alloc_inode(sb);
(gdb) s
shmem_alloc_inode (sb=0xc7009920) at mm/shmem.c:2386
2386 p = (struct shmem_inode_info *)kmem_cache_alloc(shmem_inode_cachep, GFP_KERNEL);
(gdb) n
2387 if (!p)
(gdb)
2389 return &p->vfs_inode;
(gdb)
2390 }
(gdb)
alloc_inode (sb=sb@entry=0xc7009920) at fs/inode.c:221
221 if (!inode)
(gdb)
224 if (unlikely(inode_init_always(sb, inode))) {
(gdb)
233 }
(gdb)
new_inode (sb=sb@entry=0xc7009920) at fs/inode.c:680
680 if (inode) {
(gdb)
679 inode = alloc_inode(sb);
(gdb)
680 if (inode) {
(gdb)
681 spin_lock(&inode_lock);
(gdb)
682 __inode_add_to_lists(sb, NULL, inode);
(gdb)
683 inode->i_ino = ++last_ino;
(gdb)
684 inode->i_state = 0;
(gdb)
685 spin_unlock(&inode_lock);
(gdb)
688 }
(gdb)
shmem_get_inode (sb=sb@entry=0xc7009920, mode=17407, dev=dev@entry=0,
flags=flags@entry=2097152) at mm/shmem.c:1552
1552 if (inode) {
(gdb)
1553 inode->i_mode = mode;
(gdb)
1554 inode->i_uid = current_fsuid();
(gdb)
1555 inode->i_gid = current_fsgid();
(gdb)
1556 inode->i_blocks = 0;
(gdb)
1557 inode->i_mapping->backing_dev_info = &shmem_backing_dev_info;
(gdb)
1558 inode->i_atime = inode->i_mtime = inode->i_ctime = CURRENT_TIME;
(gdb)
1559 inode->i_generation = get_seconds();
(gdb)
1560 info = SHMEM_I(inode);
(gdb)
1561 memset(info, 0, (char *)inode - (char *)info);
(gdb)
1562 spin_lock_init(&info->lock);
(gdb)
1563 info->flags = flags & VM_NORESERVE;
(gdb)
1564 INIT_LIST_HEAD(&info->swaplist);
(gdb)
1565 cache_no_acl(inode);
(gdb)
1567 switch (mode & S_IFMT) {
(gdb)
1580 inc_nlink(inode);
(gdb)
1582 inode->i_size = 2 * BOGO_DIRENT_SIZE;
(gdb)
1583 inode->i_op = &shmem_dir_inode_operations; # operations of inode
(gdb)
1584 inode->i_fop = &simple_dir_operations;
(gdb)
1597 }
(gdb)
shmem_fill_super (sb=0xc7009920, data=0x0, silent=<optimized out>)
at mm/shmem.c:2364
2364 if (!inode)
(gdb)
2366 inode->i_uid = sbinfo->uid;
(gdb)
2367 inode->i_gid = sbinfo->gid;
(gdb)
2368 root = d_alloc_root(inode);
(gdb)
(gdb) s
d_alloc_root (root_inode=root_inode@entry=0xc709e080) at fs/dcache.c:1098
1098 if (root_inode) {
(gdb) n
1101 res = d_alloc(NULL, &name);
(gdb) s
d_alloc (parent=parent@entry=0x0, name=name@entry=0xc149fbb8 <name>)
at fs/dcache.c:922
922 dentry = kmem_cache_alloc(dentry_cache, GFP_KERNEL);
(gdb)
923 if (!dentry)
(gdb)
926 if (name->len > DNAME_INLINE_LEN-1) {
(gdb)
933 dname = dentry->d_iname;
(gdb)
935 dentry->d_name.name = dname;
(gdb)
937 dentry->d_name.len = name->len;
(gdb)
938 dentry->d_name.hash = name->hash;
(gdb)
939 memcpy(dname, name->name, name->len);
(gdb)
940 dname[name->len] = 0;
(gdb)
942 atomic_set(&dentry->d_count, 1);
(gdb)
943 dentry->d_flags = DCACHE_UNHASHED;
(gdb)
944 spin_lock_init(&dentry->d_lock);
(gdb)
945 dentry->d_inode = NULL;
(gdb)
946 dentry->d_parent = NULL;
(gdb)
947 dentry->d_sb = NULL;
(gdb)
948 dentry->d_op = NULL;
(gdb)
949 dentry->d_fsdata = NULL;
(gdb)
950 dentry->d_mounted = 0;
(gdb)
951 INIT_HLIST_NODE(&dentry->d_hash);
(gdb)
952 INIT_LIST_HEAD(&dentry->d_lru);
(gdb)
953 INIT_LIST_HEAD(&dentry->d_subdirs);
(gdb)
954 INIT_LIST_HEAD(&dentry->d_alias);
(gdb)
956 if (parent) {
(gdb)
960 INIT_LIST_HEAD(&dentry->d_u.d_child);
(gdb)
963 spin_lock(&dcache_lock);
(gdb)
966 dentry_stat.nr_dentry++;
(gdb)
967 spin_unlock(&dcache_lock);
(gdb)
969 return dentry;
(gdb)
970 }
(gdb)
d_alloc_root (root_inode=root_inode@entry=0xc709e080) at fs/dcache.c:1102
1102 if (res) {
(gdb)
1103 res->d_sb = root_inode->i_sb;
(gdb)
1104 res->d_parent = res;
(gdb)
1105 d_instantiate(res, root_inode);
(gdb) s
d_instantiate (entry=0xc6c02340, inode=0xc709e080) at fs/dcache.c:1007
1007 {
(gdb) n
1008 BUG_ON(!list_empty(&entry->d_alias));
(gdb)
1009 spin_lock(&dcache_lock);
(gdb)
1010 __d_instantiate(entry, inode);
(gdb) s
__d_instantiate (dentry=dentry@entry=0xc6c02340, inode=inode@entry=0xc709e080)
at fs/dcache.c:985
985 if (inode)
(gdb) n
986 list_add(&dentry->d_alias, &inode->i_dentry);
(gdb)
987 dentry->d_inode = inode;
(gdb)
988 fsnotify_d_instantiate(dentry, inode);
(gdb) s
fsnotify_d_instantiate (inode=0xc709e080, entry=0xc6c02340) at fs/dcache.c:988
988 fsnotify_d_instantiate(dentry, inode);
(gdb) s
__fsnotify_d_instantiate (inode=0xc709e080, dentry=0xc6c02340)
at fs/dcache.c:988
988 fsnotify_d_instantiate(dentry, inode);
(gdb) n
989 }
(gdb)
d_instantiate (entry=0xc6c02340, inode=0xc709e080) at fs/dcache.c:1011
1011 spin_unlock(&dcache_lock);
(gdb)
1012 security_d_instantiate(entry, inode);
(gdb)
1013 }
(gdb)
d_alloc_root (root_inode=root_inode@entry=0xc709e080) at fs/dcache.c:1109
1109 }
(gdb)
shmem_fill_super (sb=0xc7009920, data=<optimized out>, silent=<optimized out>)
at mm/shmem.c:2369
2369 if (!root)
(gdb)
2371 sb->s_root = root;
(gdb) p root
$2 = (struct dentry *) 0xc6c02340
(gdb) p *root
$3 = {d_count = {counter = 1}, d_flags = 16, d_lock = {raw_lock = {
slock = 514}, magic = 3735899821, owner_cpu = 4294967295,
owner = 0xffffffff, dep_map = {key = 0xc1e6248c <__key.26347>,
class_cache = 0x0, name = 0xc15ef0b2 "&dentry->d_lock", cpu = 0,
ip = 3239351818}}, d_mounted = 0, d_inode = 0xc709e080, d_hash = {
next = 0x0, pprev = 0x0}, d_parent = 0xc6c02340, d_name = {hash = 0,
len = 1, name = 0xc6c023bc "/"}, d_lru = {next = 0xc6c0238c,
prev = 0xc6c0238c}, d_u = {d_child = {next = 0xc6c02394,
prev = 0xc6c02394}, d_rcu = {next = 0xc6c02394, func = 0xc6c02394}},
d_subdirs = {next = 0xc6c0239c, prev = 0xc6c0239c}, d_alias = {
next = 0xc709e098, prev = 0xc709e098}, d_time = 1802201963, d_op = 0x0,
d_sb = 0xc7009920, d_fsdata = 0x0,
d_iname = "/\000", 'k' <repeats 37 times>, "\245"}
(gdb) n
2372 return 0;
(gdb)
2379 }
(gdb)
get_sb_nodev (fs_type=<optimized out>, flags=<optimized out>,
data=data@entry=0x0, fill_super=0xc10e49b0 <shmem_fill_super>,
mnt=mnt@entry=0xc701d300) at fs/super.c:869
869 if (error) {
(gdb)
873 s->s_flags |= MS_ACTIVE;
(gdb) n
874 simple_set_mnt(mnt, s);
(gdb) s
simple_set_mnt (mnt=mnt@entry=0xc701d300, sb=sb@entry=0xc7009920)
at fs/namespace.c:394
394 mnt->mnt_sb = sb;
(gdb) n
395 mnt->mnt_root = dget(sb->s_root);
(gdb)
396 }
(gdb)
get_sb_nodev (fs_type=<optimized out>, flags=<optimized out>,
data=data@entry=0x0, fill_super=0xc10e49b0 <shmem_fill_super>,
mnt=mnt@entry=0xc701d300) at fs/super.c:875
875 return 0;
(gdb)
876 }
(gdb)
shmem_get_sb (fs_type=<optimized out>, flags=<optimized out>,
dev_name=<optimized out>, data=0x0, mnt=0xc701d300) at mm/shmem.c:2520
2520 }
(gdb)
vfs_kern_mount (type=type@entry=0xc16a4560 <tmpfs_fs_type>,
flags=flags@entry=-2147483648, name=<optimized out>, data=data@entry=0x0)
at fs/super.c:939
939 if (error < 0)
(gdb)
941 BUG_ON(!mnt->mnt_sb);
(gdb)
943 error = security_sb_kern_mount(mnt->mnt_sb, flags, secdata);
(gdb) s
security_sb_kern_mount (sb=0xc7009920, flags=flags@entry=-2147483648,
data=data@entry=0x0) at security/security.c:273
273 return security_ops->sb_kern_mount(sb, flags, data);
(gdb) s
cap_sb_kern_mount (sb=0xc7009920, flags=-2147483648, data=0x0)
at security/capability.c:65
65 }
(gdb) n
security_sb_kern_mount (sb=<optimized out>, flags=flags@entry=-2147483648,
data=data@entry=0x0) at security/security.c:274
274 }
(gdb)
vfs_kern_mount (type=type@entry=0xc16a4560 <tmpfs_fs_type>,
flags=flags@entry=-2147483648, name=<optimized out>, data=data@entry=0x0)
at fs/super.c:944
944 if (error)
(gdb)
954 WARN((mnt->mnt_sb->s_maxbytes < 0), "%s set sb->s_maxbytes to "
(gdb)
957 mnt->mnt_mountpoint = mnt->mnt_root;
(gdb)
958 mnt->mnt_parent = mnt;
(gdb)
959 up_write(&mnt->mnt_sb->s_umount);
(gdb)
960 free_secdata(secdata);
(gdb)
961 return mnt;
(gdb)
971 }
(gdb)
init_tmpfs () at mm/shmem.c:2549
2549 if (IS_ERR(shm_mnt)) {
(gdb)
2554 return 0;
(gdb)
2565 }
(gdb)
do_basic_setup () at init/main.c:787
787 driver_init();
initialize driver model
devtmpfs_init: initialize dev tmp file system
devices_init: initialize devices, it involke kset_create_and_add to create kset data. struct kset - a set of kobjects of a specific type, belonging to a specific subsystem. A kset defines a group of kobjects. They can be individually different "types" but overall these kobjects all want to be grouped together and operated on in the same manner. ksets are used to define the attribute callbacks and other common events that happen to a kobject.
do_basic_setup () at init/main.c:787
787 driver_init();
(gdb) s
driver_init () at drivers/base/init.c:23
23 devtmpfs_init(); # initialize dev tmp filesystem
(gdb) s
devtmpfs_init () at drivers/base/devtmpfs.c:369
369 char options[] = "mode=0755";
(gdb) n
371 err = register_filesystem(&dev_fs_type);
(gdb)
372 if (err) {
(gdb)
378 mnt = kern_mount_data(&dev_fs_type, options);
(gdb) s
kern_mount_data (type=type@entry=0xc16c7da0 <dev_fs_type>,
data=data@entry=0xc706bfa6) at fs/super.c:1016
1016 return vfs_kern_mount(type, MS_KERNMOUNT, type->name, data); # kern_mount_data invoke vfs_kern_mount, reference tmp file system
(gdb) n
1017 }
(gdb)
devtmpfs_init () at drivers/base/devtmpfs.c:379
379 if (IS_ERR(mnt)) {
(gdb)
387 printk(KERN_INFO "devtmpfs: initialized\n");
(gdb) p mnt
$1 = (struct vfsmount *) 0xc701d3c0
(gdb) p *mnt
$2 = {mnt_hash = {next = 0xc701d3c0, prev = 0xc701d3c0},
mnt_parent = 0xc701d3c0, mnt_mountpoint = 0xc6c02410, mnt_root = 0xc6c02410,
mnt_sb = 0xc7009d50, mnt_mounts = {next = 0xc701d3d8, prev = 0xc701d3d8},
mnt_child = {next = 0xc701d3e0, prev = 0xc701d3e0}, mnt_flags = 0,
mnt_devname = 0xc7000700 "devtmpfs", mnt_list = {next = 0xc701d3f0,
prev = 0xc701d3f0}, mnt_expire = {next = 0xc701d3f8, prev = 0xc701d3f8},
mnt_share = {next = 0xc701d400, prev = 0xc701d400}, mnt_slave_list = {
next = 0xc701d408, prev = 0xc701d408}, mnt_slave = {next = 0xc701d410,
prev = 0xc701d410}, mnt_master = 0x0, mnt_ns = 0x0, mnt_id = 5,
mnt_group_id = 0, mnt_count = {counter = 1}, mnt_expiry_mark = 0,
mnt_pinned = 0, mnt_ghosts = 0, mnt_writers = 0xc18d3a50}
(gdb) n
385 dev_mnt = mnt;
(gdb)
387 printk(KERN_INFO "devtmpfs: initialized\n");
(gdb)
388 return 0;
(gdb)
389 }
(gdb)
driver_init () at drivers/base/init.c:24
24 devices_init();
(gdb) s
devices_init () at drivers/base/core.c:1275
1275 devices_kset = kset_create_and_add("devices", &device_uevent_ops, NULL);
(gdb) break kset_create_and_add
Breakpoint 4 at 0xc1239840: file lib/kobject.c, line 837.
(gdb) c
Continuing.
Breakpoint 4, kset_create_and_add (name=name@entry=0xc15d7009 "devices",
uevent_ops=uevent_ops@entry=0xc16c78f8 <device_uevent_ops>,
parent_kobj=parent_kobj@entry=0x0) at lib/kobject.c:837
841 kset = kset_create(name, uevent_ops, parent_kobj); # allocate memory for kset and initialize it
(gdb) n
844 error = kset_register(kset);
(gdb) s
kset_register (k=k@entry=0xc70851b0) at lib/kobject.c:716
716 if (!k)
(gdb) n
719 kset_init(k);
(gdb) s
kset_init (k=k@entry=0xc70851b0) at lib/kobject.c:672
672 {
(gdb) n
673 kobject_init_internal(&k->kobj);
(gdb) s
kobject_init_internal (kobj=0xc70851dc) at lib/kobject.c:147
147 if (!kobj)
(gdb) n
149 kref_init(&kobj->kref);
(gdb)
150 INIT_LIST_HEAD(&kobj->entry);
(gdb)
151 kobj->state_in_sysfs = 0;
(gdb)
153 kobj->state_remove_uevent_sent = 0;
(gdb)
154 kobj->state_initialized = 1;
(gdb)
kset_init (k=k@entry=0xc70851b0) at lib/kobject.c:674
674 INIT_LIST_HEAD(&k->list);
(gdb)
675 spin_lock_init(&k->list_lock);
(gdb)
676 }
(gdb)
kset_register (k=k@entry=0xc70851b0) at lib/kobject.c:720
720 err = kobject_add_internal(&k->kobj);
(gdb) s
kobject_add_internal (kobj=0xc70851dc) at lib/kobject.c:163
163 if (!kobj)
(gdb) n
166 if (!kobj->name || !kobj->name[0]) {
(gdb) p *kobj
$1 = {name = 0xc7020230 "devices", entry = {next = 0xc70851e0,
prev = 0xc70851e0}, parent = 0x0, kset = 0x0,
ktype = 0xc16bb100 <kset_ktype>, sd = 0x0, kref = {refcount = {
counter = 1}}, state_initialized = 1, state_in_sysfs = 0,
state_add_uevent_sent = 0, state_remove_uevent_sent = 0, uevent_suppress = 0}
(gdb) n
172 parent = kobject_get(kobj->parent);
(gdb) s
kobject_get (kobj=0x0) at lib/kobject.c:529
529 if (kobj)
(gdb) n
kobject_add_internal (kobj=0xc70851dc) at lib/kobject.c:175
175 if (kobj->kset) {
(gdb)
182 pr_debug("kobject: '%s' (%p): %s: parent: '%s', set: '%s'\n",
(gdb)
187 error = create_dir(kobj);
(gdb) s
create_dir (kobj=0xc70851dc) at lib/kobject.c:50
50 if (kobject_name(kobj)) {
(gdb) s
51 error = sysfs_create_dir(kobj);
(gdb) s
sysfs_create_dir (kobj=kobj@entry=0xc70851dc) at fs/sysfs/dir.c:715
715 BUG_ON(!kobj);
(gdb) n
717 if (kobj->parent)
(gdb)
720 parent_sd = &sysfs_root;
(gdb)
722 error = create_dir(kobj, parent_sd, kobject_name(kobj), &sd);
(gdb) s
create_dir (kobj=kobj@entry=0xc70851dc, parent_sd=0xc16ac580 <sysfs_root>,
name=0xc7020230 "devices", p_sd=p_sd@entry=0xc706bf48)
at fs/sysfs/dir.c:675
675 {
(gdb) n
682 sd = sysfs_new_dirent(name, mode, SYSFS_DIR);
(gdb) s
sysfs_new_dirent (name=0xc7020230 "devices", mode=mode@entry=16877,
type=type@entry=1) at fs/sysfs/dir.c:318
318 if (type & SYSFS_COPY_NAME) {
(gdb)
319 name = dup_name = kstrdup(name, GFP_KERNEL);
(gdb)
320 if (!name)
(gdb)
324 sd = kmem_cache_zalloc(sysfs_dir_cachep, GFP_KERNEL);
(gdb)
325 if (!sd)
(gdb)
328 if (sysfs_alloc_ino(&sd->s_ino))
(gdb) s
sysfs_alloc_ino (pino=<optimized out>) at fs/sysfs/dir.c:264
264 *pino = ino;
(gdb) p ino
$2 = 3
(gdb) n
sysfs_new_dirent (name=0xc7020268 "devices", mode=mode@entry=16877,
type=type@entry=1) at fs/sysfs/dir.c:331
331 atomic_set(&sd->s_count, 1);
(gdb)
332 atomic_set(&sd->s_active, 0);
(gdb)
334 sd->s_name = name;
(gdb)
335 sd->s_mode = mode;
(gdb)
336 sd->s_flags = type;
(gdb)
338 return sd;
(gdb)
345 }
(gdb)
create_dir (kobj=kobj@entry=0xc70851dc, parent_sd=0xc16ac580 <sysfs_root>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf48) at fs/sysfs/dir.c:683
683 if (!sd)
(gdb)
685 sd->s_dir.kobj = kobj;
(gdb)
688 sysfs_addrm_start(&acxt, parent_sd);
(gdb) s
sysfs_addrm_start (acxt=acxt@entry=0xc706bf20,
parent_sd=parent_sd@entry=0xc16ac580 <sysfs_root>) at fs/sysfs/dir.c:374
374 memset(acxt, 0, sizeof(*acxt));
(gdb) n
375 acxt->parent_sd = parent_sd;
(gdb)
381 mutex_lock(&sysfs_mutex);
(gdb)
383 inode = ilookup5(sysfs_sb, parent_sd->s_ino, sysfs_ilookup_test,
(gdb) s
ilookup5 (sb=0xc7008430, hashval=1,
test=test@entry=0xc1170180 <sysfs_ilookup_test>,
data=data@entry=0xc16ac580 <sysfs_root>) at fs/inode.c:1001
1001 struct hlist_head *head = inode_hashtable + hash(sb, hashval);
(gdb) n
1003 return ifind(sb, head, test, data, 1);
(gdb) s
ifind (sb=0xc7008430, head=0xc213e004,
test=test@entry=0xc1170180 <sysfs_ilookup_test>,
data=data@entry=0xc16ac580 <sysfs_root>, wait=wait@entry=1)
at fs/inode.c:904
904 spin_lock(&inode_lock);
(gdb) n
905 inode = find_inode(sb, head, test, data);
(gdb) s
find_inode (sb=sb@entry=0xc7008430, head=head@entry=0xc213e004,
test=test@entry=0xc1170180 <sysfs_ilookup_test>,
data=data@entry=0xc16ac580 <sysfs_root>) at fs/inode.c:573
573 hlist_for_each_entry(inode, node, head, i_hash) {
(gdb) n
574 if (inode->i_sb != sb)
(gdb) p inode
$3 = (struct inode *) 0xc6c00000
(gdb) p *inode
$4 = {i_hash = {next = 0x0, pprev = 0xc213e004}, i_list = {
next = 0xc16a946c <inode_in_use>, prev = 0xc6c00310}, i_sb_list = {
next = 0xc7008514, prev = 0xc7008514}, i_dentry = {next = 0xc6c02064,
prev = 0xc6c02064}, i_ino = 1, i_count = {counter = 1}, i_nlink = 3,
i_uid = 0, i_gid = 0, i_rdev = 0, i_version = 0, i_size = 0,
i_size_seqcount = {sequence = 0}, i_atime = {tv_sec = 1495595682,
tv_nsec = 16001000}, i_mtime = {tv_sec = 1495595682, tv_nsec = 20001250},
i_ctime = {tv_sec = 1495595682, tv_nsec = 20001250}, i_blocks = 0,
i_blkbits = 12, i_bytes = 0, i_mode = 16877, i_lock = {raw_lock = {
slock = 0}, magic = 3735899821, owner_cpu = 4294967295,
owner = 0xffffffff, dep_map = {key = 0xc16ac550 <sysfs_fs_type+48>,
class_cache = 0x0, name = 0xc15ef1bb "&sb->s_type->i_lock_key", cpu = 0,
ip = 0}}, i_mutex = {count = {counter = 1}, wait_lock = {raw_lock = {
slock = 514}, magic = 3735899821, owner_cpu = 4294967295,
owner = 0xffffffff, dep_map = {key = 0xc190e388 <__key.15367>,
class_cache = 0x0, name = 0xc15d6430 "&lock->wait_lock", cpu = 0,
ip = 0}}, wait_list = {next = 0xc6c000c0, prev = 0xc6c000c0},
owner = 0x0, name = 0x0, magic = 0xc6c00098, dep_map = {
key = 0xc16ac560 <sysfs_fs_type+64>,
class_cache = 0xc1b5deb0 <lock_classes+145008>,
name = 0xc15ef22d "&type->i_mutex_dir_key", cpu = 0, ip = 0}},
i_alloc_sem = {count = 0, wait_lock = {raw_lock = {slock = 0},
magic = 3735899821, owner_cpu = 4294967295, owner = 0xffffffff,
---Type <return> to continue, or q <return> to quit---
dep_map = {key = 0xc1e6ccd4 <__key.14570>, class_cache = 0x0,
name = 0xc1612d38 "&sem->wait_lock", cpu = 0, ip = 0}}, wait_list = {
next = 0xc6c00110, prev = 0xc6c00110}, dep_map = {
key = 0xc16ac568 <sysfs_fs_type+72>, class_cache = 0x0,
name = 0xc15ef210 "&sb->s_type->i_alloc_sem_key", cpu = 0, ip = 0}},
i_op = 0xc14a36e0 <sysfs_dir_inode_operations>,
i_fop = 0xc14a3660 <sysfs_dir_operations>, i_sb = 0xc7008430, i_flock = 0x0,
i_mapping = 0xc6c00140, i_data = {host = 0xc6c00000, page_tree = {
height = 0, gfp_mask = 32, rnode = 0x0}, tree_lock = {raw_lock = {
slock = 0}, magic = 3735899821, owner_cpu = 4294967295,
owner = 0xffffffff, dep_map = {key = 0xc1e624ec <__key.26256>,
class_cache = 0x0, name = 0xc15ef13f "&mapping->tree_lock", cpu = 0,
ip = 0}}, i_mmap_writable = 0, i_mmap = {prio_tree_node = 0x0,
index_bits = 1, raw = 1}, i_mmap_nonlinear = {next = 0xc6c00180,
prev = 0xc6c00180}, i_mmap_lock = {raw_lock = {slock = 0},
magic = 3735899821, owner_cpu = 4294967295, owner = 0xffffffff,
dep_map = {key = 0xc1e624e4 <__key.26257>, class_cache = 0x0,
name = 0xc15ef153 "&mapping->i_mmap_lock", cpu = 0, ip = 0}},
truncate_count = 0, nrpages = 0, writeback_index = 0,
a_ops = 0xc14a3540 <sysfs_aops>, flags = 131290,
backing_dev_info = 0xc16ac140 <sysfs_backing_dev_info>, private_lock = {
raw_lock = {slock = 0}, magic = 3735899821, owner_cpu = 4294967295,
owner = 0xffffffff, dep_map = {key = 0xc1e624dc <__key.26258>,
---Type <return> to continue, or q <return> to quit---
class_cache = 0x0, name = 0xc15ef169 "&mapping->private_lock",
cpu = 0, ip = 0}}, private_list = {next = 0xc6c001e8,
prev = 0xc6c001e8}, assoc_mapping = 0x0, unmap_mutex = {count = {
counter = 1}, wait_lock = {raw_lock = {slock = 0}, magic = 3735899821,
owner_cpu = 4294967295, owner = 0xffffffff, dep_map = {
key = 0xc190e388 <__key.15367>, class_cache = 0x0,
name = 0xc15d6430 "&lock->wait_lock", cpu = 0, ip = 0}},
wait_list = {next = 0xc6c0021c, prev = 0xc6c0021c}, owner = 0x0,
name = 0x0, magic = 0xc6c001f4, dep_map = {
key = 0xc1e624d4 <__key.26259>, class_cache = 0x0,
name = 0xc15ef180 "&mapping->unmap_mutex", cpu = 0, ip = 0}}},
i_dquot = {0x0, 0x0}, i_devices = {next = 0xc6c0024c, prev = 0xc6c0024c}, {
i_pipe = 0x0, i_bdev = 0x0, i_cdev = 0x0}, i_generation = 0,
i_fsnotify_mask = 0, i_fsnotify_mark_entries = {first = 0x0},
inotify_watches = {next = 0xc6c00264, prev = 0xc6c00264}, inotify_mutex = {
count = {counter = 1}, wait_lock = {raw_lock = {slock = 0},
magic = 3735899821, owner_cpu = 4294967295, owner = 0xffffffff,
dep_map = {key = 0xc190e388 <__key.15367>, class_cache = 0x0,
name = 0xc15d6430 "&lock->wait_lock", cpu = 0, ip = 0}}, wait_list = {
next = 0xc6c00294, prev = 0xc6c00294}, owner = 0x0, name = 0x0,
magic = 0xc6c0026c, dep_map = {key = 0xc1e624cc <__key.26270>,
class_cache = 0x0, name = 0xc15ef196 "&inode->inotify_mutex", cpu = 0,
ip = 0}}, i_state = 0, dirtied_when = 0, i_flags = 0, i_writecount = {
---Type <return> to continue, or q <return> to quit---
counter = 0}, i_security = 0x0, i_acl = 0xffffffff,
i_default_acl = 0xffffffff, i_private = 0xc16ac580 <sysfs_root>}
(gdb) n
576 if (!test(inode, data))
(gdb)
578 if (inode->i_state & (I_FREEING|I_CLEAR|I_WILL_FREE)) {
(gdb)
579 __wait_on_freeing_inode(inode);
(gdb)
585 }
(gdb)
ifind (sb=0xc7008430, head=0xc213e004,
test=test@entry=0xc1170180 <sysfs_ilookup_test>,
data=data@entry=0xc16ac580 <sysfs_root>, wait=wait@entry=1)
at fs/inode.c:906
906 if (inode) {
(gdb)
907 __iget(inode);
(gdb)
908 spin_unlock(&inode_lock);
(gdb)
909 if (likely(wait))
(gdb)
910 wait_on_inode(inode);
(gdb)
915 }
(gdb)
ilookup5 (sb=<optimized out>, hashval=<optimized out>,
test=test@entry=0xc1170180 <sysfs_ilookup_test>,
data=data@entry=0xc16ac580 <sysfs_root>) at fs/inode.c:1004
1004 }
(gdb)
sysfs_addrm_start (acxt=acxt@entry=0xc706bf20,
parent_sd=parent_sd@entry=0xc16ac580 <sysfs_root>) at fs/sysfs/dir.c:385
385 if (inode) {
(gdb)
386 WARN_ON(inode->i_state & I_NEW);
(gdb)
389 acxt->parent_inode = inode;
(gdb)
395 if (!mutex_trylock(&inode->i_mutex)) {
(gdb)
401 }
(gdb)
create_dir (kobj=kobj@entry=0xc70851dc, parent_sd=0xc16ac580 <sysfs_root>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf48) at fs/sysfs/dir.c:689
689 rc = sysfs_add_one(&acxt, sd);
(gdb) s
sysfs_add_one (acxt=acxt@entry=0xc706bf20, sd=sd@entry=0xc7022058)
at fs/sysfs/dir.c:482
482 ret = __sysfs_add_one(acxt, sd);
(gdb) s
__sysfs_add_one (acxt=acxt@entry=0xc706bf20, sd=sd@entry=0xc7022058)
at fs/sysfs/dir.c:425
425 if (sysfs_find_dirent(acxt->parent_sd, sd->s_name))
(gdb) s
sysfs_find_dirent (parent_sd=0xc16ac580 <sysfs_root>,
name=0xc7020268 "devices") at fs/sysfs/dir.c:639
639 for (sd = parent_sd->s_dir.children; sd; sd = sd->s_sibling)
(gdb) n
640 if (!strcmp(sd->s_name, name))
(gdb)
642 return NULL;
(gdb)
643 }
(gdb)
__sysfs_add_one (acxt=acxt@entry=0xc706bf20, sd=sd@entry=0xc7022058)
at fs/sysfs/dir.c:428
428 sd->s_parent = sysfs_get(acxt->parent_sd);
(gdb) s
__sysfs_get (sd=0xc16ac580 <sysfs_root>) at fs/sysfs/sysfs.h:138
138 if (sd) {
(gdb) n
139 WARN_ON(!atomic_read(&sd->s_count));
(gdb)
140 atomic_inc(&sd->s_count);
(gdb)
__sysfs_add_one (acxt=acxt@entry=0xc706bf20, sd=sd@entry=0xc7022058)
at fs/sysfs/dir.c:430
430 if (sysfs_type(sd) == SYSFS_DIR && acxt->parent_inode)
(gdb)
428 sd->s_parent = sysfs_get(acxt->parent_sd);
(gdb)
430 if (sysfs_type(sd) == SYSFS_DIR && acxt->parent_inode)
(gdb)
431 inc_nlink(acxt->parent_inode);
(gdb)
433 acxt->cnt++;
(gdb)
435 sysfs_link_sibling(sd);
(gdb) s
sysfs_link_sibling (sd=0xc7022058) at fs/sysfs/dir.c:49
49 BUG_ON(sd->s_sibling);
(gdb) n
46 struct sysfs_dirent *parent_sd = sd->s_parent;
(gdb)
49 BUG_ON(sd->s_sibling);
(gdb)
55 for (pos = &parent_sd->s_dir.children; *pos; pos = &(*pos)->s_sibling) {
(gdb)
56 if (sd->s_ino < (*pos)->s_ino)
(gdb)
55 for (pos = &parent_sd->s_dir.children; *pos; pos = &(*pos)->s_sibling) {
(gdb)
59 sd->s_sibling = *pos;
(gdb)
60 *pos = sd;
(gdb)
61 }
(gdb)
__sysfs_add_one (acxt=acxt@entry=0xc706bf20, sd=sd@entry=0xc7022058)
at fs/sysfs/dir.c:437
437 return 0;
(gdb)
438 }
(gdb)
sysfs_add_one (acxt=acxt@entry=0xc706bf20, sd=sd@entry=0xc7022058)
at fs/sysfs/dir.c:483
483 if (ret == -EEXIST) {
(gdb)
482 ret = __sysfs_add_one(acxt, sd);
(gdb)
483 if (ret == -EEXIST) {
(gdb)
496 }
(gdb)
create_dir (kobj=kobj@entry=0xc70851dc, parent_sd=<optimized out>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf48) at fs/sysfs/dir.c:690
690 sysfs_addrm_finish(&acxt);
(gdb) s
sysfs_addrm_finish (acxt=acxt@entry=0xc706bf20) at fs/sysfs/dir.c:593
593 {
(gdb) n
595 mutex_unlock(&sysfs_mutex);
(gdb)
596 if (acxt->parent_inode) {
(gdb)
600 if (acxt->cnt)
(gdb)
601 inode->i_ctime = inode->i_mtime = CURRENT_TIME;
(gdb)
603 mutex_unlock(&inode->i_mutex);
(gdb)
604 iput(inode);
(gdb)
608 while (acxt->removed) {
(gdb)
619 }
(gdb)
create_dir (kobj=kobj@entry=0xc70851dc, parent_sd=<optimized out>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf48) at fs/sysfs/dir.c:692
692 if (rc == 0)
(gdb)
693 *p_sd = sd;
(gdb)
698 }
(gdb)
sysfs_create_dir (kobj=kobj@entry=0xc70851dc) at fs/sysfs/dir.c:723
723 if (!error)
(gdb)
724 kobj->sd = sd;
(gdb)
726 }
(gdb)
create_dir (kobj=0xc70851dc) at lib/kobject.c:52
52 if (!error) {
(gdb)
51 error = sysfs_create_dir(kobj);
(gdb)
52 if (!error) {
(gdb)
kobject_add_internal (kobj=0xc70851dc) at lib/kobject.c:187
187 error = create_dir(kobj);
(gdb)
204 kobj->state_in_sysfs = 1;
(gdb)
169 return -EINVAL;
(gdb)
207 }
(gdb)
kset_register (k=k@entry=0xc70851b0) at lib/kobject.c:721
721 if (err)
(gdb)
720 err = kobject_add_internal(&k->kobj);
(gdb)
721 if (err)
(gdb)
723 kobject_uevent(&k->kobj, KOBJ_ADD);
(gdb) s
kobject_uevent (kobj=kobj@entry=0xc70851dc, action=action@entry=KOBJ_ADD)
at lib/kobject_uevent.c:282
282 return kobject_uevent_env(kobj, action, NULL);
(gdb) s
kobject_uevent_env (kobj=kobj@entry=0xc70851dc, action=action@entry=KOBJ_ADD,
envp_ext=envp_ext@entry=0x0) at lib/kobject_uevent.c:93
93 const char *action_string = kobject_actions[action];
(gdb) n
103 pr_debug("kobject: '%s' (%p): %s\n",
(gdb)
91 {
(gdb)
93 const char *action_string = kobject_actions[action];
(gdb)
103 pr_debug("kobject: '%s' (%p): %s\n",
(gdb)
91 {
(gdb)
103 pr_debug("kobject: '%s' (%p): %s\n",
(gdb)
107 top_kobj = kobj;
(gdb)
108 while (!top_kobj->kset && top_kobj->parent)
(gdb)
112 pr_debug("kobject: '%s' (%p): %s: attempted to send uevent "
(gdb)
115 return -EINVAL;
(gdb)
112 pr_debug("kobject: '%s' (%p): %s: attempted to send uevent "
(gdb)
268 }
(gdb)
kobject_uevent (kobj=kobj@entry=0xc70851dc, action=action@entry=KOBJ_ADD)
at lib/kobject_uevent.c:283
283 }
(gdb)
kset_register (k=k@entry=0xc70851b0) at lib/kobject.c:725
725 }
(gdb)
kset_create_and_add (name=name@entry=0xc15d7009 "devices",
uevent_ops=uevent_ops@entry=0xc16c78f8 <device_uevent_ops>,
parent_kobj=parent_kobj@entry=0x0) at lib/kobject.c:845
845 if (error) {
(gdb)
850 }
(gdb)
devices_init () at drivers/base/core.c:1276
1276 if (!devices_kset)
(gdb)
1275 devices_kset = kset_create_and_add("devices", &device_uevent_ops, NULL);
(gdb)
1276 if (!devices_kset)
(gdb)
After kset created, device_init invoke kobject_create_and_add to create kernel objects and register it with sysfs. 3 Types of kobject created, dev for normal device, block for block device and char for charactor device. Below debug information is the procedue of creating kobject dev.
1278 dev_kobj = kobject_create_and_add("dev", NULL);
(gdb) s
kobject_create_and_add (name=name@entry=0xc1629db5 "dev",
parent=parent@entry=0x0) at lib/kobject.c:652
652 kobj = kobject_create();
(gdb) s
kobject_create () at lib/kobject.c:626
626 kobj = kzalloc(sizeof(*kobj), GFP_KERNEL);
(gdb) n
627 if (!kobj)
(gdb)
630 kobject_init(kobj, &dynamic_kobj_ktype);
(gdb) s
kobject_init (kobj=kobj@entry=0xc70860e0,
ktype=ktype@entry=0xc16bb114 <dynamic_kobj_ktype>) at lib/kobject.c:274
274 if (!kobj) {
(gdb) n
278 if (!ktype) {
(gdb)
282 if (kobj->state_initialized) {
(gdb)
289 kobject_init_internal(kobj);
(gdb) s
kobject_init_internal (kobj=0xc70860e0) at lib/kobject.c:149
149 kref_init(&kobj->kref);
(gdb) s
kref_init (kref=kref@entry=0xc70860fc) at lib/kref.c:34
34 kref_set(kref, 1);
(gdb) n
35 }
(gdb)
kobject_init_internal (kobj=0xc70860e0) at lib/kobject.c:150
150 INIT_LIST_HEAD(&kobj->entry);
(gdb)
151 kobj->state_in_sysfs = 0;
(gdb)
153 kobj->state_remove_uevent_sent = 0;
(gdb) n
154 kobj->state_initialized = 1;
(gdb)
kobject_init (kobj=kobj@entry=0xc70860e0,
ktype=ktype@entry=0xc16bb114 <dynamic_kobj_ktype>) at lib/kobject.c:290
290 kobj->ktype = ktype;
(gdb)
296 }
(gdb)
kobject_create () at lib/kobject.c:631
631 return kobj;
(gdb)
632 }
(gdb)
kobject_create_and_add (name=name@entry=0xc1629db5 "dev",
parent=parent@entry=0x0) at lib/kobject.c:653
653 if (!kobj)
(gdb)
656 retval = kobject_add(kobj, parent, "%s", name);
(gdb) s
kobject_add (kobj=kobj@entry=0xc70860e0, parent=parent@entry=0x0,
fmt=fmt@entry=0xc1627337 "%s") at lib/kobject.c:344
344 if (!kobj)
(gdb) n
347 if (!kobj->state_initialized) {
(gdb)
355 retval = kobject_add_varg(kobj, parent, fmt, args);
(gdb) s
kobject_add_varg (
vargs=0xc706bfa4 "\265\235b\301\360\021v\301`ro\301\\ro\301\274\277\006Q;\r\3
01L\\006\307G@r\301\340\277\006\307\340\250o\301\254S\\\301\001",
fmt=0xc1627337 "%s", parent=0x0, kobj=0xc70860e0) at lib/kobject.c:304
304 retval = kobject_set_name_vargs(kobj, fmt, vargs);
(gdb) n
305 if (retval) {
(gdb)
309 kobj->parent = parent;
(gdb)
310 return kobject_add_internal(kobj);
(gdb) s
kobject_add_internal (kobj=0xc70860e0) at lib/kobject.c:163
163 if (!kobj)
(gdb) p kobj
$1 = (struct kobject *) 0xc70860e0
(gdb) p *kobj
$2 = {name = 0xc70202a0 "dev", entry = {next = 0xc70860e4, prev = 0xc70860e4},
parent = 0x0, kset = 0x0, ktype = 0xc16bb114 <dynamic_kobj_ktype>, sd = 0x0,
kref = {refcount = {counter = 1}}, state_initialized = 1,
state_in_sysfs = 0, state_add_uevent_sent = 0, state_remove_uevent_sent = 0,
uevent_suppress = 0}
(gdb) n
166 if (!kobj->name || !kobj->name[0]) {
(gdb)
172 parent = kobject_get(kobj->parent);
(gdb)
175 if (kobj->kset) {
(gdb)
182 pr_debug("kobject: '%s' (%p): %s: parent: '%s', set: '%s'\n",
(gdb)
187 error = create_dir(kobj);
(gdb) s
create_dir (kobj=0xc70860e0) at lib/kobject.c:50
50 if (kobject_name(kobj)) {
(gdb) n
51 error = sysfs_create_dir(kobj);
(gdb) s
sysfs_create_dir (kobj=kobj@entry=0xc70860e0) at fs/sysfs/dir.c:715
715 BUG_ON(!kobj);
(gdb) n
717 if (kobj->parent)
(gdb)
720 parent_sd = &sysfs_root;
(gdb)
722 error = create_dir(kobj, parent_sd, kobject_name(kobj), &sd);
(gdb) s
create_dir (kobj=kobj@entry=0xc70860e0, parent_sd=0xc16ac580 <sysfs_root>,
name=0xc70202a0 "dev", p_sd=p_sd@entry=0xc706bf38) at fs/sysfs/dir.c:675
675 {
(gdb) n
682 sd = sysfs_new_dirent(name, mode, SYSFS_DIR);
(gdb) s
sysfs_new_dirent (name=0xc70202a0 "dev", mode=mode@entry=16877,
type=type@entry=1) at fs/sysfs/dir.c:318
318 if (type & SYSFS_COPY_NAME) {
(gdb)
319 name = dup_name = kstrdup(name, GFP_KERNEL);
(gdb)
320 if (!name)
(gdb)
324 sd = kmem_cache_zalloc(sysfs_dir_cachep, GFP_KERNEL);
(gdb)
325 if (!sd)
(gdb)
328 if (sysfs_alloc_ino(&sd->s_ino))
(gdb)
331 atomic_set(&sd->s_count, 1);
(gdb)
332 atomic_set(&sd->s_active, 0);
(gdb)
334 sd->s_name = name;
(gdb)
335 sd->s_mode = mode;
(gdb)
336 sd->s_flags = type;
(gdb)
338 return sd;
(gdb)
345 }
(gdb)
create_dir (kobj=kobj@entry=0xc70860e0, parent_sd=0xc16ac580 <sysfs_root>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf38) at fs/sysfs/dir.c:683
683 if (!sd)
(gdb) n
685 sd->s_dir.kobj = kobj;
(gdb)
688 sysfs_addrm_start(&acxt, parent_sd);
(gdb) s
sysfs_addrm_start (acxt=acxt@entry=0xc706bf10,
parent_sd=parent_sd@entry=0xc16ac580 <sysfs_root>) at fs/sysfs/dir.c:374
374 memset(acxt, 0, sizeof(*acxt));
(gdb) n
375 acxt->parent_sd = parent_sd;
(gdb)
381 mutex_lock(&sysfs_mutex);
(gdb)
383 inode = ilookup5(sysfs_sb, parent_sd->s_ino, sysfs_ilookup_test,
(gdb) s
ilookup5 (sb=0xc7008430, hashval=1,
test=test@entry=0xc1170180 <sysfs_ilookup_test>,
data=data@entry=0xc16ac580 <sysfs_root>) at fs/inode.c:1001
1001 struct hlist_head *head = inode_hashtable + hash(sb, hashval);
(gdb) n
1003 return ifind(sb, head, test, data, 1);
(gdb) s
ifind (sb=0xc7008430, head=0xc213e004,
test=test@entry=0xc1170180 <sysfs_ilookup_test>,
data=data@entry=0xc16ac580 <sysfs_root>, wait=wait@entry=1)
at fs/inode.c:904
904 spin_lock(&inode_lock);
(gdb)
905 inode = find_inode(sb, head, test, data);
(gdb) s
find_inode (sb=sb@entry=0xc7008430, head=head@entry=0xc213e004,
test=test@entry=0xc1170180 <sysfs_ilookup_test>,
data=data@entry=0xc16ac580 <sysfs_root>) at fs/inode.c:573
573 hlist_for_each_entry(inode, node, head, i_hash) {
(gdb)
574 if (inode->i_sb != sb)
(gdb) p inode->i_sb
$3 = (struct super_block *) 0xc7008430
(gdb) n
576 if (!test(inode, data))
(gdb)
578 if (inode->i_state & (I_FREEING|I_CLEAR|I_WILL_FREE)) {
(gdb) p inode->i_state
$4 = 0
(gdb) n
579 __wait_on_freeing_inode(inode);
(gdb)
585 }
(gdb)
ifind (sb=0xc7008430, head=0xc213e004,
test=test@entry=0xc1170180 <sysfs_ilookup_test>,
data=data@entry=0xc16ac580 <sysfs_root>, wait=wait@entry=1)
at fs/inode.c:906
906 if (inode) {
(gdb)
907 __iget(inode);
(gdb)
908 spin_unlock(&inode_lock);
(gdb)
909 if (likely(wait))
(gdb)
910 wait_on_inode(inode);
(gdb)
915 }
(gdb)
ilookup5 (sb=<optimized out>, hashval=<optimized out>,
test=test@entry=0xc1170180 <sysfs_ilookup_test>,
data=data@entry=0xc16ac580 <sysfs_root>) at fs/inode.c:1004
1004 }
(gdb)
sysfs_addrm_start (acxt=acxt@entry=0xc706bf10,
parent_sd=parent_sd@entry=0xc16ac580 <sysfs_root>) at fs/sysfs/dir.c:385
385 if (inode) {
(gdb)
386 WARN_ON(inode->i_state & I_NEW);
(gdb)
389 acxt->parent_inode = inode;
(gdb)
395 if (!mutex_trylock(&inode->i_mutex)) {
(gdb)
401 }
(gdb)
create_dir (kobj=kobj@entry=0xc70860e0, parent_sd=0xc16ac580 <sysfs_root>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf38) at fs/sysfs/dir.c:689
689 rc = sysfs_add_one(&acxt, sd);
(gdb) s
sysfs_add_one (acxt=acxt@entry=0xc706bf10, sd=sd@entry=0xc70220b0)
at fs/sysfs/dir.c:482
482 ret = __sysfs_add_one(acxt, sd);
(gdb) s
__sysfs_add_one (acxt=acxt@entry=0xc706bf10, sd=sd@entry=0xc70220b0)
at fs/sysfs/dir.c:425
425 if (sysfs_find_dirent(acxt->parent_sd, sd->s_name))
(gdb) s
sysfs_find_dirent (parent_sd=0xc16ac580 <sysfs_root>, name=0xc70202d8 "dev")
at fs/sysfs/dir.c:639
639 for (sd = parent_sd->s_dir.children; sd; sd = sd->s_sibling)
(gdb) n
640 if (!strcmp(sd->s_name, name))
(gdb) p sd->s_name
$5 = 0xc7020070 "fs"
(gdb) n
642 return NULL;
(gdb)
643 }
(gdb)
__sysfs_add_one (acxt=acxt@entry=0xc706bf10, sd=sd@entry=0xc70220b0)
at fs/sysfs/dir.c:428
428 sd->s_parent = sysfs_get(acxt->parent_sd);
(gdb)
430 if (sysfs_type(sd) == SYSFS_DIR && acxt->parent_inode)
(gdb)
431 inc_nlink(acxt->parent_inode);
(gdb)
433 acxt->cnt++;
(gdb)
435 sysfs_link_sibling(sd);
(gdb) s
sysfs_link_sibling (sd=0xc70220b0) at fs/sysfs/dir.c:46
46 struct sysfs_dirent *parent_sd = sd->s_parent;
(gdb)
49 BUG_ON(sd->s_sibling);
(gdb)
55 for (pos = &parent_sd->s_dir.children; *pos; pos = &(*pos)->s_sibling) {
(gdb)
56 if (sd->s_ino < (*pos)->s_ino)
(gdb)
55 for (pos = &parent_sd->s_dir.children; *pos; pos = &(*pos)->s_sibling) {
(gdb)
59 sd->s_sibling = *pos;
(gdb)
60 *pos = sd;
(gdb)
61 }
(gdb)
__sysfs_add_one (acxt=acxt@entry=0xc706bf10, sd=sd@entry=0xc70220b0)
at fs/sysfs/dir.c:437
437 return 0;
(gdb)
438 }
(gdb)
sysfs_add_one (acxt=acxt@entry=0xc706bf10, sd=sd@entry=0xc70220b0)
at fs/sysfs/dir.c:483
483 if (ret == -EEXIST) {
(gdb)
496 }
(gdb)
create_dir (kobj=kobj@entry=0xc70860e0, parent_sd=<optimized out>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf38) at fs/sysfs/dir.c:690
690 sysfs_addrm_finish(&acxt);
(gdb) s
sysfs_addrm_finish (acxt=acxt@entry=0xc706bf10) at fs/sysfs/dir.c:595
595 mutex_unlock(&sysfs_mutex);
(gdb) n
596 if (acxt->parent_inode) {
(gdb)
600 if (acxt->cnt)
(gdb)
601 inode->i_ctime = inode->i_mtime = CURRENT_TIME;
(gdb)
603 mutex_unlock(&inode->i_mutex);
(gdb)
604 iput(inode);
(gdb)
608 while (acxt->removed) {
(gdb)
619 }
(gdb)
create_dir (kobj=kobj@entry=0xc70860e0, parent_sd=<optimized out>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf38) at fs/sysfs/dir.c:692
692 if (rc == 0)
(gdb)
693 *p_sd = sd;
(gdb)
698 }
(gdb)
sysfs_create_dir (kobj=kobj@entry=0xc70860e0) at fs/sysfs/dir.c:723
723 if (!error)
(gdb)
724 kobj->sd = sd;
(gdb)
726 }
(gdb)
create_dir (kobj=0xc70860e0) at lib/kobject.c:52
52 if (!error) {
(gdb)
kobject_add_internal (kobj=0xc70860e0) at lib/kobject.c:187
187 error = create_dir(kobj);
(gdb)
204 kobj->state_in_sysfs = 1;
(gdb)
207 }
(gdb)
kobject_add (kobj=kobj@entry=0xc70860e0, parent=parent@entry=0x0,
fmt=fmt@entry=0xc1627337 "%s") at lib/kobject.c:359
359 }
(gdb)
kobject_create_and_add (name=name@entry=0xc1629db5 "dev",
parent=parent@entry=0x0) at lib/kobject.c:657
657 if (retval) {
(gdb)
664 }
(gdb)
devices_init () at drivers/base/core.c:1279
1279 if (!dev_kobj)
(gdb)
buses_init, classes_init create respoective kset.
firmware_init, hypervisor_init create respective kobject.
platform_bus_init create platform bus and register it.
33 platform_bus_init();
(gdb) s
platform_bus_init () at drivers/base/platform.c:959
959 early_platform_cleanup(); # clean up platform device resource from list
(gdb) s
early_platform_cleanup () at drivers/base/platform.c:1221
1221 list_for_each_entry_safe(pd, pd2, &early_platform_device_list,
(gdb) p early_platform_device_list
$6 = {next = 0xc1748758 <early_platform_device_list>,
prev = 0xc1748758 <early_platform_device_list>}
(gdb) n
1226 }
(gdb)
platform_bus_init () at drivers/base/platform.c:961
961 error = device_register(&platform_bus);
(gdb) s
device_register (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:1042
1042 device_initialize(dev);
(gdb) p *dev
$7 = {parent = 0x0, p = 0x0, kobj = {name = 0x0, entry = {next = 0x0,
prev = 0x0}, parent = 0x0, kset = 0x0, ktype = 0x0, sd = 0x0, kref = {
refcount = {counter = 0}}, state_initialized = 0, state_in_sysfs = 0,
state_add_uevent_sent = 0, state_remove_uevent_sent = 0,
uevent_suppress = 0}, init_name = 0xc15d9d57 "platform", type = 0x0,
sem = {lock = {raw_lock = {slock = 0}, magic = 0, owner_cpu = 0,
owner = 0x0, dep_map = {key = 0x0, class_cache = 0x0, name = 0x0,
cpu = 0, ip = 0}}, count = 0, wait_list = {next = 0x0, prev = 0x0}},
bus = 0x0, driver = 0x0, platform_data = 0x0, power = {power_state = {
event = 0}, can_wakeup = 0, should_wakeup = 0, status = DPM_INVALID,
entry = {next = 0x0, prev = 0x0}}, dma_mask = 0x0, coherent_dma_mask = 0,
dma_parms = 0x0, dma_pools = {next = 0x0, prev = 0x0}, dma_mem = 0x0,
archdata = {acpi_handle = 0x0}, devt = 0, devres_lock = {raw_lock = {
slock = 0}, magic = 0, owner_cpu = 0, owner = 0x0, dep_map = {key = 0x0,
class_cache = 0x0, name = 0x0, cpu = 0, ip = 0}}, devres_head = {
next = 0x0, prev = 0x0}, knode_class = {n_klist = 0x0, n_node = {
next = 0x0, prev = 0x0}, n_ref = {refcount = {counter = 0}}},
class = 0x0, groups = 0x0, release = 0x0}
(gdb) s
device_initialize (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:560
560 dev->kobj.kset = devices_kset;
(gdb) n
561 kobject_init(&dev->kobj, &device_ktype);
(gdb) s
kobject_init (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
ktype=ktype@entry=0xc16c7904 <device_ktype>) at lib/kobject.c:274
274 if (!kobj) {
(gdb) n
278 if (!ktype) {
(gdb)
282 if (kobj->state_initialized) {
(gdb)
289 kobject_init_internal(kobj);
(gdb) s
kobject_init_internal (kobj=0xc16c7b28 <platform_bus+8>) at lib/kobject.c:149
149 kref_init(&kobj->kref);
(gdb) s
kref_init (kref=kref@entry=0xc16c7b44 <platform_bus+36>) at lib/kref.c:34
34 kref_set(kref, 1);
(gdb) n
35 }
(gdb)
kobject_init_internal (kobj=0xc16c7b28 <platform_bus+8>) at lib/kobject.c:150
150 INIT_LIST_HEAD(&kobj->entry);
(gdb) n
151 kobj->state_in_sysfs = 0;
(gdb)
153 kobj->state_remove_uevent_sent = 0;
(gdb)
154 kobj->state_initialized = 1;
(gdb)
kobject_init (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
ktype=ktype@entry=0xc16c7904 <device_ktype>) at lib/kobject.c:290
290 kobj->ktype = ktype;
(gdb)
296 }
(gdb)
device_initialize (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:562
562 INIT_LIST_HEAD(&dev->dma_pools);
(gdb)
563 init_MUTEX(&dev->sem);
(gdb)
564 spin_lock_init(&dev->devres_lock);
(gdb)
565 INIT_LIST_HEAD(&dev->devres_head);
(gdb)
566 device_init_wakeup(dev, 0);
(gdb) s
device_init_wakeup (val=0, dev=<optimized out>) at include/linux/pm_wakeup.h:35
35 dev->power.can_wakeup = dev->power.should_wakeup = !!val;
(gdb) n
device_initialize (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:567
567 device_pm_init(dev);
(gdb) s
device_pm_init (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/power/main.c:58
58 dev->power.status = DPM_ON;
(gdb) n
60 }
(gdb)
device_initialize (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:569
569 }
(gdb)
device_register (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:1043
1043 return device_add(dev);
(gdb) s
device_add (dev=dev@entry=0xc16c7b20 <platform_bus>) at drivers/base/core.c:897
897 dev = get_device(dev);
(gdb) n
898 if (!dev)
(gdb)
901 if (!dev->p) {
(gdb)
902 error = device_private_init(dev);
(gdb) s
device_private_init (dev=0xc16c7b20 <platform_bus>) at drivers/base/core.c:867
867 dev->p = kzalloc(sizeof(*dev->p), GFP_KERNEL);
(gdb)
868 if (!dev->p)
(gdb)
870 dev->p->device = dev;
(gdb)
871 klist_init(&dev->p->klist_children, klist_children_get,
(gdb)
873 return 0;
(gdb)
874 }
(gdb)
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:903
903 if (error)
(gdb)
912 if (dev->init_name) {
(gdb)
913 dev_set_name(dev, "%s", dev->init_name);
(gdb)
914 dev->init_name = NULL;
(gdb)
917 if (!dev_name(dev))
(gdb) s
dev_name (dev=0xc16c7b20 <platform_bus>) at drivers/base/core.c:917
917 if (!dev_name(dev))
(gdb)
kobject_name (kobj=0xc16c7b28 <platform_bus+8>) at include/linux/kobject.h:81
81 return kobj->name;
(gdb) n
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:920
920 pr_debug("device: '%s': %s\n", dev_name(dev), __func__);
(gdb)
922 parent = get_device(dev->parent);
(gdb)
923 setup_parent(dev, parent);
(gdb) s
setup_parent (parent=0x0, dev=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:680
680 kobj = get_device_parent(dev, parent);
(gdb) s
get_device_parent (dev=dev@entry=0xc16c7b20 <platform_bus>,
parent=parent@entry=0x0) at drivers/base/core.c:605
605 if (dev->class) {
(gdb) p dev->class
$8 = (struct class *) 0x0
(gdb) n
656 if (parent)
(gdb)
658 return NULL;
(gdb)
659 }
(gdb)
setup_parent (parent=0x0, dev=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:681
681 if (kobj)
(gdb)
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:931
931 error = kobject_add(&dev->kobj, dev->kobj.parent, NULL);
(gdb) s
kobject_add (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>, parent=0x0,
fmt=fmt@entry=0x0) at lib/kobject.c:344
344 if (!kobj)
(gdb) n
347 if (!kobj->state_initialized) {
(gdb)
355 retval = kobject_add_varg(kobj, parent, fmt, args);
(gdb) s
kobject_add_varg (vargs=0xc706bf64 "W/", fmt=0x0, parent=0x0,
kobj=0xc16c7b28 <platform_bus+8>) at lib/kobject.c:304
304 retval = kobject_set_name_vargs(kobj, fmt, vargs);
(gdb) n
305 if (retval) {
(gdb)
309 kobj->parent = parent;
(gdb)
310 return kobject_add_internal(kobj);
(gdb) s
kobject_add_internal (kobj=0xc16c7b28 <platform_bus+8>) at lib/kobject.c:163
163 if (!kobj)
(gdb) n
166 if (!kobj->name || !kobj->name[0]) {
(gdb)
172 parent = kobject_get(kobj->parent);
(gdb)
175 if (kobj->kset) {
(gdb)
177 parent = kobject_get(&kobj->kset->kobj);
(gdb)
178 kobj_kset_join(kobj); # add the kobject to its kset's list
(gdb)
179 kobj->parent = parent;
(gdb)
182 pr_debug("kobject: '%s' (%p): %s: parent: '%s', set: '%s'\n",
(gdb)
187 error = create_dir(kobj);
(gdb) p *kobj
$9 = {name = 0xc70007c0 "platform", entry = {next = 0xc70851b0,
prev = 0xc70851b0}, parent = 0xc70851dc, kset = 0xc70851b0,
ktype = 0xc16c7904 <device_ktype>, sd = 0x0, kref = {refcount = {
counter = 2}}, state_initialized = 1, state_in_sysfs = 0,
state_add_uevent_sent = 0, state_remove_uevent_sent = 0, uevent_suppress = 0}
(gdb) s
create_dir (kobj=0xc16c7b28 <platform_bus+8>) at lib/kobject.c:50
50 if (kobject_name(kobj)) {
(gdb) n
51 error = sysfs_create_dir(kobj);
(gdb) s
sysfs_create_dir (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>)
at fs/sysfs/dir.c:715
715 BUG_ON(!kobj);
(gdb) n
717 if (kobj->parent)
(gdb)
718 parent_sd = kobj->parent->sd;
(gdb)
722 error = create_dir(kobj, parent_sd, kobject_name(kobj), &sd);
(gdb) s
create_dir (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>, parent_sd=0xc7022058,
name=0xc70007c0 "platform", p_sd=p_sd@entry=0xc706bef8)
at fs/sysfs/dir.c:682
682 sd = sysfs_new_dirent(name, mode, SYSFS_DIR);
(gdb) s
sysfs_new_dirent (name=0xc70007c0 "platform", mode=mode@entry=16877,
type=type@entry=1) at fs/sysfs/dir.c:318
318 if (type & SYSFS_COPY_NAME) {
(gdb)
319 name = dup_name = kstrdup(name, GFP_KERNEL);
(gdb)
320 if (!name)
(gdb)
324 sd = kmem_cache_zalloc(sysfs_dir_cachep, GFP_KERNEL);
(gdb)
325 if (!sd)
(gdb)
328 if (sysfs_alloc_ino(&sd->s_ino))
(gdb)
331 atomic_set(&sd->s_count, 1);
(gdb)
332 atomic_set(&sd->s_active, 0);
(gdb)
334 sd->s_name = name;
(gdb)
335 sd->s_mode = mode;
(gdb)
336 sd->s_flags = type;
(gdb)
338 return sd;
(gdb)
345 }
(gdb)
create_dir (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>, parent_sd=0xc7022058,
name=<optimized out>, p_sd=p_sd@entry=0xc706bef8) at fs/sysfs/dir.c:683
683 if (!sd)
(gdb)
682 sd = sysfs_new_dirent(name, mode, SYSFS_DIR);
(gdb)
683 if (!sd)
(gdb)
685 sd->s_dir.kobj = kobj;
(gdb) n
688 sysfs_addrm_start(&acxt, parent_sd);
(gdb) s
sysfs_addrm_start (acxt=acxt@entry=0xc706bed0,
parent_sd=parent_sd@entry=0xc7022058) at fs/sysfs/dir.c:374
374 memset(acxt, 0, sizeof(*acxt));
(gdb) n
375 acxt->parent_sd = parent_sd;
(gdb)
381 mutex_lock(&sysfs_mutex);
(gdb)
383 inode = ilookup5(sysfs_sb, parent_sd->s_ino, sysfs_ilookup_test,
(gdb)
385 if (inode) {
(gdb)
401 }
(gdb) n
create_dir (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>, parent_sd=0xc7022058,
name=<optimized out>, p_sd=p_sd@entry=0xc706bef8) at fs/sysfs/dir.c:689
689 rc = sysfs_add_one(&acxt, sd);
(gdb) s
sysfs_add_one (acxt=acxt@entry=0xc706bed0, sd=sd@entry=0xc70222c0)
at fs/sysfs/dir.c:482
482 ret = __sysfs_add_one(acxt, sd);
(gdb) s
__sysfs_add_one (acxt=acxt@entry=0xc706bed0, sd=sd@entry=0xc70222c0)
at fs/sysfs/dir.c:425
425 if (sysfs_find_dirent(acxt->parent_sd, sd->s_name))
(gdb) s
sysfs_find_dirent (parent_sd=0xc7022058, name=0xc7000800 "platform")
at fs/sysfs/dir.c:639
639 for (sd = parent_sd->s_dir.children; sd; sd = sd->s_sibling)
(gdb) n
642 return NULL;
(gdb)
643 }
(gdb)
__sysfs_add_one (acxt=acxt@entry=0xc706bed0, sd=sd@entry=0xc70222c0)
at fs/sysfs/dir.c:428
428 sd->s_parent = sysfs_get(acxt->parent_sd);
(gdb) s
__sysfs_get (sd=0xc7022058) at fs/sysfs/sysfs.h:138
138 if (sd) {
(gdb) n
139 WARN_ON(!atomic_read(&sd->s_count));
(gdb)
140 atomic_inc(&sd->s_count);
(gdb)
__sysfs_add_one (acxt=acxt@entry=0xc706bed0, sd=sd@entry=0xc70222c0)
at fs/sysfs/dir.c:428
428 sd->s_parent = sysfs_get(acxt->parent_sd);
(gdb)
430 if (sysfs_type(sd) == SYSFS_DIR && acxt->parent_inode)
(gdb)
433 acxt->cnt++;
(gdb)
435 sysfs_link_sibling(sd);
(gdb) s
sysfs_link_sibling (sd=0xc70222c0) at fs/sysfs/dir.c:46
46 struct sysfs_dirent *parent_sd = sd->s_parent;
(gdb)
49 BUG_ON(sd->s_sibling);
(gdb)
55 for (pos = &parent_sd->s_dir.children; *pos; pos = &(*pos)->s_sibling) {
(gdb)
59 sd->s_sibling = *pos;
(gdb)
60 *pos = sd;
(gdb)
61 }
(gdb)
__sysfs_add_one (acxt=acxt@entry=0xc706bed0, sd=sd@entry=0xc70222c0)
at fs/sysfs/dir.c:437
437 return 0;
(gdb)
438 }
(gdb)
sysfs_add_one (acxt=acxt@entry=0xc706bed0, sd=sd@entry=0xc70222c0)
at fs/sysfs/dir.c:483
483 if (ret == -EEXIST) {
(gdb)
496 }
(gdb)
create_dir (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
parent_sd=<optimized out>, name=<optimized out>,
p_sd=p_sd@entry=0xc706bef8) at fs/sysfs/dir.c:690
690 sysfs_addrm_finish(&acxt);
(gdb) s
sysfs_addrm_finish (acxt=acxt@entry=0xc706bed0) at fs/sysfs/dir.c:595
595 mutex_unlock(&sysfs_mutex);
(gdb)
596 if (acxt->parent_inode) {
(gdb)
608 while (acxt->removed) {
(gdb)
619 }
(gdb)
create_dir (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
parent_sd=<optimized out>, name=<optimized out>,
p_sd=p_sd@entry=0xc706bef8) at fs/sysfs/dir.c:692
692 if (rc == 0)
(gdb)
693 *p_sd = sd;
(gdb)
698 }
(gdb)
sysfs_create_dir (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>)
at fs/sysfs/dir.c:723
723 if (!error)
(gdb)
724 kobj->sd = sd;
(gdb)
726 }
(gdb)
create_dir (kobj=0xc16c7b28 <platform_bus+8>) at lib/kobject.c:52
52 if (!error) {
(gdb)
kobject_add_internal (kobj=0xc16c7b28 <platform_bus+8>) at lib/kobject.c:187
187 error = create_dir(kobj);
(gdb)
204 kobj->state_in_sysfs = 1;
(gdb)
207 }
(gdb)
kobject_add (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>, parent=0x0,
fmt=fmt@entry=0x0) at lib/kobject.c:359
359 }
(gdb)
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:932
932 if (error)
(gdb)
936 if (platform_notify)
(gdb) p platform_notify
$11 = (int (*)(struct device *)) 0x0
(gdb) n
939 error = device_create_file(dev, &uevent_attr); # create file for new device
(gdb) s
device_create_file (attr=0xc16c78e4 <uevent_attr>,
dev=0xc16c7b20 <platform_bus>) at drivers/base/core.c:453
453 error = sysfs_create_file(&dev->kobj, &attr->attr);
(gdb) s
sysfs_create_file (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
attr=attr@entry=0xc16c78e4 <uevent_attr>) at fs/sysfs/file.c:539
539 BUG_ON(!kobj || !kobj->sd || !attr);
(gdb) n
541 return sysfs_add_file(kobj->sd, attr, SYSFS_KOBJ_ATTR);
(gdb) s
sysfs_add_file (type=2, attr=0xc16c78e4 <uevent_attr>, dir_sd=0xc70222c0)
at fs/sysfs/file.c:527
527 return sysfs_add_file_mode(dir_sd, attr, type, attr->mode);
(gdb) s
sysfs_add_file_mode (dir_sd=0xc70222c0,
attr=attr@entry=0xc16c78e4 <uevent_attr>, type=type@entry=2, amode=420)
at fs/sysfs/file.c:508
508 sd = sysfs_new_dirent(attr->name, mode, type);
(gdb)
503 umode_t mode = (amode & S_IALLUGO) | S_IFREG;
(gdb)
508 sd = sysfs_new_dirent(attr->name, mode, type);
(gdb)
509 if (!sd)
(gdb)
508 sd = sysfs_new_dirent(attr->name, mode, type);
(gdb)
509 if (!sd)
(gdb)
511 sd->s_attr.attr = (void *)attr;
(gdb)
513 sysfs_addrm_start(&acxt, dir_sd);
(gdb)
514 rc = sysfs_add_one(&acxt, sd);
(gdb)
515 sysfs_addrm_finish(&acxt);
(gdb)
517 if (rc)
(gdb)
521 }
(gdb)
sysfs_create_file (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
attr=attr@entry=0xc16c78e4 <uevent_attr>) at fs/sysfs/file.c:543
543 }
(gdb)
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:940
940 if (error)
(gdb)
943 if (MAJOR(dev->devt)) {
(gdb)
955 error = device_add_class_symlinks(dev); # add class symbol link for device
(gdb) s
device_add_class_symlinks (dev=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:689
689 if (!dev->class)
(gdb) n
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:958
958 error = device_add_attrs(dev);
(gdb) s
device_add_attrs (dev=0xc16c7b20 <platform_bus>) at drivers/base/core.c:386
386 struct device_type *type = dev->type;
(gdb) n
395 if (type) {
(gdb) p *dev
$12 = {parent = 0x0, p = 0xc7025420, kobj = {name = 0xc70007c0 "platform",
entry = {next = 0xc70851b0, prev = 0xc70851b0}, parent = 0xc70851dc,
kset = 0xc70851b0, ktype = 0xc16c7904 <device_ktype>, sd = 0xc70222c0,
kref = {refcount = {counter = 2}}, state_initialized = 1,
state_in_sysfs = 1, state_add_uevent_sent = 0,
state_remove_uevent_sent = 0, uevent_suppress = 0}, init_name = 0x0,
type = 0x0, sem = {lock = {raw_lock = {slock = 0}, magic = 3735899821,
owner_cpu = 4294967295, owner = 0xffffffff, dep_map = {
key = 0xc1f26d64 <__key.10624>, class_cache = 0x0,
name = 0xc15d4080 "semaphore->lock", cpu = 0, ip = 0}}, count = 1,
wait_list = {next = 0xc16c7b7c <platform_bus+92>,
prev = 0xc16c7b7c <platform_bus+92>}}, bus = 0x0, driver = 0x0,
platform_data = 0x0, power = {power_state = {event = 0}, can_wakeup = 0,
should_wakeup = 0, status = DPM_ON, entry = {next = 0x0, prev = 0x0}},
dma_mask = 0x0, coherent_dma_mask = 0, dma_parms = 0x0, dma_pools = {
next = 0xc16c7bb4 <platform_bus+148>,
prev = 0xc16c7bb4 <platform_bus+148>}, dma_mem = 0x0, archdata = {
acpi_handle = 0x0}, devt = 0, devres_lock = {raw_lock = {slock = 0},
magic = 3735899821, owner_cpu = 4294967295, owner = 0xffffffff, dep_map = {
key = 0xc1f26d6c <__key.15167>, class_cache = 0x0,
name = 0xc16207a9 "&dev->devres_lock", cpu = 0, ip = 0}}, devres_head = {
next = 0xc16c7bec <platform_bus+204>,
prev = 0xc16c7bec <platform_bus+204>}, knode_class = {n_klist = 0x0,
---Type <return> to continue, or q <return> to quit---
n_node = {next = 0x0, prev = 0x0}, n_ref = {refcount = {counter = 0}}},
class = 0x0, groups = 0x0, release = 0x0}
(gdb) n
401 error = device_add_groups(dev, dev->groups);
(gdb) s
device_add_groups (dev=dev@entry=0xc16c7b20 <platform_bus>, groups=0x0)
at drivers/base/core.c:359
359 if (groups) {
(gdb) n
371 }
(gdb)
device_add_attrs (dev=0xc16c7b20 <platform_bus>) at drivers/base/core.c:402
402 if (error)
(gdb)
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:961
961 error = bus_add_device(dev);
(gdb) s
bus_add_device (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/bus.c:468
468 struct bus_type *bus = bus_get(dev->bus);
(gdb) n
489 return 0;
(gdb)
500 }
(gdb)
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:962
962 if (error)
(gdb)
964 error = dpm_sysfs_add(dev); # addl dynamic power management directory
(gdb) s
dpm_sysfs_add (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/power/sysfs.c:92
92 return sysfs_create_group(&dev->kobj, &pm_attr_group);
(gdb) s
sysfs_create_group (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
grp=grp@entry=0xc16c7e00 <pm_attr_group>) at fs/sysfs/group.c:100
100 return internal_create_group(kobj, 0, grp);
(gdb) s
internal_create_group (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
update=update@entry=0, grp=grp@entry=0xc16c7e00 <pm_attr_group>)
at fs/sysfs/group.c:65
65 BUG_ON(!kobj || (!update && !kobj->sd));
(gdb) n
71 if (grp->name) {
(gdb) p grp
$13 = (const struct attribute_group *) 0xc16c7e00 <pm_attr_group>
(gdb) p *grp
$14 = {name = 0xc16218a4 "power", is_visible = 0x0,
attrs = 0xc16c7e0c <power_attrs>}
(gdb) n
72 error = sysfs_create_subdir(kobj, grp->name, &sd);
(gdb) s
sysfs_create_subdir (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
name=0xc16218a4 "power", p_sd=p_sd@entry=0xc706bf30) at fs/sysfs/dir.c:703
703 return create_dir(kobj, kobj->sd, name, p_sd);
(gdb) n
704 }
(gdb)
internal_create_group (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
update=update@entry=0, grp=grp@entry=0xc16c7e00 <pm_attr_group>)
at fs/sysfs/group.c:73
73 if (error)
(gdb)
77 sysfs_get(sd);
(gdb)
78 error = create_files(sd, kobj, grp, update);
(gdb) s
create_files (update=0, grp=0xc16c7e00 <pm_attr_group>,
kobj=0xc16c7b28 <platform_bus+8>, dir_sd=0xc7022370) at fs/sysfs/group.c:35
35 for (i = 0, attr = grp->attrs; *attr && !error; i++, attr++) {
(gdb) n
41 if (update)
(gdb)
43 if (grp->is_visible) {
(gdb)
36 mode_t mode = 0;
(gdb)
48 error = sysfs_add_file_mode(dir_sd, *attr, SYSFS_KOBJ_ATTR,
(gdb) s
sysfs_add_file_mode (dir_sd=dir_sd@entry=0xc7022370,
attr=0xc16c7e14 <dev_attr_wakeup>, type=type@entry=2, amode=420)
at fs/sysfs/file.c:502
502 {
(gdb) n
503 umode_t mode = (amode & S_IALLUGO) | S_IFREG;
(gdb)
508 sd = sysfs_new_dirent(attr->name, mode, type);
(gdb)
509 if (!sd)
(gdb)
511 sd->s_attr.attr = (void *)attr;
(gdb)
513 sysfs_addrm_start(&acxt, dir_sd);
(gdb)
514 rc = sysfs_add_one(&acxt, sd);
(gdb)
515 sysfs_addrm_finish(&acxt);
(gdb)
517 if (rc)
(gdb)
521 }
(gdb)
create_files (update=0, grp=0xc16c7e00 <pm_attr_group>,
kobj=0xc16c7b28 <platform_bus+8>, dir_sd=0xc7022370) at fs/sysfs/group.c:50
50 if (unlikely(error))
(gdb)
35 for (i = 0, attr = grp->attrs; *attr && !error; i++, attr++) {
(gdb)
internal_create_group (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
update=update@entry=0, grp=grp@entry=0xc16c7e00 <pm_attr_group>)
at fs/sysfs/group.c:78
78 error = create_files(sd, kobj, grp, update);
(gdb)
83 sysfs_put(sd);
(gdb)
85 }
(gdb)
sysfs_create_group (kobj=kobj@entry=0xc16c7b28 <platform_bus+8>,
grp=grp@entry=0xc16c7e00 <pm_attr_group>) at fs/sysfs/group.c:101
101 }
(gdb)
dpm_sysfs_add (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/power/sysfs.c:93
93 }
(gdb)
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:999
999 bus_remove_device(dev);
(gdb)
965 if (error)
(gdb)
967 device_pm_add(dev); # add device to dpm list
(gdb) s
device_pm_add (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/power/main.c:84
84 pr_debug("PM: Adding info for %s:%s\n",
(gdb) n
83 {
(gdb)
84 pr_debug("PM: Adding info for %s:%s\n",
(gdb)
87 mutex_lock(&dpm_list_mtx);
(gdb)
88 if (dev->parent) {
(gdb)
92 } else if (transition_started) {
(gdb) p transition_started
$15 = false
(gdb) n
101 list_add_tail(&dev->power.entry, &dpm_list);
(gdb)
102 mutex_unlock(&dpm_list_mtx);
(gdb)
103 }
(gdb)
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:972
972 if (dev->bus)
976 kobject_uevent(&dev->kobj, KOBJ_ADD); # notify userspace by ending an uevent
(gdb)
977 bus_probe_device(dev); # probe drivers for a new device
(gdb) s
bus_probe_device (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/bus.c:510
510 struct bus_type *bus = dev->bus;
(gdb) n
513 if (bus && bus->p->drivers_autoprobe) {
(gdb)
517 }
(gdb)
device_add (dev=<optimized out>, dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:978
978 if (parent)
(gdb)
982 if (dev->class) {
(gdb)
996 put_device(dev);
(gdb)
1023 }
(gdb)
device_register (dev=dev@entry=0xc16c7b20 <platform_bus>)
at drivers/base/core.c:1044
1044 }
(gdb)
platform_bus_init () at drivers/base/platform.c:962
962 if (error)
(gdb)
964 error = bus_register(&platform_bus_type);
(gdb) s
bus_register (bus=bus@entry=0xc16c7aa0 <platform_bus_type>)
at drivers/base/bus.c:884
884 priv = kzalloc(sizeof(struct bus_type_private), GFP_KERNEL);
(gdb)
885 if (!priv)
(gdb)
888 priv->bus = bus;
(gdb)
889 bus->p = priv;
(gdb)
891 BLOCKING_INIT_NOTIFIER_HEAD(&priv->bus_notifier);
(gdb)
893 retval = kobject_set_name(&priv->subsys.kobj, "%s", bus->name);
(gdb)
894 if (retval)
(gdb)
897 priv->subsys.kobj.kset = bus_kset;
(gdb)
898 priv->subsys.kobj.ktype = &bus_ktype;
(gdb)
899 priv->drivers_autoprobe = 1;
(gdb)
901 retval = kset_register(&priv->subsys); # initialize and add a kset
(gdb)
902 if (retval)
(gdb)
905 retval = bus_create_file(bus, &bus_attr_uevent);
(gdb) s
bus_create_file (bus=bus@entry=0xc16c7aa0 <platform_bus_type>, # create file for bus
attr=attr@entry=0xc16c7998 <bus_attr_uevent>) at drivers/base/bus.c:126
126 if (bus_get(bus)) {
(gdb) n
124 {
(gdb)
126 if (bus_get(bus)) {
(gdb)
127 error = sysfs_create_file(&bus->p->subsys.kobj, &attr->attr);
(gdb)
128 bus_put(bus);
(gdb)
132 }
(gdb)
bus_register (bus=bus@entry=0xc16c7aa0 <platform_bus_type>)
at drivers/base/bus.c:906
906 if (retval)
(gdb)
909 priv->devices_kset = kset_create_and_add("devices", NULL, # create kset for devices and drivers associated with bus
(gdb)
911 if (!priv->devices_kset) {
(gdb)
916 priv->drivers_kset = kset_create_and_add("drivers", NULL,
(gdb)
918 if (!priv->drivers_kset) {
(gdb)
923 klist_init(&priv->klist_devices, klist_devices_get, klist_devices_put); # initlialize klist which used to iterate over the @devices_kset or @drivers_kset
(gdb)
924 klist_init(&priv->klist_drivers, NULL, NULL);
(gdb)
926 retval = add_probe_files(bus);
(gdb) s
add_probe_files (bus=0xc16c7aa0 <platform_bus_type>) at drivers/base/bus.c:609
609 retval = bus_create_file(bus, &bus_attr_drivers_probe);
(gdb) s
bus_create_file (bus=bus@entry=0xc16c7aa0 <platform_bus_type>,
attr=attr@entry=0xc16c79d4 <bus_attr_drivers_probe>)
at drivers/base/bus.c:126
126 if (bus_get(bus)) {
(gdb) n
127 error = sysfs_create_file(&bus->p->subsys.kobj, &attr->attr);
(gdb)
128 bus_put(bus);
(gdb)
132 }
(gdb)
add_probe_files (bus=0xc16c7aa0 <platform_bus_type>) at drivers/base/bus.c:610
610 if (retval)
(gdb)
613 retval = bus_create_file(bus, &bus_attr_drivers_autoprobe);
(gdb)
614 if (retval)
(gdb)
bus_register (bus=bus@entry=0xc16c7aa0 <platform_bus_type>)
at drivers/base/bus.c:930
930 retval = bus_add_attrs(bus);
(gdb) s
bus_add_attrs (bus=0xc16c7aa0 <platform_bus_type>) at drivers/base/bus.c:819
819 if (bus->bus_attrs) {
(gdb) n
bus_register (bus=bus@entry=0xc16c7aa0 <platform_bus_type>)
at drivers/base/bus.c:935
935 return 0;
(gdb)
934 pr_debug("bus: '%s': registered\n", bus->name);
(gdb)
951 }
(gdb)
platform_bus_init () at drivers/base/platform.c:965
965 if (error)
(gdb)
968 }
(gdb)
system_bus_init create kset for system bus
cpu_dev_init initialize cpu
35 cpu_dev_init();
(gdb) s
cpu_dev_init () at drivers/base/cpu.c:240
240 err = sysdev_class_register(&cpu_sysdev_class);
(gdb) s
sysdev_class_register (cls=cls@entry=0xc16c7cc0 <cpu_sysdev_class>)
at drivers/base/sys.c:136
136 pr_debug("Registering sysdev class '%s'\n", cls->name);
(gdb) n
138 INIT_LIST_HEAD(&cls->drivers);
(gdb)
139 memset(&cls->kset.kobj, 0x00, sizeof(struct kobject));
(gdb)
140 cls->kset.kobj.parent = &system_kset->kobj;
(gdb)
141 cls->kset.kobj.ktype = &ktype_sysdev_class;
(gdb)
142 cls->kset.kobj.kset = system_kset;
(gdb)
140 cls->kset.kobj.parent = &system_kset->kobj;
(gdb)
144 retval = kobject_set_name(&cls->kset.kobj, "%s", cls->name);
(gdb)
145 if (retval)
(gdb)
148 return kset_register(&cls->kset);
(gdb) s
kset_register (k=k@entry=0xc16c7cd8 <cpu_sysdev_class+24>) at lib/kobject.c:716
716 if (!k)
(gdb) n
719 kset_init(k);
(gdb)
720 err = kobject_add_internal(&k->kobj);
(gdb)
721 if (err)
(gdb)
723 kobject_uevent(&k->kobj, KOBJ_ADD);
(gdb)
725 }
(gdb)
sysdev_class_register (cls=cls@entry=0xc16c7cc0 <cpu_sysdev_class>)
at drivers/base/sys.c:149
149 }
(gdb)
cpu_dev_init () at drivers/base/cpu.c:241
241 if (!err)
(gdb)
242 err = cpu_states_init();
(gdb) s
cpu_states_init () at drivers/base/cpu.c:188
188 ret = sysdev_class_create_file(&cpu_sysdev_class,
(gdb) s
sysdev_class_create_file (c=c@entry=0xc16c7cc0 <cpu_sysdev_class>,
a=a@entry=0xc16c7c70 <attr_online_map>) at drivers/base/sys.c:119
119 return sysfs_create_file(&c->kset.kobj, &a->attr);
(gdb) s
sysfs_create_file (kobj=kobj@entry=0xc16c7d04 <cpu_sysdev_class+68>,
attr=attr@entry=0xc16c7c70 <attr_online_map>) at fs/sysfs/file.c:539
539 BUG_ON(!kobj || !kobj->sd || !attr);
(gdb) n
541 return sysfs_add_file(kobj->sd, attr, SYSFS_KOBJ_ATTR);
(gdb) s
sysfs_add_file (type=2, attr=0xc16c7c70 <attr_online_map>, dir_sd=0xc7022688)
at fs/sysfs/file.c:527
527 return sysfs_add_file_mode(dir_sd, attr, type, attr->mode);
(gdb) s
sysfs_add_file_mode (dir_sd=0xc7022688,
attr=attr@entry=0xc16c7c70 <attr_online_map>, type=type@entry=2, amode=292)
at fs/sysfs/file.c:503
503 umode_t mode = (amode & S_IALLUGO) | S_IFREG;
(gdb)
508 sd = sysfs_new_dirent(attr->name, mode, type);
(gdb)
509 if (!sd)
(gdb)
511 sd->s_attr.attr = (void *)attr;
(gdb)
513 sysfs_addrm_start(&acxt, dir_sd);
(gdb)
514 rc = sysfs_add_one(&acxt, sd);
(gdb)
515 sysfs_addrm_finish(&acxt);
(gdb)
517 if (rc)
(gdb)
521 }
(gdb)
sysfs_create_file (kobj=kobj@entry=0xc16c7d04 <cpu_sysdev_class+68>,
attr=attr@entry=0xc16c7c70 <attr_online_map>) at fs/sysfs/file.c:543
543 }
(gdb)
sysdev_class_create_file (c=c@entry=0xc16c7cc0 <cpu_sysdev_class>,
a=a@entry=0xc16c7c70 <attr_online_map>) at drivers/base/sys.c:120
120 }
(gdb)
cpu_states_init () at drivers/base/cpu.c:190
190 if (!err)
(gdb)
186 for (i = 0; i < ARRAY_SIZE(cpu_state_attr); i++) {
(gdb)
cpu_dev_init () at drivers/base/cpu.c:245
245 if (!err)
(gdb)
246 err = sched_create_sysfs_power_savings_entries(&cpu_sysdev_class);
(gdb) s
sched_create_sysfs_power_savings_entries (
cls=cls@entry=0xc16c7cc0 <cpu_sysdev_class>) at kernel/sched.c:9437
9437 if (smt_capable())
(gdb) n
9442 if (!err && mc_capable())
(gdb)
9433 {
(gdb)
9447 }
(gdb)
cpu_dev_init () at drivers/base/cpu.c:250
250 }
(gdb)
driver_init () at drivers/base/init.c:37
37 }
(gdb)
memory_dev_initinitialize the sysfs for memory devices
driver_init is complete, let's continue other routines invoked in do_basic_setup, next in do_basic_setup is init_irq_proc which is used to create folder under /proc for irq stuff.
788 init_irq_proc();
(gdb) s
init_irq_proc () at kernel/irq/proc.c:256
256 root_irq_dir = proc_mkdir("irq", NULL);
(gdb) s
proc_mkdir (name=name@entry=0xc1621630 "irq", parent=parent@entry=0x0)
at fs/proc/generic.c:701
701 return proc_mkdir_mode(name, S_IRUGO | S_IXUGO, parent);
(gdb) s
proc_mkdir_mode (name=name@entry=0xc1621630 "irq", mode=mode@entry=365,
parent=parent@entry=0x0) at fs/proc/generic.c:671
671 ent = __proc_create(&parent, name, S_IFDIR | mode, 2);
(gdb) s
__proc_create (parent=parent@entry=0xc706bfa0,
name=name@entry=0xc1621630 "irq", mode=16749, nlink=nlink@entry=2) # create directory under proc for irq
at fs/proc/generic.c:608
608 const char *fn = name;
(gdb)
612 if (!name || !strlen(name)) goto out;
(gdb)
614 if (xlate_proc_name(name, parent, &fn) != 0) # parse the name of to be created directory
(gdb) p name
$1 = 0xc1621630 "irq"
(gdb) p parent
$2 = (struct proc_dir_entry **) 0xc706bfa0
(gdb) p *parent
$3 = (struct proc_dir_entry *) 0x0
(gdb) p fn
$4 = 0xc1621630 "irq"
(gdb) n
618 if (strchr(fn, '/'))
(gdb) p fn
$5 = 0xc1621630 "irq"
(gdb) n
621 len = strlen(fn);
(gdb)
623 ent = kmalloc(sizeof(struct proc_dir_entry) + len + 1, GFP_KERNEL);
(gdb)
624 if (!ent) goto out;
(gdb)
626 memset(ent, 0, sizeof(struct proc_dir_entry));
(gdb)
627 memcpy(((char *) ent) + sizeof(struct proc_dir_entry), fn, len + 1);
(gdb)
628 ent->name = ((char *) ent) + sizeof(*ent);
(gdb)
629 ent->namelen = len;
(gdb)
630 ent->mode = mode;
(gdb)
631 ent->nlink = nlink;
(gdb)
632 atomic_set(&ent->count, 1);
(gdb)
633 ent->pde_users = 0;
(gdb)
634 spin_lock_init(&ent->pde_unload_lock);
(gdb)
635 ent->pde_unload_completion = NULL;
(gdb)
636 INIT_LIST_HEAD(&ent->pde_openers);
(gdb)
639 }
(gdb)
proc_mkdir_mode (name=name@entry=0xc1621630 "irq", mode=mode@entry=365,
parent=0xc16abe20 <proc_root>, parent@entry=0x0) at fs/proc/generic.c:672
672 if (ent) {
(gdb)
673 if (proc_register(parent, ent) < 0) {
(gdb) s
proc_register (dir=0xc16abe20 <proc_root>, dp=dp@entry=0xc70254d0)
at fs/proc/generic.c:560
560 {
(gdb) n
564 i = get_inode_number();
(gdb)
569 if (S_ISDIR(dp->mode)) {
(gdb)
570 if (dp->proc_iops == NULL) {
(gdb)
571 dp->proc_fops = &proc_dir_operations;
(gdb)
572 dp->proc_iops = &proc_dir_inode_operations;
(gdb)
574 dir->nlink++;
(gdb)
585 spin_lock(&proc_subdir_lock);
(gdb)
587 for (tmp = dir->subdir; tmp; tmp = tmp->next)
(gdb)
588 if (strcmp(tmp->name, dp->name) == 0) {
(gdb)
594 dp->next = dir->subdir;
(gdb)
597 spin_unlock(&proc_subdir_lock);
(gdb)
595 dp->parent = dir;
(gdb)
596 dir->subdir = dp;
(gdb)
597 spin_unlock(&proc_subdir_lock);
(gdb)
599 return 0;
(gdb)
600 }
(gdb)
proc_mkdir_mode (name=name@entry=0xc1621630 "irq", mode=mode@entry=365,
parent=0xc16abe20 <proc_root>, parent@entry=0x0) at fs/proc/generic.c:679
679 }
(gdb)
proc_mkdir (name=name@entry=0xc1621630 "irq", parent=parent@entry=0x0)
at fs/proc/generic.c:702
702 }
(gdb)
init_irq_proc () at kernel/irq/proc.c:257
257 if (!root_irq_dir)
(gdb)
260 register_default_affinity_proc();
(gdb) s
register_default_affinity_proc () at kernel/irq/proc.c:260
260 register_default_affinity_proc();
(gdb) s
proc_create (proc_fops=0xc149ca20 <default_affinity_proc_fops>, parent=0x0,
mode=384, name=0xc15dafe3 "irq/default_smp_affinity")
at include/linux/proc_fs.h:155
155 return proc_create_data(name, mode, parent, proc_fops, NULL);
(gdb) s
proc_create_data (name=name@entry=0xc15dafe3 "irq/default_smp_affinity",
mode=mode@entry=384, parent=parent@entry=0x0,
proc_fops=proc_fops@entry=0xc149ca20 <default_affinity_proc_fops>,
data=data@entry=0x0) at fs/proc/generic.c:740
740 if (S_ISDIR(mode)) {
(gdb) n
746 mode |= S_IFREG;
(gdb)
747 if ((mode & S_IALLUGO) == 0)
(gdb)
752 pde = __proc_create(&parent, name, mode, nlink); # create proc_dir_entry for default_smp_affinity, it's the subdirectory of irq
(gdb)
753 if (!pde)
(gdb)
755 pde->proc_fops = proc_fops;
(gdb)
756 pde->data = data;
(gdb)
757 if (proc_register(parent, pde) < 0)
(gdb)
761 kfree(pde);
(gdb)
764 }
(gdb)
init_irq_proc () at kernel/irq/proc.c:265
265 for_each_irq_desc(irq, desc) { # create other directories
(gdb)
269 register_irq_proc(irq, desc);
(gdb) p desc
$6 = (struct irq_desc *) 0xc1690b60 <irq_desc_legacy>
(gdb) p irq
$7 = 0
(gdb) n
265 for_each_irq_desc(irq, desc) {
(gdb)
269 register_irq_proc(irq, desc);
(gdb) p desc
$8 = (struct irq_desc *) 0xc1690c00 <irq_desc_legacy+160>
(gdb) p irq
$9 = 1
(gdb) n
265 for_each_irq_desc(irq, desc) {
(gdb)
269 register_irq_proc(irq, desc);
(gdb) p desc
$10 = (struct irq_desc *) 0xc1690ca0 <irq_desc_legacy+320>
(gdb) p *desc
$11 = {irq = 2, kstat_irqs = 0xc7001008,
handle_irq = 0xc10aaeb0 <handle_level_irq>, chip = 0xc1694820 <i8259A_chip>,
msi_desc = 0x0, handler_data = 0x0, chip_data = 0xc16971a8 <irq_cfgx+40>,
action = 0x0, status = 512, depth = 1, wake_depth = 0, irq_count = 0,
last_unhandled = 0, irqs_unhandled = 0, lock = {raw_lock = {slock = 514},
magic = 3735899821, owner_cpu = 4294967295, owner = 0xffffffff, dep_map = {
key = 0xc1e3376c <irq_desc_lock_class>, class_cache = 0x0,
name = 0xc15dae6f "&irq_desc_lock_class", cpu = 0, ip = 3238703088}},
affinity = {{bits = {255}}}, node = 0, pending_mask = {{bits = {0}}},
threads_active = {counter = 0}, wait_for_threads = {lock = {raw_lock = {
slock = 0}, magic = 0, owner_cpu = 0, owner = 0x0, dep_map = {
key = 0x0, class_cache = 0x0, name = 0x0, cpu = 0, ip = 0}},
task_list = {next = 0x0, prev = 0x0}}, dir = 0x0, name = 0xc162693c "XT"}
(gdb) n
265 for_each_irq_desc(irq, desc) {
(gdb)
269 register_irq_proc(irq, desc);
(gdb) p *desc
$12 = {irq = 3, kstat_irqs = 0xc700100c,
handle_irq = 0xc10ab4a0 <handle_edge_irq>, chip = 0xc16f7140 <ioapic_chip>,
msi_desc = 0x0, handler_data = 0x0, chip_data = 0xc16971bc <irq_cfgx+60>,
action = 0x0, status = 512, depth = 1, wake_depth = 0, irq_count = 0,
last_unhandled = 0, irqs_unhandled = 0, lock = {raw_lock = {slock = 1028},
magic = 3735899821, owner_cpu = 4294967295, owner = 0xffffffff, dep_map = {
key = 0xc1e3376c <irq_desc_lock_class>, class_cache = 0x0,
name = 0xc15dae6f "&irq_desc_lock_class", cpu = 0, ip = 3238703088}},
affinity = {{bits = {1}}}, node = 0, pending_mask = {{bits = {0}}},
threads_active = {counter = 0}, wait_for_threads = {lock = {raw_lock = {
slock = 0}, magic = 0, owner_cpu = 0, owner = 0x0, dep_map = {
key = 0x0, class_cache = 0x0, name = 0x0, cpu = 0, ip = 0}},
task_list = {next = 0x0, prev = 0x0}}, dir = 0x0, name = 0xc15d11ec "edge"}
(gdb) n
265 for_each_irq_desc(irq, desc) {
(gdb)
269 register_irq_proc(irq, desc);
(gdb) p *desc
$13 = {irq = 4, kstat_irqs = 0xc7001010,
handle_irq = 0xc10ab4a0 <handle_edge_irq>, chip = 0xc16f7140 <ioapic_chip>,
msi_desc = 0x0, handler_data = 0x0, chip_data = 0xc16971d0 <irq_cfgx+80>,
action = 0x0, status = 512, depth = 1, wake_depth = 0, irq_count = 0,
last_unhandled = 0, irqs_unhandled = 0, lock = {raw_lock = {slock = 1028},
magic = 3735899821, owner_cpu = 4294967295, owner = 0xffffffff, dep_map = {
key = 0xc1e3376c <irq_desc_lock_class>, class_cache = 0x0,
name = 0xc15dae6f "&irq_desc_lock_class", cpu = 0, ip = 3238703088}},
affinity = {{bits = {1}}}, node = 0, pending_mask = {{bits = {0}}},
threads_active = {counter = 0}, wait_for_threads = {lock = {raw_lock = {
slock = 0}, magic = 0, owner_cpu = 0, owner = 0x0, dep_map = {
key = 0x0, class_cache = 0x0, name = 0x0, cpu = 0, ip = 0}},
task_list = {next = 0x0, prev = 0x0}}, dir = 0x0, name = 0xc15d11ec "edge"}
(gdb) n
......
271 }
call all constructor functions linked into the kernel with routine do_ctors, the defination of __ctors_start and __ctors_end could be found in vmlinux.lds.S
#ifdef CONFIG_CONSTRUCTORS
#define KERNEL_CTORS() . = ALIGN(8); \
VMLINUX_SYMBOL(__ctors_start) = .; \
*(.ctors) \
VMLINUX_SYMBOL(__ctors_end) = .;
#else
#define KERNEL_CTORS()
#endif
do_ctors () at init/main.c:703
703 for (; call < (ctor_fn_t *) __ctors_end; call++)
(gdb) l
698 static void __init do_ctors(void)
699 {
700 #ifdef CONFIG_CONSTRUCTORS
701 ctor_fn_t *call = (ctor_fn_t *) __ctors_start;
702
703 for (; call < (ctor_fn_t *) __ctors_end; call++)
704 (*call)();
705 #endif
706 }
707
(gdb) p __ctors_start
$1 = 0xc17611f0 <__setup_str_rdinit_setup> "rdinit="
(gdb) p __ctors_end
$2 = 0xc17611f0 <__setup_str_rdinit_setup> "rdinit="
(gdb) n
the last funtion in do_basic_setup is do_initcalls which call all init functions defined in vmlinux.lsd.S shown as follow:
#define INITCALLS \
*(.initcallearly.init) \
VMLINUX_SYMBOL(__early_initcall_end) = .; \
*(.initcall0.init) \
*(.initcall0s.init) \
*(.initcall1.init) \
*(.initcall1s.init) \
*(.initcall2.init) \
*(.initcall2s.init) \
*(.initcall3.init) \
*(.initcall3s.init) \
*(.initcall4.init) \
*(.initcall4s.init) \
*(.initcall5.init) \
*(.initcall5s.init) \
*(.initcallrootfs.init) \
*(.initcall6.init) \
*(.initcall6s.init) \
*(.initcall7.init) \
*(.initcall7s.init)
#define INIT_CALLS \
VMLINUX_SYMBOL(__initcall_start) = .; \
INITCALLS \
VMLINUX_SYMBOL(__initcall_end) = .;
Invokes functions located in memory region between symbol __early_initcall_end and __initcall_end.
790 do_initcalls();
(gdb) s
do_initcalls () at init/main.c:767
767 for (call = __early_initcall_end; call < __initcall_end; call++)
(gdb) where
#0 do_initcalls () at init/main.c:767
#1 do_basic_setup () at init/main.c:790
#2 kernel_init (unused=<optimized out>) at init/main.c:890
#3 0xc1003ad7 in kernel_thread_helper ()
at arch/x86/kernel/entry_32.S:1000
(gdb) p __early_initcall_end
$1 = 0xc177d9b4 <__initcall_init_mmap_min_addr0>
(gdb) p __initcall_end
$2 = 0xc177e008 <__initcall_con_init>
(gdb)
let's check functions invoked in do_initcalls:
init_mmap_min_addr initializes mmap_min_addr which represents amount of vm to protect from userspace access by both DAC and the LSM, update mmap_min_addr = max(dac_mmap_min_addr, CONFIG_LSM_MMAP_MIN_ADDR)
(gdb) n
768 do_one_initcall(*call);
(gdb) s
do_one_initcall (fn=0xc1718601 <init_mmap_min_addr>) at init/main.c:717
717 int count = preempt_count();
(gdb)
720 if (initcall_debug) {
(gdb)
728 ret.result = fn();
(gdb) s
init_mmap_min_addr () at security/min_addr.c:48
48 update_mmap_min_addr();
(gdb) s
update_mmap_min_addr () at security/min_addr.c:19
19 mmap_min_addr = dac_mmap_min_addr;
(gdb) n
init_mmap_min_addr () at security/min_addr.c:51
51 }
(gdb) p dac_mmap_min_addr
$3 = 65536
(gdb) n
do_one_initcall (fn=0xc1718601 <init_mmap_min_addr>) at init/main.c:730
730 if (initcall_debug) {
(gdb)
740 msgbuf[0] = 0;
(gdb)
742 if (ret.result && ret.result != -ENODEV && initcall_debug)
(gdb)
745 if (preempt_count() != count) {
(gdb)
749 if (irqs_disabled()) {
(gdb)
753 if (msgbuf[0]) {
(gdb)
758 }
(gdb)
init_cpufreq_transition_notifier_list initializes an SRCU notifier head with cpufreq_transition_notifier_list as input parameter and set the flag which indicates if init_cpufreq_transition_notifier_list is called as true. The "transition" list for kernel code that needs to handle changes to devices when the CPU clock speed changes.
init_cpufreq_transition_notifier_list () at drivers/cpufreq/cpufreq.c:127
127 srcu_init_notifier_head(&cpufreq_transition_notifier_list);
(gdb) s
srcu_init_notifier_head (
nh=nh@entry=0xc1f29d00 <cpufreq_transition_notifier_list>)
at kernel/notifier.c:521
521 mutex_init(&nh->mutex);
(gdb) n
522 if (init_srcu_struct(&nh->srcu) < 0)
(gdb) s
init_srcu_struct (
sp=sp@entry=0xc1f29d50 <cpufreq_transition_notifier_list+80>)
at kernel/srcu.c:47
47 sp->completed = 0;
(gdb)
48 mutex_init(&sp->mutex);
(gdb)
49 sp->per_cpu_ref = alloc_percpu(struct srcu_struct_array);
(gdb)
50 return (sp->per_cpu_ref ? 0 : -ENOMEM);
(gdb)
51 }
(gdb)
srcu_init_notifier_head (
nh=nh@entry=0xc1f29d00 <cpufreq_transition_notifier_list>)
at kernel/notifier.c:524
524 nh->head = NULL;
(gdb)
525 }
(gdb)
init_cpufreq_transition_notifier_list () at drivers/cpufreq/cpufreq.c:128
128 init_cpufreq_transition_notifier_list_called = true;
(gdb)
130 }
(gdb)
net_ns_init is used to initialize network namespace as the function name imply
net_ns_init () at net/core/net_namespace.c:240
240 ng = net_alloc_generic(); # allocate memmory for struct net_generic
(gdb)
244 rcu_assign_pointer(init_net.gen, ng);
(gdb)
246 mutex_lock(&net_mutex);
(gdb)
247 if (setup_net(&init_net))
(gdb) s
setup_net (net=0xc1f2a8c0 <init_net>) at net/core/net_namespace.c:39
39 atomic_set(&net->count, 1);
(gdb)
45 list_for_each_entry(ops, &pernet_list, list) { # execute init functions of all opses registered in pernet_list
(gdb)
net_ns_init () at net/core/net_namespace.c:250
250 rtnl_lock();
(gdb)
(gdb)
net_ns_init () at net/core/net_namespace.c:251
251 list_add_tail_rcu(&init_net.list, &net_namespace_list); # add net namespace list
(gdb)
252 rtnl_unlock();
(gdb)
254 mutex_unlock(&net_mutex);
(gdb)
257 }
e820_mark_nvs_memory, non-volatile storage is a type of computer memory that can retrieve stored information even after having been power cycled (turned off and back on), e820_mark_nvs_memory marks ACPI NVS memory region, so that we can save/restore it during hibernation and the subsequent resume, it loops e820 memory map, if type of a specific memory region is E820_NVS, registers platform NVS memory region to save
e820_mark_nvs_memory () at arch/x86/kernel/e820.c:712
712 for (i = 0; i < e820.nr_map; i++) {
(gdb) p e820.nr_map
$6 = 8
(gdb) n
709 {
(gdb)
712 for (i = 0; i < e820.nr_map; i++) {
(gdb)
715 if (ei->type == E820_NVS)
......
712 for (i = 0; i < e820.nr_map; i++) {
(gdb)
720 }
(gdb)
cpufreq_tsc, it checks if cpu has tsc feature and if TSC ticks at a constant rate in boot cpu, registers norifier block for cpu clock rate changes.
cpufreq_tsc () at arch/x86/kernel/tsc.c:730
730 if (!cpu_has_tsc)
(gdb) n
732 if (boot_cpu_has(X86_FEATURE_CONSTANT_TSC))
(gdb)
734 cpufreq_register_notifier(&time_cpufreq_notifier_block,
(gdb) s
cpufreq_register_notifier (
nb=nb@entry=0xc1694e80 <time_cpufreq_notifier_block>,
list=list@entry=0) at drivers/cpufreq/cpufreq.c:1457
1457 WARN_ON(!init_cpufreq_transition_notifier_list_called);
(gdb) n
1459 switch (list) {
(gdb)
1461 ret = srcu_notifier_chain_register(
(gdb) s
srcu_notifier_chain_register (
nh=nh@entry=0xc1f29d00 <cpufreq_transition_notifier_list>,
n=n@entry=0xc1694e80 <time_cpufreq_notifier_block>)
at kernel/notifier.c:427
427 if (unlikely(system_state == SYSTEM_BOOTING))
(gdb)
428 return notifier_chain_register(&nh->head, n);
(gdb)
434 }
(gdb)
cpufreq_register_notifier (
nb=nb@entry=0xc1694e80 <time_cpufreq_notifier_block>,
list=list@entry=0) at drivers/cpufreq/cpufreq.c:1473
1473 }
(gdb)
cpufreq_tsc () at arch/x86/kernel/tsc.c:737
737 }
(gdb)
pci_reboot_init, it walks through table pci_reboot_dmi_table, check if dmi of current system match one of the items in the table, if yes, set reboot type and reboot.
pci_reboot_init () at arch/x86/kernel/reboot.c:485
485 dmi_check_system(pci_reboot_dmi_table);
(gdb) p pci_reboot_dmi_table
$3 = {{callback = 0xc17050f1 <set_pci_reboot>,
ident = 0xc15d1479 "Apple MacBook5", matches = {{slot = 4 '\004',
substr = "Apple Inc.", '\000' <repeats 68 times>}, {
slot = 5 '\005', substr = "MacBook5", '\000' <repeats 70 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}},
driver_data = 0x0}, {callback = 0xc17050f1 <set_pci_reboot>,
ident = 0xc15d1488 "Apple MacBookPro5", matches = {{slot = 4 '\004',
substr = "Apple Inc.", '\000' <repeats 68 times>}, {
slot = 5 '\005',
substr = "MacBookPro5", '\000' <repeats 67 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}},
driver_data = 0x0}, {callback = 0xc17050f1 <set_pci_reboot>,
ident = 0xc15d149a "Apple Macmini3,1", matches = {{slot = 4 '\004',
substr = "Apple Inc.", '\000' <repeats 68 times>}, {
slot = 5 '\005',
substr = "Macmini3,1", '\000' <repeats 68 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}},
driver_data = 0x0}, {callback = 0xc17050f1 <set_pci_reboot>,
ident = 0xc15d14ab "Apple iMac9,1", matches = {{slot = 4 '\004',
substr = "Apple Inc.", '\000' <repeats 68 times>}, {
---Type <return> to continue, or q <return> to quit---
slot = 5 '\005', substr = "iMac9,1", '\000' <repeats 71 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}},
driver_data = 0x0}, {callback = 0xc17050f1 <set_pci_reboot>,
ident = 0xc15d14b9 "Dell Latitude E5420", matches = {{slot = 4 '\004',
substr = "Dell Inc.", '\000' <repeats 69 times>}, {
slot = 5 '\005',
substr = "Latitude E5420", '\000' <repeats 64 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}},
driver_data = 0x0}, {callback = 0x0, ident = 0x0, matches = {{
slot = 0 '\000', substr = '\000' <repeats 78 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}, {
slot = 0 '\000', substr = '\000' <repeats 78 times>}},
driver_data = 0x0}}
(gdb) s
dmi_check_system (list=list@entry=0xc1739de0 <pci_reboot_dmi_table>)
at drivers/firmware/dmi_scan.c:469
469 int count = 0;
(gdb) n
472 for (d = list; !dmi_is_end_of_table(d); d++)
(gdb)
473 if (dmi_matches(d)) {
(gdb) s
dmi_matches (dmi=dmi@entry=0xc1739de0 <pci_reboot_dmi_table>)
at drivers/firmware/dmi_scan.c:427
427 {
(gdb) n
430 WARN(!dmi_initialized, KERN_ERR "dmi check: not initialized yet.\n");
(gdb)
433 int s = dmi->matches[i].slot;
(gdb)
434 if (s == DMI_NONE)
(gdb) p s
$4 = 4
(gdb) n
436 if (dmi_ident[s]
(gdb) p dmi_ident
$5 = {0x0, 0xc1f43000 <.brk.pagetables+8192> "Bochs",
0xc1f43008 <.brk.pagetables+8200> "Bochs",
0xc1f43010 <.brk.pagetables+8208> "01/01/2011",
0xc1f4301c <.brk.pagetables+8220> "QEMU",
0xc1f43024 <.brk.pagetables+8228> "Standard PC (i440FX + PIIX, 1996)",
0xc1f43048 <.brk.pagetables+8264> "pc-i440fx-trusty",
0xc1f4305c <.brk.pagetables+8284> "", 0x0, 0x0, 0x0, 0x0, 0x0, 0x0,
0xc1f43060 <.brk.pagetables+8288> "Bochs",
0xc1f43068 <.brk.pagetables+8296> "1",
0xc1f4306c <.brk.pagetables+8300> "",
0xc1f43070 <.brk.pagetables+8304> "",
0xc1f43074 <.brk.pagetables+8308> ""}
(gdb) n
437 && strstr(dmi_ident[s], dmi->matches[i].substr))
(gdb) p dmi->matches[i].substr
$6 = "Apple Inc.", '\000' <repeats 68 times>
(gdb) n
443 }
(gdb)
440 return false;
(gdb)
443 }
(gdb)
dmi_check_system (list=list@entry=0xc1739de0 <pci_reboot_dmi_table>)
at drivers/firmware/dmi_scan.c:472
472 for (d = list; !dmi_is_end_of_table(d); d++)
(gdb)
473 if (dmi_matches(d)) {
(gdb)
472 for (d = list; !dmi_is_end_of_table(d); d++)
(gdb)
473 if (dmi_matches(d)) {
(gdb)
472 for (d = list; !dmi_is_end_of_table(d); d++)
(gdb)
473 if (dmi_matches(d)) {
(gdb)
472 for (d = list; !dmi_is_end_of_table(d); d++)
(gdb)
473 if (dmi_matches(d)) {
(gdb)
472 for (d = list; !dmi_is_end_of_table(d); d++)
(gdb)
480 }
(gdb)
pci_reboot_init () at arch/x86/kernel/reboot.c:487
487 }
(gdb)
reboot_init, the procedure of this routine is similar with pci_reboot_init except the table in the routine is reboot_dmi_table instead of pci_reboot_dmi_table, the defination of table pci_reboot_init could be found in line 137 of file arch/x86/kernel/reboot.c, if dmi of current system match any items in the table, invoke set_bios_reboot which set the reboot type with BOOT_BIOS selecting BIOS-method for reboots.
the eighth one is init_lapic_sysfs, as the function imply, it initialize local apic filesystem
init_lapic_sysfs () at arch/x86/kernel/apic/apic.c:2161
2161 if (!cpu_has_apic)
(gdb)
2165 error = sysdev_class_register(&lapic_sysclass); # register local apic system device class
(gdb) s
sysdev_class_register (cls=cls@entry=0xc1696ca0 <lapic_sysclass>)
at drivers/base/sys.c:136
136 pr_debug("Registering sysdev class '%s'\n", cls->name);
(gdb) n
138 INIT_LIST_HEAD(&cls->drivers);
(gdb)
139 memset(&cls->kset.kobj, 0x00, sizeof(struct kobject));
(gdb)
140 cls->kset.kobj.parent = &system_kset->kobj;
(gdb)
141 cls->kset.kobj.ktype = &ktype_sysdev_class;
(gdb)
142 cls->kset.kobj.kset = system_kset;
(gdb)
144 retval = kobject_set_name(&cls->kset.kobj, "%s", cls->name);
(gdb)
145 if (retval)
(gdb)
148 return kset_register(&cls->kset);
(gdb)
149 }
(gdb)
init_lapic_sysfs () at arch/x86/kernel/apic/apic.c:2166
2166 if (!error)
(gdb)
2167 error = sysdev_register(&device_lapic); # register local apic device
(gdb) s
sysdev_register (sysdev=sysdev@entry=0xc1696c60 <device_lapic>)
at drivers/base/sys.c:242
242 struct sysdev_class *cls = sysdev->cls;
(gdb) n
244 if (!cls)
(gdb)
247 pr_debug("Registering sys device of class '%s'\n",
(gdb)
251 memset(&sysdev->kobj, 0x00, sizeof(struct kobject));
(gdb)
254 sysdev->kobj.kset = &cls->kset;
(gdb)
257 error = kobject_init_and_add(&sysdev->kobj, &ktype_sysdev, NULL,
(gdb)
261 if (!error) {
(gdb)
264 pr_debug("Registering sys device '%s'\n",
(gdb)
267 mutex_lock(&sysdev_drivers_lock);
(gdb)
273 list_for_each_entry(drv, &cls->drivers, entry) {
(gdb)
277 mutex_unlock(&sysdev_drivers_lock);
(gdb)
278 kobject_uevent(&sysdev->kobj, KOBJ_ADD);
(gdb)
282 }
(gdb)
init_lapic_sysfs () at arch/x86/kernel/apic/apic.c:2169
2169 }
(gdb)
init_smp_flush, it initialzes spin lock of every smp_flush_state item of array flush_state.
some reference of TLB flush could be found in Translation lookaside buffer:
Address space switch
On an address space switch, as occurs on a process switch but not on a thread switch, some TLB entries can become invalid, since the virtual-to-physical mapping is different. The simplest strategy to deal with this is to completely flush the TLB. This means that after a switch, the TLB is empty and any memory reference will be a miss, and it will be some time before things are running back at full speed. Newer CPUs use more effective strategies marking which process an entry is for. This means that if a second process runs for only a short time and jumps back to a first process, it may still have valid entries, saving the time to reload them.
For example, in the Alpha 21264, each TLB entry is tagged with an "address space number" (ASN), and only TLB entries with an ASN matching the current task are considered valid. Another example in the Intel Pentium Pro, the page global enable (PGE) flag in the register CR4 and the global (G) flag of a page-directory or page-table entry can be used to prevent frequently used pages from being automatically invalidated in the TLBs on a task switch or a load of register CR3.
While selective flushing of the TLB is an option in software managed TLBs, the only option in some hardware TLBs (for example, the TLB in the Intel 80386) is the complete flushing of the TLB on an address space switch. Other hardware TLBs (for example, the TLB in the Intel 80486 and later x86 processors, and the TLB in ARM processors) allow the flushing of individual entries from the TLB indexed by virtual address.
cpu_hotplug_pm_sync_init, it registers notifier function for cpu hotplug
cpu_hotplug_pm_sync_init () at kernel/cpu.c:514
514 pm_notifier(cpu_hotplug_pm_callback, 0);
(gdb) s
register_pm_notifier (nb=nb@entry=0xc1698ea0 <cpu_hotplug_pm_callback_nb>)
at kernel/power/main.c:31
31 return blocking_notifier_chain_register(&pm_chain_head, nb);
(gdb) s
blocking_notifier_chain_register (nh=nh@entry=0xc16a01e0 <pm_chain_head>,
n=n@entry=0xc1698ea0 <cpu_hotplug_pm_callback_nb>)
at kernel/notifier.c:212
212 {
(gdb) n
220 if (unlikely(system_state == SYSTEM_BOOTING))
(gdb)
221 return notifier_chain_register(&nh->head, n);
(gdb) s
notifier_chain_register (n=0xc1698ea0 <cpu_hotplug_pm_callback_nb>,
n@entry=0xc16a01e0 <pm_chain_head>, nl=<optimized out>)
at kernel/notifier.c:24
24 while ((*nl) != NULL) {
(gdb) n
29 n->next = *nl;
(gdb)
30 rcu_assign_pointer(*nl, n);
(gdb)
blocking_notifier_chain_register (nh=nh@entry=0xc16a01e0 <pm_chain_head>,
n=n@entry=0xc1698ea0 <cpu_hotplug_pm_callback_nb>)
at kernel/notifier.c:227
227 }
(gdb)
register_pm_notifier (nb=nb@entry=0xc1698ea0 <cpu_hotplug_pm_callback_nb>)
at kernel/power/main.c:32
32 }
(gdb)
cpu_hotplug_pm_sync_init () at kernel/cpu.c:516
516 }
(gdb)
alloc_frozen_cpus, it allocates cpu mask for frozen/slept cpu because of power management
sysctl_init, it sets parent for root_table and its child table
ksysfs_init, in the initialization it creates directory name kernel and files under the directory
ksysfs_init () at kernel/ksysfs.c:163
163 kernel_kobj = kobject_create_and_add("kernel", NULL);
(gdb) s
kobject_create_and_add (name=name@entry=0xc15cfce0 "kernel",
parent=parent@entry=0x0) at lib/kobject.c:652
652 kobj = kobject_create();
(gdb) s
kobject_create () at lib/kobject.c:626
626 kobj = kzalloc(sizeof(*kobj), GFP_KERNEL);
(gdb) n
627 if (!kobj)
(gdb)
630 kobject_init(kobj, &dynamic_kobj_ktype);
(gdb)
631 return kobj;
(gdb)
632 }
(gdb)
kobject_create_and_add (name=name@entry=0xc15cfce0 "kernel",
parent=parent@entry=0x0) at lib/kobject.c:653
653 if (!kobj)
(gdb)
656 retval = kobject_add(kobj, parent, "%s", name);
(gdb) s
kobject_add (kobj=kobj@entry=0xc7086310, parent=parent@entry=0x0,
fmt=fmt@entry=0xc1627337 "%s") at lib/kobject.c:340
340 {
(gdb) n
344 if (!kobj)
(gdb)
347 if (!kobj->state_initialized) {
(gdb)
355 retval = kobject_add_varg(kobj, parent, fmt, args);
(gdb) s
kobject_add_varg (vargs=0xc706bf7c "\340\374\\\301\263\353p\301",
fmt=0xc1627337 "%s", parent=0x0, kobj=0xc7086310) at lib/kobject.c:304
304 retval = kobject_set_name_vargs(kobj, fmt, vargs);
(gdb) s
kobject_set_name_vargs (kobj=kobj@entry=0xc7086310,
fmt=fmt@entry=0xc1627337 "%s",
vargs=vargs@entry=0xc706bf7c "\340\374\\\301\263\353p\301")
at lib/kobject.c:218
218 const char *old_name = kobj->name;
(gdb) n
221 if (kobj->name && !fmt)
(gdb)
224 kobj->name = kvasprintf(GFP_KERNEL, fmt, vargs);
(gdb)
225 if (!kobj->name)
(gdb)
229 while ((s = strchr(kobj->name, '/')))
(gdb)
232 kfree(old_name);
(gdb)
233 return 0;
(gdb)
234 }
(gdb)
kobject_add_varg (vargs=0xc706bf7c "\340\374\\\301\263\353p\301",
fmt=0xc1627337 "%s", parent=0x0, kobj=0xc7086310) at lib/kobject.c:305
305 if (retval) {
(gdb)
309 kobj->parent = parent;
(gdb)
310 return kobject_add_internal(kobj);
(gdb) s
kobject_add_internal (kobj=0xc7086310) at lib/kobject.c:163
163 if (!kobj)
(gdb) n
166 if (!kobj->name || !kobj->name[0]) {
(gdb)
172 parent = kobject_get(kobj->parent);
(gdb)
175 if (kobj->kset) {
(gdb)
182 pr_debug("kobject: '%s' (%p): %s: parent: '%s', set: '%s'\n",
(gdb)
187 error = create_dir(kobj);
(gdb) s
create_dir (kobj=0xc7086310) at lib/kobject.c:50
50 if (kobject_name(kobj)) {
(gdb) n
51 error = sysfs_create_dir(kobj);
(gdb) s
sysfs_create_dir (kobj=kobj@entry=0xc7086310) at fs/sysfs/dir.c:715
715 BUG_ON(!kobj);
(gdb)
717 if (kobj->parent)
(gdb)
720 parent_sd = &sysfs_root;
(gdb)
722 error = create_dir(kobj, parent_sd, kobject_name(kobj), &sd);
(gdb) s
create_dir (kobj=kobj@entry=0xc7086310, parent_sd=0xc16ac580 <sysfs_root>,
name=0xc70207a8 "kernel", p_sd=p_sd@entry=0xc706bf10)
at fs/sysfs/dir.c:682
682 sd = sysfs_new_dirent(name, mode, SYSFS_DIR);
(gdb)
683 if (!sd)
(gdb)
685 sd->s_dir.kobj = kobj;
(gdb)
688 sysfs_addrm_start(&acxt, parent_sd);
(gdb) s
sysfs_addrm_start (acxt=acxt@entry=0xc706bee8,
parent_sd=parent_sd@entry=0xc16ac580 <sysfs_root>)
at fs/sysfs/dir.c:374
374 memset(acxt, 0, sizeof(*acxt));
(gdb)
375 acxt->parent_sd = parent_sd;
(gdb)
381 mutex_lock(&sysfs_mutex);
(gdb)
383 inode = ilookup5(sysfs_sb, parent_sd->s_ino, sysfs_ilookup_test,
(gdb)
385 if (inode) {
(gdb)
386 WARN_ON(inode->i_state & I_NEW);
(gdb)
389 acxt->parent_inode = inode;
(gdb)
395 if (!mutex_trylock(&inode->i_mutex)) {
(gdb)
401 }
(gdb)
create_dir (kobj=kobj@entry=0xc7086310, parent_sd=0xc16ac580 <sysfs_root>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf10) at fs/sysfs/dir.c:689
689 rc = sysfs_add_one(&acxt, sd);
(gdb) s
sysfs_add_one (acxt=acxt@entry=0xc706bee8, sd=sd@entry=0xc7022948)
at fs/sysfs/dir.c:482
482 ret = __sysfs_add_one(acxt, sd);
(gdb) s
__sysfs_add_one (acxt=acxt@entry=0xc706bee8, sd=sd@entry=0xc7022948)
at fs/sysfs/dir.c:425
425 if (sysfs_find_dirent(acxt->parent_sd, sd->s_name))
(gdb) n
428 sd->s_parent = sysfs_get(acxt->parent_sd);
(gdb)
430 if (sysfs_type(sd) == SYSFS_DIR && acxt->parent_inode)
(gdb)
431 inc_nlink(acxt->parent_inode);
(gdb)
433 acxt->cnt++;
(gdb)
435 sysfs_link_sibling(sd);
(gdb)
437 return 0;
(gdb)
438 }
(gdb)
sysfs_add_one (acxt=acxt@entry=0xc706bee8, sd=sd@entry=0xc7022948)
at fs/sysfs/dir.c:483
483 if (ret == -EEXIST) {
(gdb)
496 }
(gdb)
create_dir (kobj=kobj@entry=0xc7086310, parent_sd=<optimized out>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf10) at fs/sysfs/dir.c:690
690 sysfs_addrm_finish(&acxt);
(gdb) s
sysfs_addrm_finish (acxt=acxt@entry=0xc706bee8) at fs/sysfs/dir.c:595
595 mutex_unlock(&sysfs_mutex);
(gdb)
596 if (acxt->parent_inode) {
(gdb)
600 if (acxt->cnt)
(gdb)
601 inode->i_ctime = inode->i_mtime = CURRENT_TIME;
(gdb)
603 mutex_unlock(&inode->i_mutex);
(gdb)
604 iput(inode);
(gdb)
608 while (acxt->removed) {
(gdb)
619 }
(gdb)
create_dir (kobj=kobj@entry=0xc7086310, parent_sd=<optimized out>,
name=<optimized out>, p_sd=p_sd@entry=0xc706bf10) at fs/sysfs/dir.c:692
692 if (rc == 0)
(gdb)
693 *p_sd = sd;
(gdb)
698 }
(gdb)
sysfs_create_dir (kobj=kobj@entry=0xc7086310) at fs/sysfs/dir.c:723
723 if (!error)
(gdb)
724 kobj->sd = sd;
(gdb)
726 }
(gdb)
create_dir (kobj=0xc7086310) at lib/kobject.c:52
52 if (!error) {
(gdb)
kobject_add_internal (kobj=0xc7086310) at lib/kobject.c:188
188 if (error) {
(gdb)
204 kobj->state_in_sysfs = 1;
(gdb)
207 }
(gdb)
kobject_add (kobj=kobj@entry=0xc7086310, parent=parent@entry=0x0,
fmt=fmt@entry=0xc1627337 "%s") at lib/kobject.c:359
359 }
(gdb)
kobject_create_and_add (name=name@entry=0xc15cfce0 "kernel",
parent=parent@entry=0x0) at lib/kobject.c:657
657 if (retval) {
(gdb)
664 }
(gdb)
ksysfs_init () at kernel/ksysfs.c:164
164 if (!kernel_kobj) {
(gdb)
168 error = sysfs_create_group(kernel_kobj, &kernel_attr_group); # create files which configured in kernel_attr_group
(gdb) s
sysfs_create_group (kobj=0xc7086310,
grp=grp@entry=0xc169e380 <kernel_attr_group>) at fs/sysfs/group.c:100
100 return internal_create_group(kobj, 0, grp);
(gdb) s
internal_create_group (kobj=0xc7086310, update=update@entry=0,
grp=grp@entry=0xc169e380 <kernel_attr_group>) at fs/sysfs/group.c:65
65 BUG_ON(!kobj || (!update && !kobj->sd));
(gdb) n
71 if (grp->name) {
(gdb)
76 sd = kobj->sd;
(gdb)
77 sysfs_get(sd);
(gdb)
78 error = create_files(sd, kobj, grp, update);
(gdb) s
create_files (update=0, grp=0xc169e380 <kernel_attr_group>,
kobj=0xc7086310, dir_sd=0xc7022948) at fs/sysfs/group.c:35
35 for (i = 0, attr = grp->attrs; *attr && !error; i++, attr++) {
(gdb) n
41 if (update)
(gdb)
43 if (grp->is_visible) {
(gdb)
36 mode_t mode = 0;
(gdb)
48 error = sysfs_add_file_mode(dir_sd, *attr, SYSFS_KOBJ_ATTR,
(gdb) s
sysfs_add_file_mode (dir_sd=dir_sd@entry=0xc7022948,
attr=0xc169e444 <uevent_seqnum_attr>, type=type@entry=2, amode=292)
at fs/sysfs/file.c:503
503 umode_t mode = (amode & S_IALLUGO) | S_IFREG;
(gdb) n
508 sd = sysfs_new_dirent(attr->name, mode, type);
(gdb)
509 if (!sd)
(gdb)
511 sd->s_attr.attr = (void *)attr;
(gdb)
513 sysfs_addrm_start(&acxt, dir_sd);
(gdb)
514 rc = sysfs_add_one(&acxt, sd);
(gdb)
515 sysfs_addrm_finish(&acxt);
(gdb)
517 if (rc)
(gdb)
521 }
(gdb)
create_files (update=0, grp=0xc169e380 <kernel_attr_group>,
kobj=0xc7086310, dir_sd=0xc7022948) at fs/sysfs/group.c:50
50 if (unlikely(error))
......
(gdb)
35 for (i = 0, attr = grp->attrs; *attr && !error; i++, attr++) {
(gdb)
78 error = create_files(sd, kobj, grp, update);
(gdb)
83 sysfs_put(sd);
(gdb)
85 }
(gdb)
sysfs_create_group (kobj=<optimized out>,
grp=grp@entry=0xc169e380 <kernel_attr_group>) at fs/sysfs/group.c:101
101 }
(gdb)
ksysfs_init () at kernel/ksysfs.c:169
169 if (error)
(gdb)
172 if (notes_size > 0) {
(gdb)
173 notes_attr.size = notes_size;
(gdb)
174 error = sysfs_create_bin_file(kernel_kobj, ¬es_attr);
(gdb) s
sysfs_create_bin_file (kobj=0xc7086310,
attr=attr@entry=0xc169e3c0 <notes_attr>) at fs/sysfs/bin.c:488
488 BUG_ON(!kobj || !kobj->sd || !attr);
(gdb) n
490 return sysfs_add_file(kobj->sd, &attr->attr, SYSFS_KOBJ_BIN_ATTR);
(gdb) s
sysfs_add_file (dir_sd=0xc7022948,
attr=attr@entry=0xc169e3c0 <notes_attr>, type=type@entry=4)
at fs/sysfs/file.c:527
527 return sysfs_add_file_mode(dir_sd, attr, type, attr->mode);
(gdb) s
sysfs_add_file_mode (dir_sd=0xc7022948,
attr=attr@entry=0xc169e3c0 <notes_attr>, type=type@entry=4,
amode=amode@entry=292) at fs/sysfs/file.c:503
503 umode_t mode = (amode & S_IALLUGO) | S_IFREG;
(gdb) n
508 sd = sysfs_new_dirent(attr->name, mode, type);
(gdb)
509 if (!sd)
(gdb)
511 sd->s_attr.attr = (void *)attr;
(gdb)
513 sysfs_addrm_start(&acxt, dir_sd);
(gdb)
514 rc = sysfs_add_one(&acxt, sd);
(gdb)
515 sysfs_addrm_finish(&acxt);
(gdb)
517 if (rc)
(gdb)
521 }
(gdb)
sysfs_add_file (dir_sd=<optimized out>,
attr=attr@entry=0xc169e3c0 <notes_attr>, type=type@entry=4)
at fs/sysfs/file.c:528
528 }
(gdb)
sysfs_create_bin_file (kobj=<optimized out>,
attr=attr@entry=0xc169e3c0 <notes_attr>) at fs/sysfs/bin.c:491
491 }
(gdb)
ksysfs_init () at kernel/ksysfs.c:175
175 if (error)
(gdb)
179 return 0;
(gdb)
187 }
(gdb)
async_init, it creates async_manager_thread and wake it up
the defination of kthread_run in include/linux/kthread.h, it's a macro which create and wake a thread.
/**
* kthread_run - create and wake a thread.
* @threadfn: the function to run until signal_pending(current).
* @data: data ptr for @threadfn.
* @namefmt: printf-style name for the thread.
*
* Description: Convenient wrapper for kthread_create() followed by
* wake_up_process(). Returns the kthread or ERR_PTR(-ENOMEM).
*/
#define kthread_run(threadfn, data, namefmt, ...) \
({ \
struct task_struct *__k \
= kthread_create(threadfn, data, namefmt, ## __VA_ARGS__); \
if (!IS_ERR(__k)) \
wake_up_process(__k); \
__k; \
})
init_jiffies_clocksource, it installs clock source clocksource_jiffies, the clock source defined in kernel/time/jiffies.c, the clock source will be introduced in separate chapter
struct clocksource clocksource_jiffies = {
.name = "jiffies",
.rating = 1, /* lowest valid rating*/
.read = jiffies_read,
.mask = 0xffffffff, /*32bits*/
.mult = NSEC_PER_JIFFY << JIFFIES_SHIFT, /* details above */
.shift = JIFFIES_SHIFT,
};
the entire procedure of routine init_jiffies_clocksource, all clock source added to list clocksource_list
init_jiffies_clocksource () at kernel/time/jiffies.c:69
69 return clocksource_register(&clocksource_jiffies);
(gdb) s
clocksource_register (cs=cs@entry=0xc169ee80 <clocksource_jiffies>)
at kernel/time/clocksource.c:562
562 cs->max_idle_ns = clocksource_max_deferment(cs);
(gdb) n
564 mutex_lock(&clocksource_mutex);
(gdb)
565 clocksource_enqueue(cs);
(gdb) s
clocksource_enqueue (cs=cs@entry=0xc169ee80 <clocksource_jiffies>)
at kernel/time/clocksource.c:543
543 struct list_head *entry = &clocksource_list;
(gdb) n
546 list_for_each_entry(tmp, &clocksource_list, list)
(gdb)
550 list_add(&cs->list, entry);
(gdb)
551 }
(gdb)
clocksource_register (cs=cs@entry=0xc169ee80 <clocksource_jiffies>)
at kernel/time/clocksource.c:566
566 clocksource_enqueue_watchdog(cs);
(gdb) s
clocksource_enqueue_watchdog (cs=0xc169ee80 <clocksource_jiffies>)
at kernel/time/clocksource.c:304
304 spin_lock_irqsave(&watchdog_lock, flags);
(gdb) n
305 if (cs->flags & CLOCK_SOURCE_MUST_VERIFY) {
(gdb)
311 if (cs->flags & CLOCK_SOURCE_IS_CONTINUOUS)
(gdb)
314 if (!watchdog || cs->rating > watchdog->rating) {
(gdb)
321 clocksource_start_watchdog();
(gdb) s
clocksource_start_watchdog () at kernel/time/clocksource.c:266
266 if (watchdog_running || !watchdog || list_empty(&watchdog_list))
(gdb) n
clocksource_enqueue_watchdog (cs=0xc169ee80 <clocksource_jiffies>)
at kernel/time/clocksource.c:322
322 spin_unlock_irqrestore(&watchdog_lock, flags);
(gdb)
clocksource_register (cs=cs@entry=0xc169ee80 <clocksource_jiffies>)
at kernel/time/clocksource.c:567
567 clocksource_select();
(gdb) s
clocksource_select () at kernel/time/clocksource.c:473
473 if (!finished_booting || list_empty(&clocksource_list))
(gdb) n
clocksource_register (cs=cs@entry=0xc169ee80 <clocksource_jiffies>)
at kernel/time/clocksource.c:568
568 mutex_unlock(&clocksource_mutex);
(gdb)
570 }
(gdb)
init_jiffies_clocksource () at kernel/time/jiffies.c:70
70 }
(gdb)
pm_init which is used to initialize power manager. With CONFIG_PM_RUNTIME disabled, the body of routine pm_start_workqueue invoked in pm_init is NULL, the routine creates work queue when CONFIG_PM_RUNTIME enabled. pm_init create kobject named power and create group sub folders for directory power
pm_disk_init which creates group for kobject power_kobj with disk specified attribute
swsusp_header_init which allocates page for swap device
init_zero_pfn which initialzes zero page number zero_pfn
init_zero_pfn () at mm/memory.c:119
119 zero_pfn = page_to_pfn(ZERO_PAGE(0));
(gdb)
121 }
(gdb) p zero_pfn
$3 = 6364
filelock_init which creates cache memory named file_lock_cache
the next one init_script_binfmt which registers default binfmt handler for script with script_format as input parameter
rest of routines listed here with the order it called by do_initcalls:
init_elf_binfmt registers default binfmt handler for elf with elf_format as input parameter
debugfs_init creates and adds kobject named debug, registers debug_fs_type, reference of debugfs could be found here
The procedure of securityfs_init is similar with debugfs_init
random32_init generates some initially weak seeding values to allow to start the random32() engine
virtio_init registers a bus with virtio_bus as its input parameter
regulator_init registers device class ragulator with regulator_class as its input parameter
cpufreq_core_init initializes per cpu variables and create kobject cpufreq with kobject cpu as its parent
cpuidle_init adds CPU global sysfs attributes, registers notifier for qos target PM_QOS_CPU_DMA_LATENCY changes
sock_init initializes size of socket r/w memory, skbuff slab cache and socket inode slab cache, registers and mount socket file system, initialzes net filter if CONFIG_NETFILTER enabled
netlink_proto_init initializes netlink protocal. It invokes proto_register to initialize netlink_proto which allocates cache memories and registers netlink_proto to list; later it initializes data used by netlink and registers netlink operations; finally does rt netlink initialization
bdi_class_init creates bdi class and register the class, creates a directory in the debugfs filesystem
kobject_uevent_init allocates a socket for uevent_sock and sets netlink flag of NETLINK_KOBJECT_UEVENT with NL_NONROOT_RECV, the procedure of socket creation:
kobject_uevent_init () at lib/kobject_uevent.c:324
324 uevent_sock = netlink_kernel_create(&init_net, NETLINK_KOBJECT_UEVENT,
(gdb) s
netlink_kernel_create (net=0xc1f2a8c0 <init_net>, unit=unit@entry=15,
groups=groups@entry=1, input=input@entry=0x0,
cb_mutex=cb_mutex@entry=0x0, module=module@entry=0x0)
at net/netlink/af_netlink.c:1465
1465 BUG_ON(!nl_table);
(gdb)
1467 if (unit < 0 || unit >= MAX_LINKS)
(gdb)
1470 if (sock_create_lite(PF_NETLINK, SOCK_DGRAM, unit, &sock))
(gdb) s
sock_create_lite (family=family@entry=16, type=type@entry=2,
protocol=protocol@entry=15, res=res@entry=0xc706bf74)
at net/socket.c:979
981 err = security_socket_create(family, type, protocol, 1);
(gdb) s
security_socket_create (family=family@entry=16, type=type@entry=2,
protocol=protocol@entry=15, kern=kern@entry=1)
at security/security.c:1011
1011 return security_ops->socket_create(family, type, protocol, kern);
(gdb) s
cap_socket_create (family=16, type=2, protocol=15, kern=1)
at security/capability.c:598
598 }
(gdb) n
security_socket_create (family=family@entry=16, type=type@entry=2,
protocol=protocol@entry=15, kern=kern@entry=1)
at security/security.c:1012
1012 }
(gdb)
sock_create_lite (family=family@entry=16, type=type@entry=2,
protocol=protocol@entry=15, res=res@entry=0xc706bf74)
at net/socket.c:982
982 if (err)
(gdb)
985 sock = sock_alloc();
(gdb) s
sock_alloc () at net/socket.c:486
486 inode = new_inode(sock_mnt->mnt_sb);
(gdb) n
487 if (!inode)
(gdb)
493 inode->i_mode = S_IFSOCK | S_IRWXUGO;
(gdb)
494 inode->i_uid = current_fsuid();
(gdb)
495 inode->i_gid = current_fsgid();
(gdb)
497 percpu_add(sockets_in_use, 1);
(gdb)
498 return sock;
(gdb)
499 }
(gdb)
sock_create_lite (family=family@entry=16, type=type@entry=2,
protocol=protocol@entry=15, res=res@entry=0xc706bf74)
at net/socket.c:986
986 if (!sock) {
(gdb)
991 sock->type = type;
(gdb)
992 err = security_socket_post_create(sock, family, type, protocol, 1);
(gdb) s
security_socket_post_create (sock=sock@entry=0xc6c066c0,
family=family@entry=16, type=type@entry=2, protocol=protocol@entry=15,
kern=kern@entry=1) at security/security.c:1017
1017 return security_ops->socket_post_create(sock, family, type,
(gdb) s
cap_socket_post_create (sock=0xc6c066c0, family=16, type=2, protocol=15,
kern=1) at security/capability.c:604
604 }
(gdb) n
security_socket_post_create (sock=sock@entry=0xc6c066c0,
family=family@entry=16, type=type@entry=2, protocol=protocol@entry=15,
kern=kern@entry=1) at security/security.c:1019
1019 }
(gdb)
sock_create_lite (family=family@entry=16, type=type@entry=2,
protocol=protocol@entry=15, res=res@entry=0xc706bf74)
at net/socket.c:993
993 if (err)
(gdb)
997 *res = sock;
(gdb)
1003 }
(gdb)
netlink_kernel_create (net=<optimized out>, unit=unit@entry=15,
groups=groups@entry=1, input=input@entry=0x0,
cb_mutex=cb_mutex@entry=0x0, module=module@entry=0x0)
at net/netlink/af_netlink.c:1479
1479 if (__netlink_create(&init_net, sock, cb_mutex, unit) < 0)
(gdb) s
__netlink_create (net=0xc1f2a8c0 <init_net>, sock=0xc6c066c0,
cb_mutex=cb_mutex@entry=0x0, protocol=protocol@entry=15)
at net/netlink/af_netlink.c:409
409 sock->ops = &netlink_ops;
(gdb) n
411 sk = sk_alloc(net, PF_NETLINK, GFP_KERNEL, &netlink_proto);
(gdb)
412 if (!sk)
(gdb)
415 sock_init_data(sock, sk);
(gdb)
418 if (cb_mutex)
(gdb)
421 nlk->cb_mutex = &nlk->cb_def_mutex;
(gdb)
422 mutex_init(nlk->cb_mutex);
(gdb)
424 init_waitqueue_head(&nlk->wait);
(gdb)
426 sk->sk_destruct = netlink_sock_destruct;
(gdb)
427 sk->sk_protocol = protocol;
(gdb)
428 return 0;
(gdb)
429 }
(gdb)
netlink_kernel_create (net=<optimized out>, unit=unit@entry=15,
groups=groups@entry=1, input=input@entry=0x0,
cb_mutex=cb_mutex@entry=0x0, module=module@entry=0x0)
at net/netlink/af_netlink.c:1523
1482 sk = sock->sk;
(gdb)
1485 if (groups < 32)
(gdb)
1486 groups = 32;
(gdb)
1488 listeners = kzalloc(NLGRPSZ(groups) + sizeof(struct listeners_rcu_head),
(gdb)
1490 if (!listeners)
(gdb)
1493 sk->sk_data_ready = netlink_data_ready;
(gdb)
1494 if (input)
(gdb)
1497 if (netlink_insert(sk, net, 0))
(gdb)
1501 nlk->flags |= NETLINK_KERNEL_SOCKET;
(gdb)
1503 netlink_table_grab();
(gdb)
1504 if (!nl_table[unit].registered) {
(gdb)
1505 nl_table[unit].groups = groups;
(gdb)
1506 nl_table[unit].listeners = listeners;
(gdb)
1507 nl_table[unit].cb_mutex = cb_mutex;
(gdb)
1508 nl_table[unit].module = module;
(gdb)
1509 nl_table[unit].registered = 1;
(gdb)
1514 netlink_table_ungrab();
(gdb)
1515 return sk;
(gdb)
1525 }
pcibus_class_init registers and initializes pcibus_class
pci_driver_init registers bus pci_bus_type
backlight_class_init creates backlight class and initializes it
tty_class_init creates tty class and initializes it
vtconsole_class_init creates vtconsole class and creates device and initializes the device for every driver in registered_con_driver
spi_init allocates memory for spi read/write buffer, registers spi_bus_type and creates class spi_master_class
i2c_init registers bus i2c_bus_type and add driver for i2c
eisa_init registers bus eisa_bus_type
amd_postcore_init if vender of cpu is not AMD, ignore it
arch_kdebugfs_init creates directory for x86 with debug filesystem and creates deubg directory and files for boot parameters if CONFIG_DEBUG_BOOT_PARAMS enabled
init_pit_clocksource registers PIT clock source if needed
mtrr_if_init create mtrr folder under /proc if supported
ffh_cstate_init allocates per cpu variable cpu_cstate_entry, about cstate, it could be found here
arch_init_ftrace_syscalls allocates memory for syscalls_metadata array and initializes it with the meta data of every syscall
kdump_buf_page_init allocates pages for kdump buffer
acpi_pci_init registers acpi bus type acpi_pci_bus and sets acpi platform power manager acpi_pci_platform_pm
dmi_id_init allocates memory and adds device for dmi_dev
pci_arch_init checks pci type and initializes raw_pci_ops and raw_pci_ext_ops, check system DMI data of pciprobe_dmi_table and can_skip_pciprobe_dmi_table
topology_init initializes topology of cpu
mtrr_init_finialize if CPU has no MTRR and seems to not support SMP. They have specific drivers mtrr_sysdev_driver, we use a tricky method to support suspend/resume for them
param_sysfs_init creates module_kset and adds module_parameters to sysfs for "modules" built into the kernel
pm_sysrq_init registers key and operation for power off
audit_watch_init initializes an inotify handle with operations audit_inotify_ops
init_slow_work initializes slow work facility, slow work is a thread pool implementation which can be used by kernel subsystems to run tasks which are expected to take a fair amount of time in their execution.
default_bdi_init initializes default backing device interface synchronization, initializes default backing device info and registers it
init_bio initializes block I/O
fsnotify_init initializes sleep-RCU structure fsnotify_grp_srcu which is used to protect reads while running the fsnotify_groups list
fsnotify_notification_init allocates caches and initializes q_overflow_event for fsnotify notification
cryptomgr_init registers notify block cryptomgr_notifier for crypto
blk_settings_init initializes frame page number limitation for block device
blk_ioc_init allocates cache memory for block device I/O context
blk_softirq_init initializes softirq for block device
blk_iopoll_setup initializes softirq for block device iopoll
genhd_device_init initializes hard disk device
blk_dev_integrity_init initializes block device integrity which allocate cache memory named blkdev_integrity
gpiolib_debugfs_init creates gpio file under debugfs
pci_slot_init initializes pci slot, creates pci_slots_kset inside this routine
fbmem_init initializes frame buffer device, it creates folder fb under /proc, creates and registers cdev for frame buffer and creates fb class
acpi_init initializes acpi, it creates and adds kobject and so on
dock_init initializes dock
acpi_pci_root_init registers acpi_pci_root_drivers
acpi_pci_link_init registers acpi_pci_link_driver
pnp_init registers bus pnp_bus_type
misc_init creates folder fs under /proc, creates misc class and registers misc charactor device
vga_arb_device_init registers device vga_arb_device, registers notifier for pci bus and adds all pci devices satisfying vga class in the arbiter by default
cn_init, wm8400_module_init, wm831x_i2c_init, wm8350_i2c_init, twl4030_init, ezx_pcap_init, da903x_init, ab3100_i2c_init, init_scsi, ata_init, nop_usb_xceiv_init, usb_init, serio_init, input_init, rtc_init, power_supply_class_init, hwmon_init, thermal_init, md_init, mmc_init, leds_init, pci_subsys_init, proto_init, net_dev_init, neigh_init, fib_rules_init, pktsched_init, tc_filter_init, tc_action_init, genl_init, cipso_v4_init, wireless_nlevent_init, netlbl_init, rfkill_init, sysctl_init, print_all_ICs, hpet_late_init, init_k8_nbs, clocksource_done_booting, ftrace_init_debugfs, rb_init_debugfs, tracer_init_debugfs, init_trace_printk_function_export, event_trace_init, init_pipe_fs, eventpoll_init, anon_inode_init, tomoyo_initerface_init, blk_scsi_ioctl_init, acpi_event_init, pnpacpi_init, pnp_system_init, chr_dev_init, firmware_class_init, cpufreq_gov_performance_init, init_acpi_pm_clocksource, pcibios_assign_resources, sysctl_core_init, inet_init, af_unix_init, pci_apply_final_quirks, populate_rootfs, pci_iommu_init, i8259A_init_sysfs, sbf_init, i8237A_init_sysfs, add_rtc_cmos, cache_sysfs_init, mce_init_device, threshold_init_device, thermal_throttle_init_device, speedstep_init, ioapic_init_sysfs, add_pcspkr, microcode_init, start_periodic_check_for_corruption, start_pageattr_test, pt_dump_init, crc32c_intel_mod_init, init_sched_debug_procfs, proc_schedstat_init, proc_execdomains_init, ioresources_init, uid_cache_init, init_posix_timers, init_posix_cpu_timers, nsproxy_cache_init, create_proc_profile, timekeeping_init_device, init_clocksource_sysfs, init_timer_list_procfs, init_tstats_procfs, lockdep_proc_init, futex_init, init_rttest, proc_dma_init, proc_modules_init, kallsyms_init, snapshot_device_init, crash_save_vmcoreinfo_init, crash_notes_memory_init, backtrace_regression_test, audit_init, audit_tree_init, init_kprobes, hung_task_init, rcu_torture_init, utsname_sysctl_init, init_tracepoints, init_lstats_procfs, ftrace_mod_cmd_init, init_events, init_sched_switch_trace, init_function_trace, init_wakeup_tracer, stack_trace_init, init_mmio_trace, init_graph_trace, init_blk_tracer, perf_event_sysfs_init, init_per_zone_wmark_min, kswapd_init, setup_vmstat, mm_sysfs_init, proc_vmalloc_init, init_emergency_pool, procswaps_init, hugetlb_init, ksm_init, slab_proc_init, slab_sysfs_init, fasync_init, proc_filesystems_init, dnotify_init, inotify_setup, inotify_user_setup, aio_setup, proc_locks_init, init_mbcache, dquot_init, proc_cmdline_init, proc_cpuinfo_init, proc_devices_init, proc_interrupts_init, proc_loadavg_init, proc_meminfo_init, proc_stat_init, proc_uptime_init, proc_version_init, proc_softirqs_init, proc_kcore_init, vmcore_init, proc_kmsg_init, proc_page_init, init_devpts_fs, init_ext4_fs, journal_init, init_squashfs_fs, init_ramfs_fs, init_hugetlbfs_fs, init_fat_fs, init_vfat_fs, init_nls_cp437, fuse_init, ipc_init, ipc_sysctl_init, key_proc_init, selinux_nf_ip_init, init_sel_fs, selnl_init, sel_netif_init, sel_netnode_init, sel_netport_init, aurule_init, init_smk_fs, crypto_wq_init, crypto_algapi_init, skcipher_module_init, chainiv_module_init, eseqiv_module_init, hmac_module_init, md5_mod_init, sha1_generic_mod_init, sha256_generic_mod_init, sha512_generic_mod_init, crypto_ecb_module_init, crypto_cbc_module_init, aes_init, crc32c_mod_init, lzo_mod_init, krng_mod_init, proc_genhd_init, bsg_init, noop_init, as_init, deadline_init, cfq_init, debug_objects_init_debugfs, percpu_counter_startup, audit_classes_init, dynamic_debug_init, pci_proc_init, pcie_portdrv_init, aer_service_init, pci_hotplug_init, pcied_init, acpiphp_init, fb_console_init, imsttfb_init, asiliantfb_init, vesafb_init, efifb_init, acpi_reserve_resources, irqrouter_init_sysfs, acpi_ac_init, acpi_button_init, acpi_fan_init, acpi_pci_slot_init, acpi_processor_init, acpi_container_init, acpi_thermal_init, acpi_battery_init, isapnp_init, init
Breakpoint 3, do_one_initcall (fn=0xc1722231 <init>) at init/main.c:716
716 {
(gdb) bt
#0 do_one_initcall (fn=0xc1722231 <init>) at init/main.c:716
#1 0xc16fa930 in do_initcalls () at init/main.c:768
#2 do_basic_setup () at init/main.c:790
#3 kernel_init (unused=<optimized out>) at init/main.c:890
#4 0xc1003ad7 in kernel_thread_helper ()
at arch/x86/kernel/entry_32.S:1000
rand_initialize, tty_init, sysrq_init, init
Breakpoint 3, do_one_initcall (fn=0xc1722fcf <init>) at init/main.c:716
716 {
(gdb) bt
#0 do_one_initcall (fn=0xc1722fcf <init>) at init/main.c:716
#1 0xc16fa930 in do_initcalls () at init/main.c:768
#2 do_basic_setup () at init/main.c:790
#3 kernel_init (unused=<optimized out>) at init/main.c:890
#4 0xc1003ad7 in kernel_thread_helper ()
at arch/x86/kernel/entry_32.S:1000
hpet_init, agp_init, agp_amdk7_init, agp_amd64_init, agp_intel_init, agp_nvidia_init, agp_via_init, init_tis, cn_proc_init, serial8250_init, serial8250_pnp_init, serial8250_pci_init, isa_bus_init, topology_sysfs_init, brd_init, loop_init, scsi_dh_init, init_sd, init_sr, init_sg, piix_init, sis_init, ata_generic_init, net_olddevs_init, ppp_init, tun_init, cdrom_init, ehci_hcd_init, ohci_hcd_mod_init, uhci_hcd_init, i8042_init, mousedev_init, evdev_init, atkbd_init, uinput_init, cmos_init, i2c_dev_init, dm_init, pci_eisa_init_module, virtual_eisa_root_init, cpufreq_stats_init, cpufreq_gov_powersave_init, cpufreq_gov_userspace_init, cpufreq_gov_dbs_init, cpufreq_gov_dbs_init, init_ladder, init_menu, efivars_init, staging_init, flow_cache_init, blackhole_module_init, sysctl_ipv4_init, init_syncookies, ipv4_netfilter_init, cubictcp_register, inet6_init, packet_init, dcbnl_init, mce_debugfs_init, severities_debugfs_init, powernowk8_init, acpi_cpufreq_init, centrino_init, hpet_insert_resource, update_mp_table, lapic_insert_resource, print_ipi_mode, init_lapic_nmi_sysfs, io_apic_bug_finalize, check_early_ioremap_leak, pat_memtype_list_init, sched_init_debug, init_oops_id, disable_boot_consoles, pm_qos_power_init, software_resume, debugfs_kprobe_init, taskstats_init, clear_boot_tracer, fail_page_alloc_debugfs, max_swapfiles_check, failslab_debugfs_init, init_ima, fail_make_request_debugfs, fail_io_timeout_debugfs, random32_reseed, pci_resource_alignment_sysfs_init, pci_sysfs_init, regulator_init_complete, random_int_secret_init, scsi_complete_async_scans, rtc_hctosys, edd_init, memmap_init, pci_mmcfg_late_insert_resources, net_secret_init, tcp_congestion_default, initialize_hashrnd
That's all of the init funtions invoked in do_initcalls
Check if there is user specified init path exists, if yes, do the rest of the initialization. If not, use the default path
/initto continue.Try to access the init path, find the dentry from init path.
The final function invoked in
kernel_init:init_post
wait for all asynchronous functions to complete, async_synchronize_full returns when there are no asynchronous function calls in the system. Of course, another one could always be submitted immediately thereafter. An asynchronous function call infrastructure
static noinline int init_post(void)
__releases(kernel_lock)
{
/* need to finish all async __init code before freeing the memory */
async_synchronize_full();
free init memory, unlock kernel, the lock invoked in the beginning of kernel_init, mark kernel code and kernel read-only data section memory with write protecting, update the system_state
free_initmem();
unlock_kernel();
mark_rodata_ro();
system_state = SYSTEM_RUNNING;
MNMA configuration is disabled in my enviroment, ignore numa_default_policy, open /dev/console, its fd is 0, map the file /dev/console to 1 and 2 seperately with invoking sys_dup twice, it's the source of stdout, stdin and stderr
numa_default_policy();
if (sys_open((const char __user *) "/dev/console", O_RDWR, 0) < 0)
sys_open (filename=filename@entry=0xc15c52bb "/dev/console",
flags=flags@entry=2, mode=mode@entry=0) at fs/open.c:1060
1060 ret = do_sys_open(AT_FDCWD, filename, flags, mode);
(gdb) s
do_sys_open (dfd=dfd@entry=-100,
filename=filename@entry=0xc15c52bb "/dev/console",
flags=flags@entry=2, mode=mode@entry=0) at fs/open.c:1032
1032 {
(gdb) n
1033 char *tmp = getname(filename);
(gdb)
1036 if (!IS_ERR(tmp)) {
(gdb)
1037 fd = get_unused_fd_flags(flags); # allocate file descriptor
(gdb)
1038 if (fd >= 0) {
(gdb)
1039 struct file *f = do_filp_open(dfd, tmp, flags, mode, 0);
(gdb)
1040 if (IS_ERR(f)) {
(gdb)
1044 fsnotify_open(f->f_path.dentry);
(gdb)
1045 fd_install(fd, f); # map fd and file
(gdb)
1048 putname(tmp);
(gdb)
1051 }
(gdb)
sys_open (filename=filename@entry=0xc15c52bb "/dev/console",
flags=flags@entry=2, mode=mode@entry=0) at fs/open.c:1064
1064 }
(gdb)
printk(KERN_WARNING "Warning: unable to open an initial console.\n");
(void) sys_dup(0);
(void) sys_dup(0);
mark the init process as unkillable, we can find the answer why you can't kill init
current->signal->flags |= SIGNAL_UNKILLABLE;
and the defination of the macro SIGNAL_UNKILLABLE
#define SIGNAL_UNKILLABLE 0x00000040 /* for init: ignore fatal signals */
execute init script /init
if (ramdisk_execute_command) {
run_init_process(ramdisk_execute_command);
(gdb) s
run_init_process (init_filename=0xc15c52f0 "/init") at init/main.c:804
804 kernel_execve(init_filename, argv_init, envp_init);
(gdb) s
803 argv_init[0] = init_filename;
(gdb) n
804 kernel_execve(init_filename, argv_init, envp_init);
(gdb) s
kernel_execve (filename=0xc15c52f0 "/init",
argv=argv@entry=0xc1694240 <argv_init>,
envp=envp@entry=0xc16941a0 <envp_init>)
at arch/x86/kernel/sys_i386_32.c:217
217 {
(gdb) n
219 asm volatile ("push %%ebx ; movl %2,%%ebx ; int $0x80 ; pop %%ebx"
printk(KERN_WARNING "Failed to execute %s\n",
ramdisk_execute_command);
}
if no /init, execute other script
/*
* We try each of these until one succeeds.
*
* The Bourne shell can be used instead of init if we are
* trying to recover a really broken machine.
*/
if (execute_command) {
run_init_process(execute_command);
printk(KERN_WARNING "Failed to execute %s. Attempting "
"defaults...\n", execute_command);
}
run_init_process("/sbin/init");
run_init_process("/etc/init");
run_init_process("/bin/init");
run_init_process("/bin/sh");
panic("No init found. Try passing init= option to kernel.");
}
Links
- Optimizing preemption
- completions - wait for completion handling
- Dynamic Power Management
- sysfs
- Sleepable RCU
- Network namespaces
- Introducing Linux Network Namespaces
- Non-volatile memory
- nvSRAM
- SMP flush
- Translation lookaside buffer
- Translation Lookaside Buffer (TLB)
- Power Management In The Linux Kernel
- The Sysctl Interface
- Clock sources, Clock events, sched_clock() and delay timers
- debugfs
- Netlink
- i2c
- MTRR (Memory Type Range Register) control
- CPU performance states (P-states) and CPU operating states (C-states)
- The slow work mechanism
- SLOW WORK ITEM EXECUTION THREAD POOL
- An asynchronous function call infrastructure
- why you can't kill init
- GCC-Inline-Assembly