Based on kernel version 4.9. Page generated on 2016-12-21 14:34 EST.
1 An ad-hoc collection of notes on IA64 MCA and INIT processing. Feel 2 free to update it with notes about any area that is not clear. 3 4 --- 5 6 MCA/INIT are completely asynchronous. They can occur at any time, when 7 the OS is in any state. Including when one of the cpus is already 8 holding a spinlock. Trying to get any lock from MCA/INIT state is 9 asking for deadlock. Also the state of structures that are protected 10 by locks is indeterminate, including linked lists. 11 12 --- 13 14 The complicated ia64 MCA process. All of this is mandated by Intel's 15 specification for ia64 SAL, error recovery and unwind, it is not as 16 if we have a choice here. 17 18 * MCA occurs on one cpu, usually due to a double bit memory error. 19 This is the monarch cpu. 20 21 * SAL sends an MCA rendezvous interrupt (which is a normal interrupt) 22 to all the other cpus, the slaves. 23 24 * Slave cpus that receive the MCA interrupt call down into SAL, they 25 end up spinning disabled while the MCA is being serviced. 26 27 * If any slave cpu was already spinning disabled when the MCA occurred 28 then it cannot service the MCA interrupt. SAL waits ~20 seconds then 29 sends an unmaskable INIT event to the slave cpus that have not 30 already rendezvoused. 31 32 * Because MCA/INIT can be delivered at any time, including when the cpu 33 is down in PAL in physical mode, the registers at the time of the 34 event are _completely_ undefined. In particular the MCA/INIT 35 handlers cannot rely on the thread pointer, PAL physical mode can 36 (and does) modify TP. It is allowed to do that as long as it resets 37 TP on return. However MCA/INIT events expose us to these PAL 38 internal TP changes. Hence curr_task(). 39 40 * If an MCA/INIT event occurs while the kernel was running (not user 41 space) and the kernel has called PAL then the MCA/INIT handler cannot 42 assume that the kernel stack is in a fit state to be used. Mainly 43 because PAL may or may not maintain the stack pointer internally. 44 Because the MCA/INIT handlers cannot trust the kernel stack, they 45 have to use their own, per-cpu stacks. The MCA/INIT stacks are 46 preformatted with just enough task state to let the relevant handlers 47 do their job. 48 49 * Unlike most other architectures, the ia64 struct task is embedded in 50 the kernel stack. So switching to a new kernel stack means that 51 we switch to a new task as well. Because various bits of the kernel 52 assume that current points into the struct task, switching to a new 53 stack also means a new value for current. 54 55 * Once all slaves have rendezvoused and are spinning disabled, the 56 monarch is entered. The monarch now tries to diagnose the problem 57 and decide if it can recover or not. 58 59 * Part of the monarch's job is to look at the state of all the other 60 tasks. The only way to do that on ia64 is to call the unwinder, 61 as mandated by Intel. 62 63 * The starting point for the unwind depends on whether a task is 64 running or not. That is, whether it is on a cpu or is blocked. The 65 monarch has to determine whether or not a task is on a cpu before it 66 knows how to start unwinding it. The tasks that received an MCA or 67 INIT event are no longer running, they have been converted to blocked 68 tasks. But (and its a big but), the cpus that received the MCA 69 rendezvous interrupt are still running on their normal kernel stacks! 70 71 * To distinguish between these two cases, the monarch must know which 72 tasks are on a cpu and which are not. Hence each slave cpu that 73 switches to an MCA/INIT stack, registers its new stack using 74 set_curr_task(), so the monarch can tell that the _original_ task is 75 no longer running on that cpu. That gives us a decent chance of 76 getting a valid backtrace of the _original_ task. 77 78 * MCA/INIT can be nested, to a depth of 2 on any cpu. In the case of a 79 nested error, we want diagnostics on the MCA/INIT handler that 80 failed, not on the task that was originally running. Again this 81 requires set_curr_task() so the MCA/INIT handlers can register their 82 own stack as running on that cpu. Then a recursive error gets a 83 trace of the failing handler's "task". 84 85  My (Keith Owens) original design called for ia64 to separate its 86 struct task and the kernel stacks. Then the MCA/INIT data would be 87 chained stacks like i386 interrupt stacks. But that required 88 radical surgery on the rest of ia64, plus extra hard wired TLB 89 entries with its associated performance degradation. David 90 Mosberger vetoed that approach. Which meant that separate kernel 91 stacks meant separate "tasks" for the MCA/INIT handlers. 92 93 --- 94 95 INIT is less complicated than MCA. Pressing the nmi button or using 96 the equivalent command on the management console sends INIT to all 97 cpus. SAL picks one of the cpus as the monarch and the rest are 98 slaves. All the OS INIT handlers are entered at approximately the same 99 time. The OS monarch prints the state of all tasks and returns, after 100 which the slaves return and the system resumes. 101 102 At least that is what is supposed to happen. Alas there are broken 103 versions of SAL out there. Some drive all the cpus as monarchs. Some 104 drive them all as slaves. Some drive one cpu as monarch, wait for that 105 cpu to return from the OS then drive the rest as slaves. Some versions 106 of SAL cannot even cope with returning from the OS, they spin inside 107 SAL on resume. The OS INIT code has workarounds for some of these 108 broken SAL symptoms, but some simply cannot be fixed from the OS side. 109 110 --- 111 112 The scheduler hooks used by ia64 (curr_task, set_curr_task) are layer 113 violations. Unfortunately MCA/INIT start off as massive layer 114 violations (can occur at _any_ time) and they build from there. 115 116 At least ia64 makes an attempt at recovering from hardware errors, but 117 it is a difficult problem because of the asynchronous nature of these 118 errors. When processing an unmaskable interrupt we sometimes need 119 special code to cope with our inability to take any locks. 120 121 --- 122 123 How is ia64 MCA/INIT different from x86 NMI? 124 125 * x86 NMI typically gets delivered to one cpu. MCA/INIT gets sent to 126 all cpus. 127 128 * x86 NMI cannot be nested. MCA/INIT can be nested, to a depth of 2 129 per cpu. 130 131 * x86 has a separate struct task which points to one of multiple kernel 132 stacks. ia64 has the struct task embedded in the single kernel 133 stack, so switching stack means switching task. 134 135 * x86 does not call the BIOS so the NMI handler does not have to worry 136 about any registers having changed. MCA/INIT can occur while the cpu 137 is in PAL in physical mode, with undefined registers and an undefined 138 kernel stack. 139 140 * i386 backtrace is not very sensitive to whether a process is running 141 or not. ia64 unwind is very, very sensitive to whether a process is 142 running or not. 143 144 --- 145 146 What happens when MCA/INIT is delivered what a cpu is running user 147 space code? 148 149 The user mode registers are stored in the RSE area of the MCA/INIT on 150 entry to the OS and are restored from there on return to SAL, so user 151 mode registers are preserved across a recoverable MCA/INIT. Since the 152 OS has no idea what unwind data is available for the user space stack, 153 MCA/INIT never tries to backtrace user space. Which means that the OS 154 does not bother making the user space process look like a blocked task, 155 i.e. the OS does not copy pt_regs and switch_stack to the user space 156 stack. Also the OS has no idea how big the user space RSE and memory 157 stacks are, which makes it too risky to copy the saved state to a user 158 mode stack. 159 160 --- 161 162 How do we get a backtrace on the tasks that were running when MCA/INIT 163 was delivered? 164 165 mca.c:::ia64_mca_modify_original_stack(). That identifies and 166 verifies the original kernel stack, copies the dirty registers from 167 the MCA/INIT stack's RSE to the original stack's RSE, copies the 168 skeleton struct pt_regs and switch_stack to the original stack, fills 169 in the skeleton structures from the PAL minstate area and updates the 170 original stack's thread.ksp. That makes the original stack look 171 exactly like any other blocked task, i.e. it now appears to be 172 sleeping. To get a backtrace, just start with thread.ksp for the 173 original task and unwind like any other sleeping task. 174 175 --- 176 177 How do we identify the tasks that were running when MCA/INIT was 178 delivered? 179 180 If the previous task has been verified and converted to a blocked 181 state, then sos->prev_task on the MCA/INIT stack is updated to point to 182 the previous task. You can look at that field in dumps or debuggers. 183 To help distinguish between the handler and the original tasks, 184 handlers have _TIF_MCA_INIT set in thread_info.flags. 185 186 The sos data is always in the MCA/INIT handler stack, at offset 187 MCA_SOS_OFFSET. You can get that value from mca_asm.h or calculate it 188 as KERNEL_STACK_SIZE - sizeof(struct pt_regs) - sizeof(struct 189 ia64_sal_os_state), with 16 byte alignment for all structures. 190 191 Also the comm field of the MCA/INIT task is modified to include the pid 192 of the original task, for humans to use. For example, a comm field of 193 'MCA 12159' means that pid 12159 was running when the MCA was 194 delivered.