153 lines · plain
1.. SPDX-License-Identifier: GPL-2.02 3=============4Kernel Stacks5=============6 7Kernel stacks on x86-64 bit8===========================9 10Most of the text from Keith Owens, hacked by AK11 12x86_64 page size (PAGE_SIZE) is 4K.13 14Like all other architectures, x86_64 has a kernel stack for every15active thread. These thread stacks are THREAD_SIZE (4*PAGE_SIZE) big.16These stacks contain useful data as long as a thread is alive or a17zombie. While the thread is in user space the kernel stack is empty18except for the thread_info structure at the bottom.19 20In addition to the per thread stacks, there are specialized stacks21associated with each CPU. These stacks are only used while the kernel22is in control on that CPU; when a CPU returns to user space the23specialized stacks contain no useful data. The main CPU stacks are:24 25* Interrupt stack. IRQ_STACK_SIZE26 27 Used for external hardware interrupts. If this is the first external28 hardware interrupt (i.e. not a nested hardware interrupt) then the29 kernel switches from the current task to the interrupt stack. Like30 the split thread and interrupt stacks on i386, this gives more room31 for kernel interrupt processing without having to increase the size32 of every per thread stack.33 34 The interrupt stack is also used when processing a softirq.35 36Switching to the kernel interrupt stack is done by software based on a37per CPU interrupt nest counter. This is needed because x86-64 "IST"38hardware stacks cannot nest without races.39 40x86_64 also has a feature which is not available on i386, the ability41to automatically switch to a new stack for designated events such as42double fault or NMI, which makes it easier to handle these unusual43events on x86_64. This feature is called the Interrupt Stack Table44(IST). There can be up to 7 IST entries per CPU. The IST code is an45index into the Task State Segment (TSS). The IST entries in the TSS46point to dedicated stacks; each stack can be a different size.47 48An IST is selected by a non-zero value in the IST field of an49interrupt-gate descriptor. When an interrupt occurs and the hardware50loads such a descriptor, the hardware automatically sets the new stack51pointer based on the IST value, then invokes the interrupt handler. If52the interrupt came from user mode, then the interrupt handler prologue53will switch back to the per-thread stack. If software wants to allow54nested IST interrupts then the handler must adjust the IST values on55entry to and exit from the interrupt handler. (This is occasionally56done, e.g. for debug exceptions.)57 58Events with different IST codes (i.e. with different stacks) can be59nested. For example, a debug interrupt can safely be interrupted by an60NMI. arch/x86_64/kernel/entry.S::paranoidentry adjusts the stack61pointers on entry to and exit from all IST events, in theory allowing62IST events with the same code to be nested. However in most cases, the63stack size allocated to an IST assumes no nesting for the same code.64If that assumption is ever broken then the stacks will become corrupt.65 66The currently assigned IST stacks are:67 68* ESTACK_DF. EXCEPTION_STKSZ (PAGE_SIZE).69 70 Used for interrupt 8 - Double Fault Exception (#DF).71 72 Invoked when handling one exception causes another exception. Happens73 when the kernel is very confused (e.g. kernel stack pointer corrupt).74 Using a separate stack allows the kernel to recover from it well enough75 in many cases to still output an oops.76 77* ESTACK_NMI. EXCEPTION_STKSZ (PAGE_SIZE).78 79 Used for non-maskable interrupts (NMI).80 81 NMI can be delivered at any time, including when the kernel is in the82 middle of switching stacks. Using IST for NMI events avoids making83 assumptions about the previous state of the kernel stack.84 85* ESTACK_DB. EXCEPTION_STKSZ (PAGE_SIZE).86 87 Used for hardware debug interrupts (interrupt 1) and for software88 debug interrupts (INT3).89 90 When debugging a kernel, debug interrupts (both hardware and91 software) can occur at any time. Using IST for these interrupts92 avoids making assumptions about the previous state of the kernel93 stack.94 95 To handle nested #DB correctly there exist two instances of DB stacks. On96 #DB entry the IST stackpointer for #DB is switched to the second instance97 so a nested #DB starts from a clean stack. The nested #DB switches98 the IST stackpointer to a guard hole to catch triple nesting.99 100* ESTACK_MCE. EXCEPTION_STKSZ (PAGE_SIZE).101 102 Used for interrupt 18 - Machine Check Exception (#MC).103 104 MCE can be delivered at any time, including when the kernel is in the105 middle of switching stacks. Using IST for MCE events avoids making106 assumptions about the previous state of the kernel stack.107 108For more details see the Intel IA32 or AMD AMD64 architecture manuals.109 110 111Printing backtraces on x86112==========================113 114The question about the '?' preceding function names in an x86 stacktrace115keeps popping up, here's an indepth explanation. It helps if the reader116stares at print_context_stack() and the whole machinery in and around117arch/x86/kernel/dumpstack.c.118 119Adapted from Ingo's mail, Message-ID: <20150521101614.GA10889@gmail.com>:120 121We always scan the full kernel stack for return addresses stored on122the kernel stack(s) [1]_, from stack top to stack bottom, and print out123anything that 'looks like' a kernel text address.124 125If it fits into the frame pointer chain, we print it without a question126mark, knowing that it's part of the real backtrace.127 128If the address does not fit into our expected frame pointer chain we129still print it, but we print a '?'. It can mean two things:130 131 - either the address is not part of the call chain: it's just stale132 values on the kernel stack, from earlier function calls. This is133 the common case.134 135 - or it is part of the call chain, but the frame pointer was not set136 up properly within the function, so we don't recognize it.137 138This way we will always print out the real call chain (plus a few more139entries), regardless of whether the frame pointer was set up correctly140or not - but in most cases we'll get the call chain right as well. The141entries printed are strictly in stack order, so you can deduce more142information from that as well.143 144The most important property of this method is that we _never_ lose145information: we always strive to print _all_ addresses on the stack(s)146that look like kernel text addresses, so if debug information is wrong,147we still print out the real call chain as well - just with more question148marks than ideal.149 150.. [1] For things like IRQ and IST stacks, we also scan those stacks, in151 the right order, and try to cross from one stack into another152 reconstructing the call chain. This works most of the time.153