Based on kernel version 3.19. Page generated on 2015-02-13 21:23 EST.
1 This file documents some of the kernel entries in 2 arch/x86/kernel/entry_64.S. A lot of this explanation is adapted from 3 an email from Ingo Molnar: 4 5 http://lkml.kernel.org/r/<20110529191055.GC9835%40elte.hu> 6 7 The x86 architecture has quite a few different ways to jump into 8 kernel code. Most of these entry points are registered in 9 arch/x86/kernel/traps.c and implemented in arch/x86/kernel/entry_64.S 10 for 64-bit, arch/x86/kernel/entry_32.S for 32-bit and finally 11 arch/x86/ia32/ia32entry.S which implements the 32-bit compatibility 12 syscall entry points and thus provides for 32-bit processes the 13 ability to execute syscalls when running on 64-bit kernels. 14 15 The IDT vector assignments are listed in arch/x86/include/asm/irq_vectors.h. 16 17 Some of these entries are: 18 19 - system_call: syscall instruction from 64-bit code. 20 21 - ia32_syscall: int 0x80 from 32-bit or 64-bit code; compat syscall 22 either way. 23 24 - ia32_syscall, ia32_sysenter: syscall and sysenter from 32-bit 25 code 26 27 - interrupt: An array of entries. Every IDT vector that doesn't 28 explicitly point somewhere else gets set to the corresponding 29 value in interrupts. These point to a whole array of 30 magically-generated functions that make their way to do_IRQ with 31 the interrupt number as a parameter. 32 33 - APIC interrupts: Various special-purpose interrupts for things 34 like TLB shootdown. 35 36 - Architecturally-defined exceptions like divide_error. 37 38 There are a few complexities here. The different x86-64 entries 39 have different calling conventions. The syscall and sysenter 40 instructions have their own peculiar calling conventions. Some of 41 the IDT entries push an error code onto the stack; others don't. 42 IDT entries using the IST alternative stack mechanism need their own 43 magic to get the stack frames right. (You can find some 44 documentation in the AMD APM, Volume 2, Chapter 8 and the Intel SDM, 45 Volume 3, Chapter 6.) 46 47 Dealing with the swapgs instruction is especially tricky. Swapgs 48 toggles whether gs is the kernel gs or the user gs. The swapgs 49 instruction is rather fragile: it must nest perfectly and only in 50 single depth, it should only be used if entering from user mode to 51 kernel mode and then when returning to user-space, and precisely 52 so. If we mess that up even slightly, we crash. 53 54 So when we have a secondary entry, already in kernel mode, we *must 55 not* use SWAPGS blindly - nor must we forget doing a SWAPGS when it's 56 not switched/swapped yet. 57 58 Now, there's a secondary complication: there's a cheap way to test 59 which mode the CPU is in and an expensive way. 60 61 The cheap way is to pick this info off the entry frame on the kernel 62 stack, from the CS of the ptregs area of the kernel stack: 63 64 xorl %ebx,%ebx 65 testl $3,CS+8(%rsp) 66 je error_kernelspace 67 SWAPGS 68 69 The expensive (paranoid) way is to read back the MSR_GS_BASE value 70 (which is what SWAPGS modifies): 71 72 movl $1,%ebx 73 movl $MSR_GS_BASE,%ecx 74 rdmsr 75 testl %edx,%edx 76 js 1f /* negative -> in kernel */ 77 SWAPGS 78 xorl %ebx,%ebx 79 1: ret 80 81 and the whole paranoid non-paranoid macro complexity is about whether 82 to suffer that RDMSR cost. 83 84 If we are at an interrupt or user-trap/gate-alike boundary then we can 85 use the faster check: the stack will be a reliable indicator of 86 whether SWAPGS was already done: if we see that we are a secondary 87 entry interrupting kernel mode execution, then we know that the GS 88 base has already been switched. If it says that we interrupted 89 user-space execution then we must do the SWAPGS. 90 91 But if we are in an NMI/MCE/DEBUG/whatever super-atomic entry context, 92 which might have triggered right after a normal entry wrote CS to the 93 stack but before we executed SWAPGS, then the only safe way to check 94 for GS is the slower method: the RDMSR. 95 96 So we try only to mark those entry methods 'paranoid' that absolutely 97 need the more expensive check for the GS base - and we generate all 98 'normal' entry points with the regular (faster) entry macros.