From f95d47caae5302a63d92be9a0292abc90e2a14e1 Mon Sep 17 00:00:00 2001 From: Jeremy Fitzhardinge Date: Thu, 7 Dec 2006 02:14:02 +0100 Subject: [PATCH] i386: Use %gs as the PDA base-segment in the kernel This patch is the meat of the PDA change. This patch makes several related changes: 1: Most significantly, %gs is now used in the kernel. This means that on entry, the old value of %gs is saved away, and it is reloaded with __KERNEL_PDA. 2: entry.S constructs the stack in the shape of struct pt_regs, and this is passed around the kernel so that the process's saved register state can be accessed. Unfortunately struct pt_regs doesn't currently have space for %gs (or %fs). This patch extends pt_regs to add space for gs (no space is allocated for %fs, since it won't be used, and it would just complicate the code in entry.S to work around the space). 3: Because %gs is now saved on the stack like %ds, %es and the integer registers, there are a number of places where it no longer needs to be handled specially; namely context switch, and saving/restoring the register state in a signal context. 4: And since kernel threads run in kernel space and call normal kernel code, they need to be created with their %gs == __KERNEL_PDA. Signed-off-by: Jeremy Fitzhardinge Signed-off-by: Andi Kleen Cc: Chuck Ebbert <76306.1226@compuserve.com> Cc: Zachary Amsden Cc: Jan Beulich Cc: Andi Kleen Signed-off-by: Andrew Morton --- kernel/fork.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel') diff --git a/kernel/fork.c b/kernel/fork.c index 8cdd3e72ba55..fd22245e3881 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1303,7 +1303,7 @@ fork_out: return ERR_PTR(retval); } -struct pt_regs * __devinit __attribute__((weak)) idle_regs(struct pt_regs *regs) +noinline struct pt_regs * __devinit __attribute__((weak)) idle_regs(struct pt_regs *regs) { memset(regs, 0, sizeof(struct pt_regs)); return regs; -- cgit v1.2.3 From 0741f4d207a644482d7a040f05cd264c98cf7ee8 Mon Sep 17 00:00:00 2001 From: Chuck Ebbert <76306.1226@compuserve.com> Date: Thu, 7 Dec 2006 02:14:11 +0100 Subject: [PATCH] x86: add sysctl for kstack_depth_to_print Add sysctl for kstack_depth_to_print. This lets users change the amount of raw stack data printed in dump_stack() without having to reboot. Signed-off-by: Chuck Ebbert <76306.1226@compuserve.com> Signed-off-by: Andi Kleen --- Documentation/sysctl/kernel.txt | 8 ++++++++ arch/i386/kernel/traps.c | 2 +- arch/x86_64/kernel/traps.c | 2 +- include/asm-x86_64/stacktrace.h | 2 ++ kernel/sysctl.c | 9 +++++++++ 5 files changed, 21 insertions(+), 2 deletions(-) (limited to 'kernel') diff --git a/Documentation/sysctl/kernel.txt b/Documentation/sysctl/kernel.txt index 0bc7f1e3c9e6..5922e84d9133 100644 --- a/Documentation/sysctl/kernel.txt +++ b/Documentation/sysctl/kernel.txt @@ -27,6 +27,7 @@ show up in /proc/sys/kernel: - hotplug - java-appletviewer [ binfmt_java, obsolete ] - java-interpreter [ binfmt_java, obsolete ] +- kstack_depth_to_print [ X86 only ] - l2cr [ PPC only ] - modprobe ==> Documentation/kmod.txt - msgmax @@ -170,6 +171,13 @@ This flag controls the L2 cache of G3 processor boards. If ============================================================== +kstack_depth_to_print: (X86 only) + +Controls the number of words to print when dumping the raw +kernel stack. + +============================================================== + osrelease, ostype & version: # cat osrelease diff --git a/arch/i386/kernel/traps.c b/arch/i386/kernel/traps.c index 7b2f9f022089..1d48a75fa338 100644 --- a/arch/i386/kernel/traps.c +++ b/arch/i386/kernel/traps.c @@ -91,7 +91,7 @@ asmlinkage void alignment_check(void); asmlinkage void spurious_interrupt_bug(void); asmlinkage void machine_check(void); -static int kstack_depth_to_print = 24; +int kstack_depth_to_print = 24; #ifdef CONFIG_STACK_UNWIND static int call_trace = 1; #else diff --git a/arch/x86_64/kernel/traps.c b/arch/x86_64/kernel/traps.c index 264db33476ab..75ceccee178c 100644 --- a/arch/x86_64/kernel/traps.c +++ b/arch/x86_64/kernel/traps.c @@ -108,7 +108,7 @@ static inline void preempt_conditional_cli(struct pt_regs *regs) preempt_enable_no_resched(); } -static int kstack_depth_to_print = 12; +int kstack_depth_to_print = 12; #ifdef CONFIG_STACK_UNWIND static int call_trace = 1; #else diff --git a/include/asm-x86_64/stacktrace.h b/include/asm-x86_64/stacktrace.h index 5eb9799bef76..6f0b54594307 100644 --- a/include/asm-x86_64/stacktrace.h +++ b/include/asm-x86_64/stacktrace.h @@ -1,6 +1,8 @@ #ifndef _ASM_STACKTRACE_H #define _ASM_STACKTRACE_H 1 +extern int kstack_depth_to_print; + /* Generic stack tracer with callbacks */ struct stacktrace_ops { diff --git a/kernel/sysctl.c b/kernel/sysctl.c index 09e569f4792b..6fc5e17086f4 100644 --- a/kernel/sysctl.c +++ b/kernel/sysctl.c @@ -54,6 +54,7 @@ extern int proc_nr_files(ctl_table *table, int write, struct file *filp, #ifdef CONFIG_X86 #include +#include #endif #if defined(CONFIG_SYSCTL) @@ -707,6 +708,14 @@ static ctl_table kern_table[] = { .mode = 0444, .proc_handler = &proc_dointvec, }, + { + .ctl_name = CTL_UNNUMBERED, + .procname = "kstack_depth_to_print", + .data = &kstack_depth_to_print, + .maxlen = sizeof(int), + .mode = 0644, + .proc_handler = &proc_dointvec, + }, #endif #if defined(CONFIG_MMU) { -- cgit v1.2.3 From e2124bb8d369a4bc1afde1959040e33d71c41d5e Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Thu, 7 Dec 2006 02:14:12 +0100 Subject: [PATCH] unwinder: Use probe_kernel_address instead of __get_user in kernel/unwind.c This avoids trouble with the page fault handler if the fault happens inside an interrupt context. Suggested by Linus Cc: jbeulich@novell.com Signed-off-by: Andi Kleen --- kernel/unwind.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) (limited to 'kernel') diff --git a/kernel/unwind.c b/kernel/unwind.c index ed0a21d4a902..af48168a3afb 100644 --- a/kernel/unwind.c +++ b/kernel/unwind.c @@ -14,6 +14,7 @@ #include #include #include +#include #include #include #include @@ -550,7 +551,7 @@ static unsigned long read_pointer(const u8 **pLoc, return 0; } if ((ptrType & DW_EH_PE_indirect) - && __get_user(value, (unsigned long *)value)) + && probe_kernel_address((unsigned long *)value, value)) return 0; *pLoc = ptr.p8; @@ -982,18 +983,19 @@ int unwind(struct unwind_frame_info *frame) & (sizeof(unsigned long) - 1))) { unsigned long link; - if (!__get_user(link, + if (!probe_kernel_address( (unsigned long *)(UNW_FP(frame) - + FRAME_LINK_OFFSET)) + + FRAME_LINK_OFFSET), + link) # if FRAME_RETADDR_OFFSET < 0 && link > bottom && link < UNW_FP(frame) # else && link > UNW_FP(frame) && link < bottom # endif && !(link & (sizeof(link) - 1)) - && !__get_user(UNW_PC(frame), + && !probe_kernel_address( (unsigned long *)(UNW_FP(frame) - + FRAME_RETADDR_OFFSET))) { + + FRAME_RETADDR_OFFSET), UNW_PC(frame))) { UNW_SP(frame) = UNW_FP(frame) + FRAME_RETADDR_OFFSET # if FRAME_RETADDR_OFFSET < 0 - @@ -1104,7 +1106,7 @@ int unwind(struct unwind_frame_info *frame) return -EIO; switch(reg_info[i].width) { #define CASE(n) case sizeof(u##n): \ - __get_user(FRAME_REG(i, u##n), (u##n *)addr); \ + probe_kernel_address((u##n *)addr, FRAME_REG(i, u##n)); \ break CASES; #undef CASE -- cgit v1.2.3 From eef5e0d185fc049bda11fa14ba286fbd357da896 Mon Sep 17 00:00:00 2001 From: Andi Kleen Date: Thu, 7 Dec 2006 02:14:12 +0100 Subject: [PATCH] unwinder: Remove lockdep disabling of nested locks for unwinder Shouldn't be needed anymore since __kernel_text_address is used unconditionally on x86-64 Signed-off-by: Andi Kleen --- kernel/lockdep.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'kernel') diff --git a/kernel/lockdep.c b/kernel/lockdep.c index c9fefdb1a7db..9bb8d784eb02 100644 --- a/kernel/lockdep.c +++ b/kernel/lockdep.c @@ -228,11 +228,7 @@ static int save_trace(struct stack_trace *trace) trace->skip = 3; trace->all_contexts = 0; - /* Make sure to not recurse in case the the unwinder needs to tak -e locks. */ - lockdep_off(); save_stack_trace(trace, NULL); - lockdep_on(); trace->max_entries = trace->nr_entries; -- cgit v1.2.3 From 359ad0d4015a9ab39243f2ebc4eb07915bd618b2 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 7 Dec 2006 02:14:13 +0100 Subject: [PATCH] unwinder: more sanity checks in Dwarf2 unwinder Tighten the requirements on both input to and output from the Dwarf2 unwinder. Signed-off-by: Jan Beulich Signed-off-by: Andi Kleen --- arch/i386/kernel/traps.c | 7 +++++++ arch/x86_64/kernel/traps.c | 7 +++++++ include/asm-i386/unwind.h | 12 ++++-------- include/asm-x86_64/unwind.h | 8 ++------ kernel/unwind.c | 16 +++++++++++++++- 5 files changed, 35 insertions(+), 15 deletions(-) (limited to 'kernel') diff --git a/arch/i386/kernel/traps.c b/arch/i386/kernel/traps.c index 86d8476be4fe..c447807e2a45 100644 --- a/arch/i386/kernel/traps.c +++ b/arch/i386/kernel/traps.c @@ -161,12 +161,19 @@ dump_trace_unwind(struct unwind_frame_info *info, void *data) { struct ops_and_data *oad = (struct ops_and_data *)data; int n = 0; + unsigned long sp = UNW_SP(info); + if (arch_unw_user_mode(info)) + return -1; while (unwind(info) == 0 && UNW_PC(info)) { n++; oad->ops->address(oad->data, UNW_PC(info)); if (arch_unw_user_mode(info)) break; + if ((sp & ~(PAGE_SIZE - 1)) == (UNW_SP(info) & ~(PAGE_SIZE - 1)) + && sp > UNW_SP(info)) + break; + sp = UNW_SP(info); } return n; } diff --git a/arch/x86_64/kernel/traps.c b/arch/x86_64/kernel/traps.c index 9864d195c408..4fdd162f0bef 100644 --- a/arch/x86_64/kernel/traps.c +++ b/arch/x86_64/kernel/traps.c @@ -225,12 +225,19 @@ static int dump_trace_unwind(struct unwind_frame_info *info, void *context) { struct ops_and_data *oad = (struct ops_and_data *)context; int n = 0; + unsigned long sp = UNW_SP(info); + if (arch_unw_user_mode(info)) + return -1; while (unwind(info) == 0 && UNW_PC(info)) { n++; oad->ops->address(oad->data, UNW_PC(info)); if (arch_unw_user_mode(info)) break; + if ((sp & ~(PAGE_SIZE - 1)) == (UNW_SP(info) & ~(PAGE_SIZE - 1)) + && sp > UNW_SP(info)) + break; + sp = UNW_SP(info); } return n; } diff --git a/include/asm-i386/unwind.h b/include/asm-i386/unwind.h index 601fc67bd775..aa2c931e30db 100644 --- a/include/asm-i386/unwind.h +++ b/include/asm-i386/unwind.h @@ -79,17 +79,13 @@ extern asmlinkage int arch_unwind_init_running(struct unwind_frame_info *, void *arg), void *arg); -static inline int arch_unw_user_mode(const struct unwind_frame_info *info) +static inline int arch_unw_user_mode(/*const*/ struct unwind_frame_info *info) { -#if 0 /* This can only work when selector register and EFLAGS saves/restores - are properly annotated (and tracked in UNW_REGISTER_INFO). */ - return user_mode_vm(&info->regs); -#else - return info->regs.eip < PAGE_OFFSET + return user_mode_vm(&info->regs) + || info->regs.eip < PAGE_OFFSET || (info->regs.eip >= __fix_to_virt(FIX_VDSO) - && info->regs.eip < __fix_to_virt(FIX_VDSO) + PAGE_SIZE) + && info->regs.eip < __fix_to_virt(FIX_VDSO) + PAGE_SIZE) || info->regs.esp < PAGE_OFFSET; -#endif } #else diff --git a/include/asm-x86_64/unwind.h b/include/asm-x86_64/unwind.h index 2e7ff10fd775..2f6349e48717 100644 --- a/include/asm-x86_64/unwind.h +++ b/include/asm-x86_64/unwind.h @@ -87,14 +87,10 @@ extern int arch_unwind_init_running(struct unwind_frame_info *, static inline int arch_unw_user_mode(const struct unwind_frame_info *info) { -#if 0 /* This can only work when selector register saves/restores - are properly annotated (and tracked in UNW_REGISTER_INFO). */ - return user_mode(&info->regs); -#else - return (long)info->regs.rip >= 0 + return user_mode(&info->regs) + || (long)info->regs.rip >= 0 || (info->regs.rip >= VSYSCALL_START && info->regs.rip < VSYSCALL_END) || (long)info->regs.rsp >= 0; -#endif } #else diff --git a/kernel/unwind.c b/kernel/unwind.c index af48168a3afb..7e721f104105 100644 --- a/kernel/unwind.c +++ b/kernel/unwind.c @@ -95,6 +95,7 @@ static const struct { typedef unsigned long uleb128_t; typedef signed long sleb128_t; +#define sleb128abs __builtin_labs static struct unwind_table { struct { @@ -787,7 +788,7 @@ int unwind(struct unwind_frame_info *frame) #define FRAME_REG(r, t) (((t *)frame)[reg_info[r].offs]) const u32 *fde = NULL, *cie = NULL; const u8 *ptr = NULL, *end = NULL; - unsigned long pc = UNW_PC(frame) - frame->call_frame; + unsigned long pc = UNW_PC(frame) - frame->call_frame, sp; unsigned long startLoc = 0, endLoc = 0, cfa; unsigned i; signed ptrType = -1; @@ -936,6 +937,9 @@ int unwind(struct unwind_frame_info *frame) state.dataAlign = get_sleb128(&ptr, end); if (state.codeAlign == 0 || state.dataAlign == 0 || ptr >= end) cie = NULL; + else if (UNW_PC(frame) % state.codeAlign + || UNW_SP(frame) % sleb128abs(state.dataAlign)) + return -EPERM; else { retAddrReg = state.version <= 1 ? *ptr++ : get_uleb128(&ptr, end); /* skip augmentation */ @@ -968,6 +972,8 @@ int unwind(struct unwind_frame_info *frame) #ifdef CONFIG_FRAME_POINTER unsigned long top, bottom; + if ((UNW_SP(frame) | UNW_FP(frame)) % sizeof(unsigned long)) + return -EPERM; top = STACK_TOP(frame->task); bottom = STACK_BOTTOM(frame->task); # if FRAME_RETADDR_OFFSET < 0 @@ -1018,6 +1024,7 @@ int unwind(struct unwind_frame_info *frame) || state.regs[retAddrReg].where == Nowhere || state.cfa.reg >= ARRAY_SIZE(reg_info) || reg_info[state.cfa.reg].width != sizeof(unsigned long) + || FRAME_REG(state.cfa.reg, unsigned long) % sizeof(unsigned long) || state.cfa.offs % sizeof(unsigned long)) return -EIO; /* update frame */ @@ -1038,6 +1045,8 @@ int unwind(struct unwind_frame_info *frame) #else # define CASES CASE(8); CASE(16); CASE(32); CASE(64) #endif + pc = UNW_PC(frame); + sp = UNW_SP(frame); for (i = 0; i < ARRAY_SIZE(state.regs); ++i) { if (REG_INVALID(i)) { if (state.regs[i].where == Nowhere) @@ -1118,6 +1127,11 @@ int unwind(struct unwind_frame_info *frame) } } + if (UNW_PC(frame) % state.codeAlign + || UNW_SP(frame) % sleb128abs(state.dataAlign) + || (pc == UNW_PC(frame) && sp == UNW_SP(frame))) + return -EIO; + return 0; #undef CASES #undef FRAME_REG -- cgit v1.2.3 From 6d0185ea611276fdf81991d7774d396bdc1ae392 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 7 Dec 2006 02:14:13 +0100 Subject: [PATCH] unwinder: Add debugging output to the Dwarf2 unwinder Add debugging printks to the unwinder to allow easier debugging when something goes wrong with it. This can be controlled with the new unwinder_debug=N option Most output is given by N=1 AK: Added documentation of unwinder_debug= Signed-off-by: Jan Beulich Signed-off-by: Andi Kleen --- Documentation/kernel-parameters.txt | 3 + kernel/unwind.c | 113 ++++++++++++++++++++++++++++++------ 2 files changed, 99 insertions(+), 17 deletions(-) (limited to 'kernel') diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 4e90aa427aea..d34fd6a28faa 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -1735,6 +1735,9 @@ and is between 256 and 4096 characters. It is defined in the file norandmaps Don't use address space randomization Equivalent to echo 0 > /proc/sys/kernel/randomize_va_space + unwind_debug=N N > 0 will enable dwarf2 unwinder debugging + This is useful to get more information why + you got a "dwarf2 unwinder stuck" ______________________________________________________________________ diff --git a/kernel/unwind.c b/kernel/unwind.c index 7e721f104105..209e248517db 100644 --- a/kernel/unwind.c +++ b/kernel/unwind.c @@ -137,6 +137,17 @@ struct unwind_state { static const struct cfa badCFA = { ARRAY_SIZE(reg_info), 1 }; +static unsigned unwind_debug; +static int __init unwind_debug_setup(char *s) +{ + unwind_debug = simple_strtoul(s, NULL, 0); + return 1; +} +__setup("unwind_debug=", unwind_debug_setup); +#define dprintk(lvl, fmt, args...) \ + ((void)(lvl > unwind_debug \ + || printk(KERN_DEBUG "unwind: " fmt "\n", ##args))) + static struct unwind_table *find_table(unsigned long pc) { struct unwind_table *table; @@ -281,6 +292,7 @@ static void __init setup_unwind_table(struct unwind_table *table, hdrSize = 4 + sizeof(unsigned long) + sizeof(unsigned int) + 2 * n * sizeof(unsigned long); + dprintk(2, "Binary lookup table size for %s: %lu bytes", table->name, hdrSize); header = alloc(hdrSize); if (!header) return; @@ -500,13 +512,17 @@ static unsigned long read_pointer(const u8 **pLoc, const unsigned long *pul; } ptr; - if (ptrType < 0 || ptrType == DW_EH_PE_omit) + if (ptrType < 0 || ptrType == DW_EH_PE_omit) { + dprintk(1, "Invalid pointer encoding %02X (%p,%p).", ptrType, *pLoc, end); return 0; + } ptr.p8 = *pLoc; switch(ptrType & DW_EH_PE_FORM) { case DW_EH_PE_data2: - if (end < (const void *)(ptr.p16u + 1)) + if (end < (const void *)(ptr.p16u + 1)) { + dprintk(1, "Data16 overrun (%p,%p).", ptr.p8, end); return 0; + } if(ptrType & DW_EH_PE_signed) value = get_unaligned(ptr.p16s++); else @@ -514,8 +530,10 @@ static unsigned long read_pointer(const u8 **pLoc, break; case DW_EH_PE_data4: #ifdef CONFIG_64BIT - if (end < (const void *)(ptr.p32u + 1)) + if (end < (const void *)(ptr.p32u + 1)) { + dprintk(1, "Data32 overrun (%p,%p).", ptr.p8, end); return 0; + } if(ptrType & DW_EH_PE_signed) value = get_unaligned(ptr.p32s++); else @@ -527,8 +545,10 @@ static unsigned long read_pointer(const u8 **pLoc, BUILD_BUG_ON(sizeof(u32) != sizeof(value)); #endif case DW_EH_PE_native: - if (end < (const void *)(ptr.pul + 1)) + if (end < (const void *)(ptr.pul + 1)) { + dprintk(1, "DataUL overrun (%p,%p).", ptr.p8, end); return 0; + } value = get_unaligned(ptr.pul++); break; case DW_EH_PE_leb128: @@ -536,10 +556,14 @@ static unsigned long read_pointer(const u8 **pLoc, value = ptrType & DW_EH_PE_signed ? get_sleb128(&ptr.p8, end) : get_uleb128(&ptr.p8, end); - if ((const void *)ptr.p8 > end) + if ((const void *)ptr.p8 > end) { + dprintk(1, "DataLEB overrun (%p,%p).", ptr.p8, end); return 0; + } break; default: + dprintk(2, "Cannot decode pointer type %02X (%p,%p).", + ptrType, ptr.p8, end); return 0; } switch(ptrType & DW_EH_PE_ADJUST) { @@ -549,11 +573,16 @@ static unsigned long read_pointer(const u8 **pLoc, value += (unsigned long)*pLoc; break; default: + dprintk(2, "Cannot adjust pointer type %02X (%p,%p).", + ptrType, *pLoc, end); return 0; } if ((ptrType & DW_EH_PE_indirect) - && probe_kernel_address((unsigned long *)value, value)) + && probe_kernel_address((unsigned long *)value, value)) { + dprintk(1, "Cannot read indirect value %lx (%p,%p).", + value, *pLoc, end); return 0; + } *pLoc = ptr.p8; return value; @@ -702,8 +731,10 @@ static int processCFI(const u8 *start, state->label = NULL; return 1; } - if (state->stackDepth >= MAX_STACK_DEPTH) + if (state->stackDepth >= MAX_STACK_DEPTH) { + dprintk(1, "State stack overflow (%p,%p).", ptr.p8, end); return 0; + } state->stack[state->stackDepth++] = ptr.p8; break; case DW_CFA_restore_state: @@ -718,8 +749,10 @@ static int processCFI(const u8 *start, result = processCFI(start, end, 0, ptrType, state); state->loc = loc; state->label = label; - } else + } else { + dprintk(1, "State stack underflow (%p,%p).", ptr.p8, end); return 0; + } break; case DW_CFA_def_cfa: state->cfa.reg = get_uleb128(&ptr.p8, end); @@ -751,6 +784,7 @@ static int processCFI(const u8 *start, break; case DW_CFA_GNU_window_save: default: + dprintk(1, "Unrecognized CFI op %02X (%p,%p).", ptr.p8[-1], ptr.p8 - 1, end); result = 0; break; } @@ -766,12 +800,17 @@ static int processCFI(const u8 *start, set_rule(*ptr.p8++ & 0x3f, Nowhere, 0, state); break; } - if (ptr.p8 > end) + if (ptr.p8 > end) { + dprintk(1, "Data overrun (%p,%p).", ptr.p8, end); result = 0; + } if (result && targetLoc != 0 && targetLoc < state->loc) return 1; } + if (result && ptr.p8 < end) + dprintk(1, "Data underrun (%p,%p).", ptr.p8, end); + return result && ptr.p8 == end && (targetLoc == 0 @@ -843,6 +882,8 @@ int unwind(struct unwind_frame_info *frame) hdr[3]); } } + if(hdr && !fde) + dprintk(3, "Binary lookup for %lx failed.", pc); if (fde != NULL) { cie = cie_for_fde(fde, table); @@ -864,6 +905,8 @@ int unwind(struct unwind_frame_info *frame) fde = NULL; } else fde = NULL; + if(!fde) + dprintk(1, "Binary lookup result for %lx discarded.", pc); } if (fde == NULL) { for (fde = table->address, tableSize = table->size; @@ -895,6 +938,8 @@ int unwind(struct unwind_frame_info *frame) if (pc >= startLoc && pc < endLoc) break; } + if(!fde) + dprintk(3, "Linear lookup for %lx failed.", pc); } } if (cie != NULL) { @@ -928,6 +973,8 @@ int unwind(struct unwind_frame_info *frame) if (ptr >= end || *ptr) cie = NULL; } + if(!cie) + dprintk(1, "CIE unusable (%p,%p).", ptr, end); ++ptr; } if (cie != NULL) { @@ -938,9 +985,11 @@ int unwind(struct unwind_frame_info *frame) if (state.codeAlign == 0 || state.dataAlign == 0 || ptr >= end) cie = NULL; else if (UNW_PC(frame) % state.codeAlign - || UNW_SP(frame) % sleb128abs(state.dataAlign)) + || UNW_SP(frame) % sleb128abs(state.dataAlign)) { + dprintk(1, "Input pointer(s) misaligned (%lx,%lx).", + UNW_PC(frame), UNW_SP(frame)); return -EPERM; - else { + } else { retAddrReg = state.version <= 1 ? *ptr++ : get_uleb128(&ptr, end); /* skip augmentation */ if (((const char *)(cie + 2))[1] == 'z') { @@ -954,6 +1003,8 @@ int unwind(struct unwind_frame_info *frame) || reg_info[retAddrReg].width != sizeof(unsigned long)) cie = NULL; } + if(!cie) + dprintk(1, "CIE validation failed (%p,%p).", ptr, end); } if (cie != NULL) { state.cieStart = ptr; @@ -967,6 +1018,8 @@ int unwind(struct unwind_frame_info *frame) if ((ptr += augSize) > end) fde = NULL; } + if(!fde) + dprintk(1, "FDE validation failed (%p,%p).", ptr, end); } if (cie == NULL || fde == NULL) { #ifdef CONFIG_FRAME_POINTER @@ -1025,8 +1078,10 @@ int unwind(struct unwind_frame_info *frame) || state.cfa.reg >= ARRAY_SIZE(reg_info) || reg_info[state.cfa.reg].width != sizeof(unsigned long) || FRAME_REG(state.cfa.reg, unsigned long) % sizeof(unsigned long) - || state.cfa.offs % sizeof(unsigned long)) + || state.cfa.offs % sizeof(unsigned long)) { + dprintk(1, "Unusable unwind info (%p,%p).", ptr, end); return -EIO; + } /* update frame */ #ifndef CONFIG_AS_CFI_SIGNAL_FRAME if(frame->call_frame @@ -1051,6 +1106,8 @@ int unwind(struct unwind_frame_info *frame) if (REG_INVALID(i)) { if (state.regs[i].where == Nowhere) continue; + dprintk(1, "Cannot restore register %u (%d).", + i, state.regs[i].where); return -EIO; } switch(state.regs[i].where) { @@ -1059,8 +1116,11 @@ int unwind(struct unwind_frame_info *frame) case Register: if (state.regs[i].value >= ARRAY_SIZE(reg_info) || REG_INVALID(state.regs[i].value) - || reg_info[i].width > reg_info[state.regs[i].value].width) + || reg_info[i].width > reg_info[state.regs[i].value].width) { + dprintk(1, "Cannot restore register %u from register %lu.", + i, state.regs[i].value); return -EIO; + } switch(reg_info[state.regs[i].value].width) { #define CASE(n) \ case sizeof(u##n): \ @@ -1070,6 +1130,9 @@ int unwind(struct unwind_frame_info *frame) CASES; #undef CASE default: + dprintk(1, "Unsupported register size %u (%lu).", + reg_info[state.regs[i].value].width, + state.regs[i].value); return -EIO; } break; @@ -1094,12 +1157,17 @@ int unwind(struct unwind_frame_info *frame) CASES; #undef CASE default: + dprintk(1, "Unsupported register size %u (%u).", + reg_info[i].width, i); return -EIO; } break; case Value: - if (reg_info[i].width != sizeof(unsigned long)) + if (reg_info[i].width != sizeof(unsigned long)) { + dprintk(1, "Unsupported value size %u (%u).", + reg_info[i].width, i); return -EIO; + } FRAME_REG(i, unsigned long) = cfa + state.regs[i].value * state.dataAlign; break; @@ -1111,8 +1179,11 @@ int unwind(struct unwind_frame_info *frame) % sizeof(unsigned long) || addr < startLoc || addr + sizeof(unsigned long) < addr - || addr + sizeof(unsigned long) > endLoc) + || addr + sizeof(unsigned long) > endLoc) { + dprintk(1, "Bad memory location %lx (%lx).", + addr, state.regs[i].value); return -EIO; + } switch(reg_info[i].width) { #define CASE(n) case sizeof(u##n): \ probe_kernel_address((u##n *)addr, FRAME_REG(i, u##n)); \ @@ -1120,6 +1191,8 @@ int unwind(struct unwind_frame_info *frame) CASES; #undef CASE default: + dprintk(1, "Unsupported memory size %u (%u).", + reg_info[i].width, i); return -EIO; } } @@ -1128,9 +1201,15 @@ int unwind(struct unwind_frame_info *frame) } if (UNW_PC(frame) % state.codeAlign - || UNW_SP(frame) % sleb128abs(state.dataAlign) - || (pc == UNW_PC(frame) && sp == UNW_SP(frame))) + || UNW_SP(frame) % sleb128abs(state.dataAlign)) { + dprintk(1, "Output pointer(s) misaligned (%lx,%lx).", + UNW_PC(frame), UNW_SP(frame)); return -EIO; + } + if (pc == UNW_PC(frame) && sp == UNW_SP(frame)) { + dprintk(1, "No progress (%lx,%lx).", pc, sp); + return -EIO; + } return 0; #undef CASES -- cgit v1.2.3 From c65f38d911aa301cea109d38d40925750dd6c2da Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 7 Dec 2006 02:14:19 +0100 Subject: [PATCH] unwinder: fully support linker generated .eh_frame_hdr section Now that binutils' ld is able to properly populate .eh_frame_hdr in the Linux kernel case, here's a patch to add some functionality to the Dwarf2 unwinder to actually be able to make use of this (applies on firstfloor tree with the previously sent patch to add debug output, but not on plain 2.6.19). Signed-off-by: Jan Beulich Signed-off-by: Andi Kleen --- kernel/unwind.c | 66 ++++++++++++++++++++++++++++++++++++++++----------------- 1 file changed, 47 insertions(+), 19 deletions(-) (limited to 'kernel') diff --git a/kernel/unwind.c b/kernel/unwind.c index 209e248517db..08645aa7c2d6 100644 --- a/kernel/unwind.c +++ b/kernel/unwind.c @@ -164,7 +164,9 @@ static struct unwind_table *find_table(unsigned long pc) static unsigned long read_pointer(const u8 **pLoc, const void *end, - signed ptrType); + signed ptrType, + unsigned long text_base, + unsigned long data_base); static void init_unwind_table(struct unwind_table *table, const char *name, @@ -189,10 +191,13 @@ static void init_unwind_table(struct unwind_table *table, /* See if the linker provided table looks valid. */ if (header_size <= 4 || header_start[0] != 1 - || (void *)read_pointer(&ptr, end, header_start[1]) != table_start - || header_start[2] == DW_EH_PE_omit - || read_pointer(&ptr, end, header_start[2]) <= 0 - || header_start[3] == DW_EH_PE_omit) + || (void *)read_pointer(&ptr, end, header_start[1], 0, 0) + != table_start + || !read_pointer(&ptr, end, header_start[2], 0, 0) + || !read_pointer(&ptr, end, header_start[3], 0, + (unsigned long)header_start) + || !read_pointer(&ptr, end, header_start[3], 0, + (unsigned long)header_start)) header_start = NULL; table->hdrsz = header_size; smp_wmb(); @@ -282,7 +287,7 @@ static void __init setup_unwind_table(struct unwind_table *table, ptr = (const u8 *)(fde + 2); if (!read_pointer(&ptr, (const u8 *)(fde + 1) + *fde, - ptrType)) + ptrType, 0, 0)) return; ++n; } @@ -317,7 +322,7 @@ static void __init setup_unwind_table(struct unwind_table *table, ptr = (const u8 *)(fde + 2); header->table[n].start = read_pointer(&ptr, (const u8 *)(fde + 1) + *fde, - fde_pointer_type(cie)); + fde_pointer_type(cie), 0, 0); header->table[n].fde = (unsigned long)fde; ++n; } @@ -500,7 +505,9 @@ static const u32 *cie_for_fde(const u32 *fde, const struct unwind_table *table) static unsigned long read_pointer(const u8 **pLoc, const void *end, - signed ptrType) + signed ptrType, + unsigned long text_base, + unsigned long data_base) { unsigned long value = 0; union { @@ -572,6 +579,22 @@ static unsigned long read_pointer(const u8 **pLoc, case DW_EH_PE_pcrel: value += (unsigned long)*pLoc; break; + case DW_EH_PE_textrel: + if (likely(text_base)) { + value += text_base; + break; + } + dprintk(2, "Text-relative encoding %02X (%p,%p), but zero text base.", + ptrType, *pLoc, end); + return 0; + case DW_EH_PE_datarel: + if (likely(data_base)) { + value += data_base; + break; + } + dprintk(2, "Data-relative encoding %02X (%p,%p), but zero data base.", + ptrType, *pLoc, end); + return 0; default: dprintk(2, "Cannot adjust pointer type %02X (%p,%p).", ptrType, *pLoc, end); @@ -625,7 +648,8 @@ static signed fde_pointer_type(const u32 *cie) case 'P': { signed ptrType = *ptr++; - if (!read_pointer(&ptr, end, ptrType) || ptr > end) + if (!read_pointer(&ptr, end, ptrType, 0, 0) + || ptr > end) return -1; } break; @@ -685,7 +709,8 @@ static int processCFI(const u8 *start, case DW_CFA_nop: break; case DW_CFA_set_loc: - if ((state->loc = read_pointer(&ptr.p8, end, ptrType)) == 0) + state->loc = read_pointer(&ptr.p8, end, ptrType, 0, 0); + if (state->loc == 0) result = 0; break; case DW_CFA_advance_loc1: @@ -854,9 +879,9 @@ int unwind(struct unwind_frame_info *frame) ptr = hdr + 4; end = hdr + table->hdrsz; if (tableSize - && read_pointer(&ptr, end, hdr[1]) + && read_pointer(&ptr, end, hdr[1], 0, 0) == (unsigned long)table->address - && (i = read_pointer(&ptr, end, hdr[2])) > 0 + && (i = read_pointer(&ptr, end, hdr[2], 0, 0)) > 0 && i == (end - ptr) / (2 * tableSize) && !((end - ptr) % (2 * tableSize))) { do { @@ -864,7 +889,8 @@ int unwind(struct unwind_frame_info *frame) startLoc = read_pointer(&cur, cur + tableSize, - hdr[3]); + hdr[3], 0, + (unsigned long)hdr); if (pc < startLoc) i /= 2; else { @@ -875,11 +901,13 @@ int unwind(struct unwind_frame_info *frame) if (i == 1 && (startLoc = read_pointer(&ptr, ptr + tableSize, - hdr[3])) != 0 + hdr[3], 0, + (unsigned long)hdr)) != 0 && pc >= startLoc) fde = (void *)read_pointer(&ptr, ptr + tableSize, - hdr[3]); + hdr[3], 0, + (unsigned long)hdr); } } if(hdr && !fde) @@ -894,13 +922,13 @@ int unwind(struct unwind_frame_info *frame) && (ptrType = fde_pointer_type(cie)) >= 0 && read_pointer(&ptr, (const u8 *)(fde + 1) + *fde, - ptrType) == startLoc) { + ptrType, 0, 0) == startLoc) { if (!(ptrType & DW_EH_PE_indirect)) ptrType &= DW_EH_PE_FORM|DW_EH_PE_signed; endLoc = startLoc + read_pointer(&ptr, (const u8 *)(fde + 1) + *fde, - ptrType); + ptrType, 0, 0); if(pc >= endLoc) fde = NULL; } else @@ -926,7 +954,7 @@ int unwind(struct unwind_frame_info *frame) ptr = (const u8 *)(fde + 2); startLoc = read_pointer(&ptr, (const u8 *)(fde + 1) + *fde, - ptrType); + ptrType, 0, 0); if (!startLoc) continue; if (!(ptrType & DW_EH_PE_indirect)) @@ -934,7 +962,7 @@ int unwind(struct unwind_frame_info *frame) endLoc = startLoc + read_pointer(&ptr, (const u8 *)(fde + 1) + *fde, - ptrType); + ptrType, 0, 0); if (pc >= startLoc && pc < endLoc) break; } -- cgit v1.2.3 From b65780e123ba9b762276482bbfb52836e4d41fd9 Mon Sep 17 00:00:00 2001 From: Jan Beulich Date: Thu, 7 Dec 2006 02:14:19 +0100 Subject: [PATCH] unwinder: move .eh_frame to RODATA The .eh_frame section contents is never written to, so it can as well benefit from CONFIG_DEBUG_RODATA. Diff-ed against firstfloor tree. Signed-off-by: Jan Beulich Signed-off-by: Andi Kleen --- arch/i386/kernel/vmlinux.lds.S | 9 --------- arch/x86_64/kernel/vmlinux.lds.S | 9 --------- include/asm-generic/vmlinux.lds.h | 17 ++++++++++++----- kernel/unwind.c | 2 +- 4 files changed, 13 insertions(+), 24 deletions(-) (limited to 'kernel') diff --git a/arch/i386/kernel/vmlinux.lds.S b/arch/i386/kernel/vmlinux.lds.S index 25581e87c60d..56e6ad5cb045 100644 --- a/arch/i386/kernel/vmlinux.lds.S +++ b/arch/i386/kernel/vmlinux.lds.S @@ -102,15 +102,6 @@ SECTIONS _edata = .; /* End of data section */ } -#ifdef CONFIG_STACK_UNWIND - . = ALIGN(4); - .eh_frame : AT(ADDR(.eh_frame) - LOAD_OFFSET) { - __start_unwind = .; - *(.eh_frame) - __end_unwind = .; - } -#endif - . = ALIGN(THREAD_SIZE); /* init_task */ .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) { *(.data.init_task) diff --git a/arch/x86_64/kernel/vmlinux.lds.S b/arch/x86_64/kernel/vmlinux.lds.S index d9534e750d4f..6a1f8f491e5d 100644 --- a/arch/x86_64/kernel/vmlinux.lds.S +++ b/arch/x86_64/kernel/vmlinux.lds.S @@ -51,15 +51,6 @@ SECTIONS RODATA -#ifdef CONFIG_STACK_UNWIND - . = ALIGN(8); - .eh_frame : AT(ADDR(.eh_frame) - LOAD_OFFSET) { - __start_unwind = .; - *(.eh_frame) - __end_unwind = .; - } -#endif - . = ALIGN(PAGE_SIZE); /* Align data segment to page size boundary */ /* Data */ .data : AT(ADDR(.data) - LOAD_OFFSET) { diff --git a/include/asm-generic/vmlinux.lds.h b/include/asm-generic/vmlinux.lds.h index 9f4747780dac..4d4c62d11059 100644 --- a/include/asm-generic/vmlinux.lds.h +++ b/include/asm-generic/vmlinux.lds.h @@ -119,8 +119,7 @@ *(__ksymtab_strings) \ } \ \ - /* Unwind data binary search table */ \ - EH_FRAME_HDR \ + EH_FRAME \ \ /* Built-in module parameters. */ \ __param : AT(ADDR(__param) - LOAD_OFFSET) { \ @@ -162,15 +161,23 @@ VMLINUX_SYMBOL(__kprobes_text_end) = .; #ifdef CONFIG_STACK_UNWIND - /* Unwind data binary search table */ -#define EH_FRAME_HDR \ +#define EH_FRAME \ + /* Unwind data binary search table */ \ + . = ALIGN(8); \ .eh_frame_hdr : AT(ADDR(.eh_frame_hdr) - LOAD_OFFSET) { \ VMLINUX_SYMBOL(__start_unwind_hdr) = .; \ *(.eh_frame_hdr) \ VMLINUX_SYMBOL(__end_unwind_hdr) = .; \ + } \ + /* Unwind data */ \ + . = ALIGN(8); \ + .eh_frame : AT(ADDR(.eh_frame) - LOAD_OFFSET) { \ + VMLINUX_SYMBOL(__start_unwind) = .; \ + *(.eh_frame) \ + VMLINUX_SYMBOL(__end_unwind) = .; \ } #else -#define EH_FRAME_HDR +#define EH_FRAME #endif /* DWARF debug sections. diff --git a/kernel/unwind.c b/kernel/unwind.c index 08645aa7c2d6..09c261329249 100644 --- a/kernel/unwind.c +++ b/kernel/unwind.c @@ -19,7 +19,7 @@ #include #include -extern char __start_unwind[], __end_unwind[]; +extern const char __start_unwind[], __end_unwind[]; extern const u8 __start_unwind_hdr[], __end_unwind_hdr[]; #define MAX_STACK_DEPTH 8 -- cgit v1.2.3