From 69111bac42f5ceacdd22e30947837ceb2c4493ed Mon Sep 17 00:00:00 2001 From: Christoph Lameter Date: Tue, 21 Oct 2014 15:23:25 -0500 Subject: powerpc: Replace __get_cpu_var uses This still has not been merged and now powerpc is the only arch that does not have this change. Sorry about missing linuxppc-dev before. V2->V2 - Fix up to work against 3.18-rc1 __get_cpu_var() is used for multiple purposes in the kernel source. One of them is address calculation via the form &__get_cpu_var(x). This calculates the address for the instance of the percpu variable of the current processor based on an offset. Other use cases are for storing and retrieving data from the current processors percpu area. __get_cpu_var() can be used as an lvalue when writing data or on the right side of an assignment. __get_cpu_var() is defined as : __get_cpu_var() always only does an address determination. However, store and retrieve operations could use a segment prefix (or global register on other platforms) to avoid the address calculation. this_cpu_write() and this_cpu_read() can directly take an offset into a percpu area and use optimized assembly code to read and write per cpu variables. This patch converts __get_cpu_var into either an explicit address calculation using this_cpu_ptr() or into a use of this_cpu operations that use the offset. Thereby address calculations are avoided and less registers are used when code is generated. At the end of the patch set all uses of __get_cpu_var have been removed so the macro is removed too. The patch set includes passes over all arches as well. Once these operations are used throughout then specialized macros can be defined in non -x86 arches as well in order to optimize per cpu access by f.e. using a global register that may be set to the per cpu base. Transformations done to __get_cpu_var() 1. Determine the address of the percpu instance of the current processor. DEFINE_PER_CPU(int, y); int *x = &__get_cpu_var(y); Converts to int *x = this_cpu_ptr(&y); 2. Same as #1 but this time an array structure is involved. DEFINE_PER_CPU(int, y[20]); int *x = __get_cpu_var(y); Converts to int *x = this_cpu_ptr(y); 3. Retrieve the content of the current processors instance of a per cpu variable. DEFINE_PER_CPU(int, y); int x = __get_cpu_var(y) Converts to int x = __this_cpu_read(y); 4. Retrieve the content of a percpu struct DEFINE_PER_CPU(struct mystruct, y); struct mystruct x = __get_cpu_var(y); Converts to memcpy(&x, this_cpu_ptr(&y), sizeof(x)); 5. Assignment to a per cpu variable DEFINE_PER_CPU(int, y) __get_cpu_var(y) = x; Converts to __this_cpu_write(y, x); 6. Increment/Decrement etc of a per cpu variable DEFINE_PER_CPU(int, y); __get_cpu_var(y)++ Converts to __this_cpu_inc(y) Cc: Benjamin Herrenschmidt CC: Paul Mackerras Signed-off-by: Christoph Lameter [mpe: Fix build errors caused by set/or_softirq_pending(), and rework assignment in __set_breakpoint() to use memcpy().] Signed-off-by: Michael Ellerman --- arch/powerpc/kernel/process.c | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) (limited to 'arch/powerpc/kernel/process.c') diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c index 923cd2daba89..91e132b495bf 100644 --- a/arch/powerpc/kernel/process.c +++ b/arch/powerpc/kernel/process.c @@ -499,7 +499,7 @@ static inline int set_dawr(struct arch_hw_breakpoint *brk) void __set_breakpoint(struct arch_hw_breakpoint *brk) { - __get_cpu_var(current_brk) = *brk; + memcpy(this_cpu_ptr(¤t_brk), brk, sizeof(*brk)); if (cpu_has_feature(CPU_FTR_DAWR)) set_dawr(brk); @@ -842,7 +842,7 @@ struct task_struct *__switch_to(struct task_struct *prev, * schedule DABR */ #ifndef CONFIG_HAVE_HW_BREAKPOINT - if (unlikely(!hw_brk_match(&__get_cpu_var(current_brk), &new->thread.hw_brk))) + if (unlikely(!hw_brk_match(this_cpu_ptr(¤t_brk), &new->thread.hw_brk))) __set_breakpoint(&new->thread.hw_brk); #endif /* CONFIG_HAVE_HW_BREAKPOINT */ #endif @@ -856,7 +856,7 @@ struct task_struct *__switch_to(struct task_struct *prev, * Collect processor utilization data per process */ if (firmware_has_feature(FW_FEATURE_SPLPAR)) { - struct cpu_usage *cu = &__get_cpu_var(cpu_usage_array); + struct cpu_usage *cu = this_cpu_ptr(&cpu_usage_array); long unsigned start_tb, current_tb; start_tb = old_thread->start_tb; cu->current_tb = current_tb = mfspr(SPRN_PURR); @@ -866,7 +866,7 @@ struct task_struct *__switch_to(struct task_struct *prev, #endif /* CONFIG_PPC64 */ #ifdef CONFIG_PPC_BOOK3S_64 - batch = &__get_cpu_var(ppc64_tlb_batch); + batch = this_cpu_ptr(&ppc64_tlb_batch); if (batch->active) { current_thread_info()->local_flags |= _TLF_LAZY_MMU; if (batch->index) @@ -889,7 +889,7 @@ struct task_struct *__switch_to(struct task_struct *prev, #ifdef CONFIG_PPC_BOOK3S_64 if (current_thread_info()->local_flags & _TLF_LAZY_MMU) { current_thread_info()->local_flags &= ~_TLF_LAZY_MMU; - batch = &__get_cpu_var(ppc64_tlb_batch); + batch = this_cpu_ptr(&ppc64_tlb_batch); batch->active = 1; } #endif /* CONFIG_PPC_BOOK3S_64 */ -- cgit v1.2.3-70-g09d2 From 7b051f665c32db3339ac9b6d9806ef508d09647f Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Mon, 13 Oct 2014 20:27:15 +1100 Subject: powerpc: Use probe_kernel_address in show_instructions We really don't want to take a pagefault in show_instructions, so use probe_kernel_address instead of __get_user. Signed-off-by: Anton Blanchard Signed-off-by: Michael Ellerman --- arch/powerpc/kernel/process.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) (limited to 'arch/powerpc/kernel/process.c') diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c index 91e132b495bf..8c2691e445bd 100644 --- a/arch/powerpc/kernel/process.c +++ b/arch/powerpc/kernel/process.c @@ -37,9 +37,9 @@ #include #include #include +#include #include -#include #include #include #include @@ -921,12 +921,8 @@ static void show_instructions(struct pt_regs *regs) pc = (unsigned long)phys_to_virt(pc); #endif - /* We use __get_user here *only* to avoid an OOPS on a - * bad address because the pc *should* only be a - * kernel address. - */ if (!__kernel_text_address(pc) || - __get_user(instr, (unsigned int __user *)pc)) { + probe_kernel_address((unsigned int __user *)pc, instr)) { printk(KERN_CONT "XXXXXXXX "); } else { if (regs->nip == pc) -- cgit v1.2.3-70-g09d2 From 7d56c65a6ff9065c459fc63c509950d8ea66e00c Mon Sep 17 00:00:00 2001 From: Anton Blanchard Date: Wed, 17 Sep 2014 17:07:03 +1000 Subject: powerpc/ftrace: Remove mod_return_to_handler mod_return_to_handler is the same as return_to_handler, except it handles the change of the TOC (r2). Add this into return_to_handler and remove mod_return_to_handler. Signed-off-by: Anton Blanchard Signed-off-by: Michael Ellerman --- arch/powerpc/kernel/entry_64.S | 24 +----------------------- arch/powerpc/kernel/ftrace.c | 14 ++------------ arch/powerpc/kernel/process.c | 9 +-------- 3 files changed, 4 insertions(+), 43 deletions(-) (limited to 'arch/powerpc/kernel/process.c') diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S index 0905c8da90f1..66f7ed32848f 100644 --- a/arch/powerpc/kernel/entry_64.S +++ b/arch/powerpc/kernel/entry_64.S @@ -1240,28 +1240,6 @@ _GLOBAL(ftrace_graph_caller) blr _GLOBAL(return_to_handler) - /* need to save return values */ - std r4, -24(r1) - std r3, -16(r1) - std r31, -8(r1) - mr r31, r1 - stdu r1, -112(r1) - - bl ftrace_return_to_handler - nop - - /* return value has real return address */ - mtlr r3 - - ld r1, 0(r1) - ld r4, -24(r1) - ld r3, -16(r1) - ld r31, -8(r1) - - /* Jump back to real return address */ - blr - -_GLOBAL(mod_return_to_handler) /* need to save return values */ std r4, -32(r1) std r3, -24(r1) @@ -1272,7 +1250,7 @@ _GLOBAL(mod_return_to_handler) stdu r1, -112(r1) /* - * We are in a module using the module's TOC. + * We might be called from a module. * Switch to our TOC to run inside the core kernel. */ ld r2, PACATOC(r13) diff --git a/arch/powerpc/kernel/ftrace.c b/arch/powerpc/kernel/ftrace.c index 390311c0f03d..abf79213c356 100644 --- a/arch/powerpc/kernel/ftrace.c +++ b/arch/powerpc/kernel/ftrace.c @@ -510,10 +510,6 @@ int ftrace_disable_ftrace_graph_caller(void) } #endif /* CONFIG_DYNAMIC_FTRACE */ -#ifdef CONFIG_PPC64 -extern void mod_return_to_handler(void); -#endif - /* * Hook the return address and push it in the stack of return addrs * in current thread info. @@ -523,7 +519,7 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) unsigned long old; int faulted; struct ftrace_graph_ent trace; - unsigned long return_hooker = (unsigned long)&return_to_handler; + unsigned long return_hooker; if (unlikely(ftrace_graph_is_dead())) return; @@ -531,13 +527,7 @@ void prepare_ftrace_return(unsigned long *parent, unsigned long self_addr) if (unlikely(atomic_read(¤t->tracing_graph_pause))) return; -#ifdef CONFIG_PPC64 - /* non core kernel code needs to save and restore the TOC */ - if (REGION_ID(self_addr) != KERNEL_REGION_ID) - return_hooker = (unsigned long)&mod_return_to_handler; -#endif - - return_hooker = ppc_function_entry((void *)return_hooker); + return_hooker = ppc_function_entry(return_to_handler); /* * Protect against fault, even if it shouldn't diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c index 8c2691e445bd..f6b82152e7aa 100644 --- a/arch/powerpc/kernel/process.c +++ b/arch/powerpc/kernel/process.c @@ -1527,13 +1527,6 @@ void show_stack(struct task_struct *tsk, unsigned long *stack) int curr_frame = current->curr_ret_stack; extern void return_to_handler(void); unsigned long rth = (unsigned long)return_to_handler; - unsigned long mrth = -1; -#ifdef CONFIG_PPC64 - extern void mod_return_to_handler(void); - rth = *(unsigned long *)rth; - mrth = (unsigned long)mod_return_to_handler; - mrth = *(unsigned long *)mrth; -#endif #endif sp = (unsigned long) stack; @@ -1558,7 +1551,7 @@ void show_stack(struct task_struct *tsk, unsigned long *stack) if (!firstframe || ip != lr) { printk("["REG"] ["REG"] %pS", sp, ip, (void *)ip); #ifdef CONFIG_FUNCTION_GRAPH_TRACER - if ((ip == rth || ip == mrth) && curr_frame >= 0) { + if ((ip == rth) && curr_frame >= 0) { printk(" (%pS)", (void *)current->ret_stack[curr_frame].ret); curr_frame--; -- cgit v1.2.3-70-g09d2 From 59994fb01a102a448ba758c9b824a29b4a99cc1b Mon Sep 17 00:00:00 2001 From: Vineeth Vijayan Date: Fri, 14 Nov 2014 14:42:05 +0530 Subject: powerpc: Use generic PIE randomization Back in 2009 we merged 501cb16d3cfd "Randomise PIEs", which added support for randomizing PIE (Position Independent Executable) binaries. That commit added randomize_et_dyn(), which correctly randomized the addresses, but failed to honor PF_RANDOMIZE. That means it was not possible to disable PIE randomization via the personality flag, or /proc/sys/kernel/randomize_va_space. Since then there has been generic support for PIE randomization added to binfmt_elf.c, selectable via ARCH_BINFMT_ELF_RANDOMIZE_PIE. Enabling that allows us to drop randomize_et_dyn(), which means we start honoring PF_RANDOMIZE correctly. It also causes a fairly major change to how we layout PIE binaries. Currently we will place the binary at 512MB-520MB for 32 bit binaries, or 512MB-1.5GB for 64 bit binaries, eg: $ cat /proc/$$/maps 4e550000-4e580000 r-xp 00000000 08:02 129813 /bin/dash 4e580000-4e590000 rw-p 00020000 08:02 129813 /bin/dash 10014110000-10014140000 rw-p 00000000 00:00 0 [heap] 3fffaa3f0000-3fffaa5a0000 r-xp 00000000 08:02 921 /lib/powerpc64le-linux-gnu/libc-2.19.so 3fffaa5a0000-3fffaa5b0000 rw-p 001a0000 08:02 921 /lib/powerpc64le-linux-gnu/libc-2.19.so 3fffaa5c0000-3fffaa5d0000 rw-p 00000000 00:00 0 3fffaa5d0000-3fffaa5f0000 r-xp 00000000 00:00 0 [vdso] 3fffaa5f0000-3fffaa620000 r-xp 00000000 08:02 1246 /lib/powerpc64le-linux-gnu/ld-2.19.so 3fffaa620000-3fffaa630000 rw-p 00020000 08:02 1246 /lib/powerpc64le-linux-gnu/ld-2.19.so 3ffffc340000-3ffffc370000 rw-p 00000000 00:00 0 [stack] With this commit applied we don't do any special randomisation for the binary, and instead rely on mmap randomisation. This means the binary ends up at high addresses, eg: $ cat /proc/$$/maps 3fff99820000-3fff999d0000 r-xp 00000000 08:02 921 /lib/powerpc64le-linux-gnu/libc-2.19.so 3fff999d0000-3fff999e0000 rw-p 001a0000 08:02 921 /lib/powerpc64le-linux-gnu/libc-2.19.so 3fff999f0000-3fff99a00000 rw-p 00000000 00:00 0 3fff99a00000-3fff99a20000 r-xp 00000000 00:00 0 [vdso] 3fff99a20000-3fff99a50000 r-xp 00000000 08:02 1246 /lib/powerpc64le-linux-gnu/ld-2.19.so 3fff99a50000-3fff99a60000 rw-p 00020000 08:02 1246 /lib/powerpc64le-linux-gnu/ld-2.19.so 3fff99a60000-3fff99a90000 r-xp 00000000 08:02 129813 /bin/dash 3fff99a90000-3fff99aa0000 rw-p 00020000 08:02 129813 /bin/dash 3fffc3de0000-3fffc3e10000 rw-p 00000000 00:00 0 [stack] 3fffc55e0000-3fffc5610000 rw-p 00000000 00:00 0 [heap] Although this should be OK, it's possible it might break badly written binaries that make assumptions about the address space layout. Signed-off-by: Vineeth Vijayan [mpe: Rewrite changelog] Signed-off-by: Michael Ellerman --- arch/powerpc/Kconfig | 1 + arch/powerpc/include/asm/elf.h | 3 +-- arch/powerpc/kernel/process.c | 9 --------- 3 files changed, 2 insertions(+), 11 deletions(-) (limited to 'arch/powerpc/kernel/process.c') diff --git a/arch/powerpc/Kconfig b/arch/powerpc/Kconfig index 46227336aacb..421df5cc6d18 100644 --- a/arch/powerpc/Kconfig +++ b/arch/powerpc/Kconfig @@ -88,6 +88,7 @@ config PPC select ARCH_MIGHT_HAVE_PC_PARPORT select ARCH_MIGHT_HAVE_PC_SERIO select BINFMT_ELF + select ARCH_BINFMT_ELF_RANDOMIZE_PIE select OF select OF_EARLY_FLATTREE select OF_RESERVED_MEM diff --git a/arch/powerpc/include/asm/elf.h b/arch/powerpc/include/asm/elf.h index 888d8f3f2524..57d289acb803 100644 --- a/arch/powerpc/include/asm/elf.h +++ b/arch/powerpc/include/asm/elf.h @@ -28,8 +28,7 @@ the loader. We need to make sure that it is out of the way of the program that it will "exec", and that there is sufficient room for the brk. */ -extern unsigned long randomize_et_dyn(unsigned long base); -#define ELF_ET_DYN_BASE (randomize_et_dyn(0x20000000)) +#define ELF_ET_DYN_BASE 0x20000000 #define ELF_CORE_EFLAGS (is_elf2_task() ? 2 : 0) diff --git a/arch/powerpc/kernel/process.c b/arch/powerpc/kernel/process.c index f6b82152e7aa..b4cc7bef6b16 100644 --- a/arch/powerpc/kernel/process.c +++ b/arch/powerpc/kernel/process.c @@ -1654,12 +1654,3 @@ unsigned long arch_randomize_brk(struct mm_struct *mm) return ret; } -unsigned long randomize_et_dyn(unsigned long base) -{ - unsigned long ret = PAGE_ALIGN(base + brk_rnd()); - - if (ret < base) - return base; - - return ret; -} -- cgit v1.2.3-70-g09d2