Skip to content

L
linux
Commit 2d4a7167 authored by Ingo Molnar's avatar Ingo Molnar
Browse files
Options

x86, mm: fault.c cleanup 

Impact: cleanup, no code changed

Clean up various small details, which can be correctness checked
automatically:

 - tidy up the include file section
 - eliminate unnecessary includes
 - introduce show_signal_msg() to clean up code flow
 - standardize the code flow
 - standardize comments and other style details
 - more cleanups, pointed out by checkpatch

No code changed on either 32-bit nor 64-bit:

arch/x86/mm/fault.o:

   text	   data	    bss	    dec	    hex	filename
   4632	     32	     24	   4688	   1250	fault.o.before
   4632	     32	     24	   4688	   1250	fault.o.after

the md5 changed due to a change in a single instruction:

   2e8a8241e7f0d69706776a5a26c90bc0  fault.o.before.asm
   c5c3d36e725586eb74f0e10692f0193e  fault.o.after.asm

Because a __LINE__ reference in a WARN_ONCE() has changed.

On 32-bit a few stack offsets changed - no code size difference
nor any functionality difference.

Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: default avatarIngo Molnar <mingo@elte.hu>
parent c9e1585b
Show whitespace changes
Inline Side-by-side
Showing with 331 additions and 215 deletions
arch/x86/mm/fault.c
View file @ 2d4a7167
/* /*
* Copyright (C) 1995 Linus Torvalds * Copyright (C) 1995 Linus Torvalds
* Copyright (C) 2001,2002 Andi Kleen, SuSE Labs. * Copyright (C) 2001, 2002 Andi Kleen, SuSE Labs.
*/ */
#include <linux/signal.h>
#include <linux/sched.h>
#include <linux/kernel.h>
#include <linux/errno.h>
#include <linux/string.h>
#include <linux/types.h>
#include <linux/ptrace.h>
#include <linux/mmiotrace.h>
#include <linux/mman.h>
#include <linux/mm.h>
#include <linux/smp.h>
#include <linux/interrupt.h> #include <linux/interrupt.h>
#include <linux/init.h> #include <linux/mmiotrace.h>
#include <linux/tty.h> #include <linux/bootmem.h>
#include <linux/vt_kern.h> /* For unblank_screen() */
#include <linux/compiler.h> #include <linux/compiler.h>
#include <linux/highmem.h> #include <linux/highmem.h>
#include <linux/bootmem.h> /* for max_low_pfn */
#include <linux/vmalloc.h>
#include <linux/module.h>
#include <linux/kprobes.h> #include <linux/kprobes.h>
#include <linux/uaccess.h> #include <linux/uaccess.h>
#include <linux/vmalloc.h>
#include <linux/vt_kern.h>
#include <linux/signal.h>
#include <linux/kernel.h>
#include <linux/ptrace.h>
#include <linux/string.h>
#include <linux/module.h>
#include <linux/kdebug.h> #include <linux/kdebug.h>
#include <linux/errno.h>
#include <linux/magic.h> #include <linux/magic.h>
#include <linux/sched.h>
#include <linux/types.h>
#include <linux/init.h>
#include <linux/mman.h>
#include <linux/tty.h>
#include <linux/smp.h>
#include <linux/mm.h>
#include <asm-generic/sections.h>
#include <asm/system.h>
#include <asm/desc.h>
#include <asm/segment.h>
#include <asm/pgalloc.h>
#include <asm/smp.h>
#include <asm/tlbflush.h> #include <asm/tlbflush.h>
#include <asm/pgalloc.h>
#include <asm/segment.h>
#include <asm/system.h>
#include <asm/proto.h> #include <asm/proto.h>
#include <asm-generic/sections.h>
#include <asm/traps.h> #include <asm/traps.h>
#include <asm/desc.h>
/* /*
* Page fault error code bits * Page fault error code bits:
* bit 0 == 0 means no page found, 1 means protection fault *
* bit 1 == 0 means read, 1 means write * bit 0 == 0: no page found 1: protection fault
* bit 2 == 0 means kernel, 1 means user-mode * bit 1 == 0: read access 1: write access
* bit 3 == 1 means use of reserved bit detected * bit 2 == 0: kernel-mode access 1: user-mode access
* bit 4 == 1 means fault was an instruction fetch * bit 3 == 1: use of reserved bit detected
* bit 4 == 1: fault was an instruction fetch
*/ */
#define PF_PROT (1<<0) enum x86_pf_error_code {
#define PF_WRITE (1<<1)
#define PF_USER (1<<2) PF_PROT = 1 << 0,
#define PF_RSVD (1<<3) PF_WRITE = 1 << 1,
#define PF_INSTR (1<<4) PF_USER = 1 << 2,
PF_RSVD = 1 << 3,
PF_INSTR = 1 << 4,
};
static inline int kmmio_fault(struct pt_regs *regs, unsigned long addr) static inline int kmmio_fault(struct pt_regs *regs, unsigned long addr)
{ {
...@@ -82,23 +85,27 @@ static inline int notify_page_fault(struct pt_regs *regs) ...@@ -82,23 +85,27 @@ static inline int notify_page_fault(struct pt_regs *regs)
} }
/* /*
* X86_32 * Prefetch quirks:
*
* 32-bit mode:
*
* Sometimes AMD Athlon/Opteron CPUs report invalid exceptions on prefetch. * Sometimes AMD Athlon/Opteron CPUs report invalid exceptions on prefetch.
* Check that here and ignore it. * Check that here and ignore it.
* *
* X86_64 * 64-bit mode:
*
* Sometimes the CPU reports invalid exceptions on prefetch. * Sometimes the CPU reports invalid exceptions on prefetch.
* Check that here and ignore it. * Check that here and ignore it.
* *
* Opcode checker based on code by Richard Brunner * Opcode checker based on code by Richard Brunner.
*/ */
static int is_prefetch(struct pt_regs *regs, unsigned long error_code, static int
unsigned long addr) is_prefetch(struct pt_regs *regs, unsigned long error_code, unsigned long addr)
{ {
unsigned char *max_instr;
unsigned char *instr; unsigned char *instr;
int scan_more = 1; int scan_more = 1;
int prefetch = 0; int prefetch = 0;
unsigned char *max_instr;
/* /*
* If it was a exec (instruction fetch) fault on NX page, then * If it was a exec (instruction fetch) fault on NX page, then
...@@ -114,9 +121,9 @@ static int is_prefetch(struct pt_regs *regs, unsigned long error_code, ...@@ -114,9 +121,9 @@ static int is_prefetch(struct pt_regs *regs, unsigned long error_code,
return 0; return 0;
while (scan_more && instr < max_instr) { while (scan_more && instr < max_instr) {
unsigned char opcode;
unsigned char instr_hi; unsigned char instr_hi;
unsigned char instr_lo; unsigned char instr_lo;
unsigned char opcode;
if (probe_kernel_address(instr, opcode)) if (probe_kernel_address(instr, opcode))
break; break;
...@@ -173,8 +180,9 @@ static int is_prefetch(struct pt_regs *regs, unsigned long error_code, ...@@ -173,8 +180,9 @@ static int is_prefetch(struct pt_regs *regs, unsigned long error_code,
return prefetch; return prefetch;
} }
static void force_sig_info_fault(int si_signo, int si_code, static void
unsigned long address, struct task_struct *tsk) force_sig_info_fault(int si_signo, int si_code, unsigned long address,
struct task_struct *tsk)
{ {
siginfo_t info; siginfo_t info;
...@@ -182,6 +190,7 @@ static void force_sig_info_fault(int si_signo, int si_code, ...@@ -182,6 +190,7 @@ static void force_sig_info_fault(int si_signo, int si_code,
info.si_errno = 0; info.si_errno = 0;
info.si_code = si_code; info.si_code = si_code;
info.si_addr = (void __user *)address; info.si_addr = (void __user *)address;
force_sig_info(si_signo, &info, tsk); force_sig_info(si_signo, &info, tsk);
} }
...@@ -189,6 +198,7 @@ static void force_sig_info_fault(int si_signo, int si_code, ...@@ -189,6 +198,7 @@ static void force_sig_info_fault(int si_signo, int si_code,
static int bad_address(void *p) static int bad_address(void *p)
{ {
unsigned long dummy; unsigned long dummy;
return probe_kernel_address((unsigned long *)p, dummy); return probe_kernel_address((unsigned long *)p, dummy);
} }
#endif #endif
...@@ -200,6 +210,7 @@ static void dump_pagetable(unsigned long address) ...@@ -200,6 +210,7 @@ static void dump_pagetable(unsigned long address)
page = read_cr3(); page = read_cr3();
page = ((__typeof__(page) *) __va(page))[address >> PGDIR_SHIFT]; page = ((__typeof__(page) *) __va(page))[address >> PGDIR_SHIFT];
#ifdef CONFIG_X86_PAE #ifdef CONFIG_X86_PAE
printk("*pdpt = %016Lx ", page); printk("*pdpt = %016Lx ", page);
if ((page >> PAGE_SHIFT) < max_low_pfn if ((page >> PAGE_SHIFT) < max_low_pfn
...@@ -218,11 +229,12 @@ static void dump_pagetable(unsigned long address) ...@@ -218,11 +229,12 @@ static void dump_pagetable(unsigned long address)
* We must not directly access the pte in the highpte * We must not directly access the pte in the highpte
* case if the page table is located in highmem. * case if the page table is located in highmem.
* And let's rather not kmap-atomic the pte, just in case * And let's rather not kmap-atomic the pte, just in case
* it's allocated already. * it's allocated already:
*/ */
if ((page >> PAGE_SHIFT) < max_low_pfn if ((page >> PAGE_SHIFT) < max_low_pfn
&& (page & _PAGE_PRESENT) && (page & _PAGE_PRESENT)
&& !(page & _PAGE_PSE)) { && !(page & _PAGE_PSE)) {
page &= PAGE_MASK; page &= PAGE_MASK;
page = ((__typeof__(page) *) __va(page))[(address >> PAGE_SHIFT) page = ((__typeof__(page) *) __va(page))[(address >> PAGE_SHIFT)
& (PTRS_PER_PTE - 1)]; & (PTRS_PER_PTE - 1)];
...@@ -239,26 +251,38 @@ static void dump_pagetable(unsigned long address) ...@@ -239,26 +251,38 @@ static void dump_pagetable(unsigned long address)
pgd = (pgd_t *)read_cr3(); pgd = (pgd_t *)read_cr3();
pgd = __va((unsigned long)pgd & PHYSICAL_PAGE_MASK); pgd = __va((unsigned long)pgd & PHYSICAL_PAGE_MASK);
pgd += pgd_index(address); pgd += pgd_index(address);
if (bad_address(pgd)) goto bad; if (bad_address(pgd))
goto bad;
printk("PGD %lx ", pgd_val(*pgd)); printk("PGD %lx ", pgd_val(*pgd));
if (!pgd_present(*pgd)) goto ret;
if (!pgd_present(*pgd))
goto out;
pud = pud_offset(pgd, address); pud = pud_offset(pgd, address);
if (bad_address(pud)) goto bad; if (bad_address(pud))
goto bad;
printk("PUD %lx ", pud_val(*pud)); printk("PUD %lx ", pud_val(*pud));
if (!pud_present(*pud) || pud_large(*pud)) if (!pud_present(*pud) || pud_large(*pud))
goto ret; goto out;
pmd = pmd_offset(pud, address); pmd = pmd_offset(pud, address);
if (bad_address(pmd)) goto bad; if (bad_address(pmd))
goto bad;
printk("PMD %lx ", pmd_val(*pmd)); printk("PMD %lx ", pmd_val(*pmd));
if (!pmd_present(*pmd) || pmd_large(*pmd)) goto ret; if (!pmd_present(*pmd) || pmd_large(*pmd))
goto out;
pte = pte_offset_kernel(pmd, address); pte = pte_offset_kernel(pmd, address);
if (bad_address(pte)) goto bad; if (bad_address(pte))
goto bad;
printk("PTE %lx", pte_val(*pte)); printk("PTE %lx", pte_val(*pte));
ret: out:
printk("\n"); printk("\n");
return; return;
bad: bad:
...@@ -285,7 +309,6 @@ static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address) ...@@ -285,7 +309,6 @@ static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address)
* and redundant with the set_pmd() on non-PAE. As would * and redundant with the set_pmd() on non-PAE. As would
* set_pud. * set_pud.
*/ */
pud = pud_offset(pgd, address); pud = pud_offset(pgd, address);
pud_k = pud_offset(pgd_k, address); pud_k = pud_offset(pgd_k, address);
if (!pud_present(*pud_k)) if (!pud_present(*pud_k))
...@@ -295,11 +318,14 @@ static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address) ...@@ -295,11 +318,14 @@ static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address)
pmd_k = pmd_offset(pud_k, address); pmd_k = pmd_offset(pud_k, address);
if (!pmd_present(*pmd_k)) if (!pmd_present(*pmd_k))
return NULL; return NULL;
if (!pmd_present(*pmd)) { if (!pmd_present(*pmd)) {
set_pmd(pmd, *pmd_k); set_pmd(pmd, *pmd_k);
arch_flush_lazy_mmu_mode(); arch_flush_lazy_mmu_mode();
} else } else {
BUG_ON(pmd_page(*pmd) != pmd_page(*pmd_k)); BUG_ON(pmd_page(*pmd) != pmd_page(*pmd_k));
}
return pmd_k; return pmd_k;
} }
#endif #endif
...@@ -312,29 +338,37 @@ KERN_ERR "******* Please consider a BIOS update.\n" ...@@ -312,29 +338,37 @@ KERN_ERR "******* Please consider a BIOS update.\n"
KERN_ERR "******* Disabling USB legacy in the BIOS may also help.\n"; KERN_ERR "******* Disabling USB legacy in the BIOS may also help.\n";
#endif #endif
/* Workaround for K8 erratum #93 & buggy BIOS. /*
BIOS SMM functions are required to use a specific workaround * Workaround for K8 erratum #93 & buggy BIOS.
to avoid corruption of the 64bit RIP register on C stepping K8. *
A lot of BIOS that didn't get tested properly miss this. * BIOS SMM functions are required to use a specific workaround
The OS sees this as a page fault with the upper 32bits of RIP cleared. * to avoid corruption of the 64bit RIP register on C stepping K8.
Try to work around it here. *
Note we only handle faults in kernel here. * A lot of BIOS that didn't get tested properly miss this.
Does nothing for X86_32 *
* The OS sees this as a page fault with the upper 32bits of RIP cleared.
* Try to work around it here.
*
* Note we only handle faults in kernel here.
* Does nothing on 32-bit.
*/ */
static int is_errata93(struct pt_regs *regs, unsigned long address) static int is_errata93(struct pt_regs *regs, unsigned long address)
{ {
#ifdef CONFIG_X86_64 #ifdef CONFIG_X86_64
static int warned; static int once;
if (address != regs->ip) if (address != regs->ip)
return 0; return 0;
if ((address >> 32) != 0) if ((address >> 32) != 0)
return 0; return 0;
address |= 0xffffffffUL << 32; address |= 0xffffffffUL << 32;
if ((address >= (u64)_stext && address <= (u64)_etext) || if ((address >= (u64)_stext && address <= (u64)_etext) ||
(address >= MODULES_VADDR && address <= MODULES_END)) { (address >= MODULES_VADDR && address <= MODULES_END)) {
if (!warned) { if (!once) {
printk(errata93_warning); printk(errata93_warning);
warned = 1; once = 1;
} }
regs->ip = address; regs->ip = address;
return 1; return 1;
...@@ -344,16 +378,17 @@ static int is_errata93(struct pt_regs *regs, unsigned long address) ...@@ -344,16 +378,17 @@ static int is_errata93(struct pt_regs *regs, unsigned long address)
} }
/* /*
* Work around K8 erratum #100 K8 in compat mode occasionally jumps to illegal * Work around K8 erratum #100 K8 in compat mode occasionally jumps
* addresses >4GB. We catch this in the page fault handler because these * to illegal addresses >4GB.
* addresses are not reachable. Just detect this case and return. Any code *
* We catch this in the page fault handler because these addresses
* are not reachable. Just detect this case and return. Any code
* segment in LDT is compatibility mode. * segment in LDT is compatibility mode.
*/ */
static int is_errata100(struct pt_regs *regs, unsigned long address) static int is_errata100(struct pt_regs *regs, unsigned long address)
{ {
#ifdef CONFIG_X86_64 #ifdef CONFIG_X86_64
if ((regs->cs == __USER32_CS || (regs->cs & (1<<2))) && if ((regs->cs == __USER32_CS || (regs->cs & (1<<2))) && (address >> 32))
(address >> 32))
return 1; return 1;
#endif #endif
return 0; return 0;
...@@ -363,8 +398,9 @@ static int is_f00f_bug(struct pt_regs *regs, unsigned long address) ...@@ -363,8 +398,9 @@ static int is_f00f_bug(struct pt_regs *regs, unsigned long address)
{ {
#ifdef CONFIG_X86_F00F_BUG #ifdef CONFIG_X86_F00F_BUG
unsigned long nr; unsigned long nr;
/* /*
* Pentium F0 0F C7 C8 bug workaround. * Pentium F0 0F C7 C8 bug workaround:
*/ */
if (boot_cpu_data.f00f_bug) { if (boot_cpu_data.f00f_bug) {
nr = (address - idt_descr.address) >> 3; nr = (address - idt_descr.address) >> 3;
...@@ -378,7 +414,8 @@ static int is_f00f_bug(struct pt_regs *regs, unsigned long address) ...@@ -378,7 +414,8 @@ static int is_f00f_bug(struct pt_regs *regs, unsigned long address)
return 0; return 0;
} }
static void show_fault_oops(struct pt_regs *regs, unsigned long error_code, static void
show_fault_oops(struct pt_regs *regs, unsigned long error_code,
unsigned long address) unsigned long address)
{ {
#ifdef CONFIG_X86_32 #ifdef CONFIG_X86_32
...@@ -389,13 +426,15 @@ static void show_fault_oops(struct pt_regs *regs, unsigned long error_code, ...@@ -389,13 +426,15 @@ static void show_fault_oops(struct pt_regs *regs, unsigned long error_code,
#ifdef CONFIG_X86_PAE #ifdef CONFIG_X86_PAE
if (error_code & PF_INSTR) { if (error_code & PF_INSTR) {
unsigned int level; unsigned int level;
pte_t *pte = lookup_address(address, &level); pte_t *pte = lookup_address(address, &level);
if (pte && pte_present(*pte) && !pte_exec(*pte)) if (pte && pte_present(*pte) && !pte_exec(*pte)) {
printk(KERN_CRIT "kernel tried to execute " printk(KERN_CRIT "kernel tried to execute "
"NX-protected page - exploit attempt? " "NX-protected page - exploit attempt? "
"(uid: %d)\n", current_uid()); "(uid: %d)\n", current_uid());
} }
}
#endif #endif
printk(KERN_ALERT "BUG: unable to handle kernel "); printk(KERN_ALERT "BUG: unable to handle kernel ");
...@@ -403,34 +442,45 @@ static void show_fault_oops(struct pt_regs *regs, unsigned long error_code, ...@@ -403,34 +442,45 @@ static void show_fault_oops(struct pt_regs *regs, unsigned long error_code,
printk(KERN_CONT "NULL pointer dereference"); printk(KERN_CONT "NULL pointer dereference");
else else
printk(KERN_CONT "paging request"); printk(KERN_CONT "paging request");
printk(KERN_CONT " at %p\n", (void *) address); printk(KERN_CONT " at %p\n", (void *) address);
printk(KERN_ALERT "IP:"); printk(KERN_ALERT "IP:");
printk_address(regs->ip, 1); printk_address(regs->ip, 1);
dump_pagetable(address); dump_pagetable(address);
} }
#ifdef CONFIG_X86_64 #ifdef CONFIG_X86_64
static noinline void pgtable_bad(struct pt_regs *regs, static noinline void
unsigned long error_code, unsigned long address) pgtable_bad(struct pt_regs *regs, unsigned long error_code,
unsigned long address)
{ {
unsigned long flags = oops_begin(); struct task_struct *tsk;
int sig = SIGKILL; unsigned long flags;
struct task_struct *tsk = current; int sig;
flags = oops_begin();
tsk = current;
sig = SIGKILL;
printk(KERN_ALERT "%s: Corrupted page table at address %lx\n", printk(KERN_ALERT "%s: Corrupted page table at address %lx\n",
tsk->comm, address); tsk->comm, address);
dump_pagetable(address); dump_pagetable(address);
tsk->thread.cr2 = address; tsk->thread.cr2 = address;
tsk->thread.trap_no = 14; tsk->thread.trap_no = 14;
tsk->thread.error_code = error_code; tsk->thread.error_code = error_code;
if (__die("Bad pagetable", regs, error_code)) if (__die("Bad pagetable", regs, error_code))
sig = 0; sig = 0;
oops_end(flags, regs, sig); oops_end(flags, regs, sig);
} }
#endif #endif
static noinline void no_context(struct pt_regs *regs, static noinline void
unsigned long error_code, unsigned long address) no_context(struct pt_regs *regs, unsigned long error_code,
unsigned long address)
{ {
struct task_struct *tsk = current; struct task_struct *tsk = current;
unsigned long *stackend; unsigned long *stackend;
...@@ -445,12 +495,14 @@ static noinline void no_context(struct pt_regs *regs, ...@@ -445,12 +495,14 @@ static noinline void no_context(struct pt_regs *regs,
return; return;
/* /*
* X86_32 * 32-bit:
*
* Valid to do another page fault here, because if this fault * Valid to do another page fault here, because if this fault
* had been triggered by is_prefetch fixup_exception would have * had been triggered by is_prefetch fixup_exception would have
* handled it. * handled it.
* *
* X86_64 * 64-bit:
*
* Hall of shame of CPU/BIOS bugs. * Hall of shame of CPU/BIOS bugs.
*/ */
if (is_prefetch(regs, error_code, address)) if (is_prefetch(regs, error_code, address))
...@@ -461,7 +513,7 @@ static noinline void no_context(struct pt_regs *regs, ...@@ -461,7 +513,7 @@ static noinline void no_context(struct pt_regs *regs,
/* /*
* Oops. The kernel tried to access some bad page. We'll have to * Oops. The kernel tried to access some bad page. We'll have to
* terminate things with extreme prejudice. * terminate things with extreme prejudice:
*/ */
#ifdef CONFIG_X86_32 #ifdef CONFIG_X86_32
bust_spinlocks(1); bust_spinlocks(1);
...@@ -487,28 +539,54 @@ static noinline void no_context(struct pt_regs *regs, ...@@ -487,28 +539,54 @@ static noinline void no_context(struct pt_regs *regs,
sig = SIGKILL; sig = SIGKILL;
if (__die("Oops", regs, error_code)) if (__die("Oops", regs, error_code))
sig = 0; sig = 0;
/* Executive summary in case the body of the oops scrolled away */ /* Executive summary in case the body of the oops scrolled away */
printk(KERN_EMERG "CR2: %016lx\n", address); printk(KERN_EMERG "CR2: %016lx\n", address);
oops_end(flags, regs, sig); oops_end(flags, regs, sig);
#endif #endif
} }
static void __bad_area_nosemaphore(struct pt_regs *regs, /*
unsigned long error_code, unsigned long address, * Print out info about fatal segfaults, if the show_unhandled_signals
int si_code) * sysctl is set:
*/
static inline void
show_signal_msg(struct pt_regs *regs, unsigned long error_code,
unsigned long address, struct task_struct *tsk)
{
if (!unhandled_signal(tsk, SIGSEGV))
return;
if (!printk_ratelimit())
return;
printk(KERN_CONT "%s%s[%d]: segfault at %lx ip %p sp %p error %lx",
task_pid_nr(tsk) > 1 ? KERN_INFO : KERN_EMERG,
tsk->comm, task_pid_nr(tsk), address,
(void *)regs->ip, (void *)regs->sp, error_code);
print_vma_addr(KERN_CONT " in ", regs->ip);
printk(KERN_CONT "\n");
}
static void
__bad_area_nosemaphore(struct pt_regs *regs, unsigned long error_code,
unsigned long address, int si_code)
{ {
struct task_struct *tsk = current; struct task_struct *tsk = current;
/* User mode accesses just cause a SIGSEGV */ /* User mode accesses just cause a SIGSEGV */
if (error_code & PF_USER) { if (error_code & PF_USER) {
/* /*
* It's possible to have interrupts off here. * It's possible to have interrupts off here:
*/ */
local_irq_enable(); local_irq_enable();
/* /*
* Valid to do another page fault here because this one came * Valid to do another page fault here because this one came
* from user space. * from user space:
*/ */
if (is_prefetch(regs, error_code, address)) if (is_prefetch(regs, error_code, address))
return; return;
...@@ -516,22 +594,16 @@ static void __bad_area_nosemaphore(struct pt_regs *regs, ...@@ -516,22 +594,16 @@ static void __bad_area_nosemaphore(struct pt_regs *regs,
if (is_errata100(regs, address)) if (is_errata100(regs, address))
return; return;
if (show_unhandled_signals && unhandled_signal(tsk, SIGSEGV) && if (unlikely(show_unhandled_signals))
printk_ratelimit()) { show_signal_msg(regs, error_code, address, tsk);
printk(
"%s%s[%d]: segfault at %lx ip %p sp %p error %lx",
task_pid_nr(tsk) > 1 ? KERN_INFO : KERN_EMERG,
tsk->comm, task_pid_nr(tsk), address,
(void *) regs->ip, (void *) regs->sp, error_code);
print_vma_addr(" in ", regs->ip);
printk("\n");
}
/* Kernel addresses are always protection faults: */
tsk->thread.cr2 = address; tsk->thread.cr2 = address;
/* Kernel addresses are always protection faults */
tsk->thread.error_code = error_code | (address >= TASK_SIZE); tsk->thread.error_code = error_code | (address >= TASK_SIZE);
tsk->thread.trap_no = 14; tsk->thread.trap_no = 14;
force_sig_info_fault(SIGSEGV, si_code, address, tsk); force_sig_info_fault(SIGSEGV, si_code, address, tsk);
return; return;
} }
...@@ -541,15 +613,16 @@ static void __bad_area_nosemaphore(struct pt_regs *regs, ...@@ -541,15 +613,16 @@ static void __bad_area_nosemaphore(struct pt_regs *regs,
no_context(regs, error_code, address); no_context(regs, error_code, address);
} }
static noinline void bad_area_nosemaphore(struct pt_regs *regs, static noinline void
unsigned long error_code, unsigned long address) bad_area_nosemaphore(struct pt_regs *regs, unsigned long error_code,
unsigned long address)
{ {
__bad_area_nosemaphore(regs, error_code, address, SEGV_MAPERR); __bad_area_nosemaphore(regs, error_code, address, SEGV_MAPERR);
} }
static void __bad_area(struct pt_regs *regs, static void
unsigned long error_code, unsigned long address, __bad_area(struct pt_regs *regs, unsigned long error_code,
int si_code) unsigned long address, int si_code)
{ {
struct mm_struct *mm = current->mm; struct mm_struct *mm = current->mm;
...@@ -562,67 +635,77 @@ static void __bad_area(struct pt_regs *regs, ...@@ -562,67 +635,77 @@ static void __bad_area(struct pt_regs *regs,
__bad_area_nosemaphore(regs, error_code, address, si_code); __bad_area_nosemaphore(regs, error_code, address, si_code);
} }
static noinline void bad_area(struct pt_regs *regs, static noinline void
unsigned long error_code, unsigned long address) bad_area(struct pt_regs *regs, unsigned long error_code, unsigned long address)
{ {
__bad_area(regs, error_code, address, SEGV_MAPERR); __bad_area(regs, error_code, address, SEGV_MAPERR);
} }
static noinline void bad_area_access_error(struct pt_regs *regs, static noinline void
unsigned long error_code, unsigned long address) bad_area_access_error(struct pt_regs *regs, unsigned long error_code,
unsigned long address)
{ {
__bad_area(regs, error_code, address, SEGV_ACCERR); __bad_area(regs, error_code, address, SEGV_ACCERR);
} }
/* TODO: fixup for "mm-invoke-oom-killer-from-page-fault.patch" */ /* TODO: fixup for "mm-invoke-oom-killer-from-page-fault.patch" */
static void out_of_memory(struct pt_regs *regs, static void
unsigned long error_code, unsigned long address) out_of_memory(struct pt_regs *regs, unsigned long error_code,
unsigned long address)
{ {
/* /*
* We ran out of memory, call the OOM killer, and return the userspace * We ran out of memory, call the OOM killer, and return the userspace
* (which will retry the fault, or kill us if we got oom-killed). * (which will retry the fault, or kill us if we got oom-killed):
*/ */
up_read(&current->mm->mmap_sem); up_read(&current->mm->mmap_sem);
pagefault_out_of_memory(); pagefault_out_of_memory();
} }
static void do_sigbus(struct pt_regs *regs, static void
unsigned long error_code, unsigned long address) do_sigbus(struct pt_regs *regs, unsigned long error_code, unsigned long address)
{ {
struct task_struct *tsk = current; struct task_struct *tsk = current;
struct mm_struct *mm = tsk->mm; struct mm_struct *mm = tsk->mm;
up_read(&mm->mmap_sem); up_read(&mm->mmap_sem);
/* Kernel mode? Handle exceptions or die */ /* Kernel mode? Handle exceptions or die: */
if (!(error_code & PF_USER)) if (!(error_code & PF_USER))
no_context(regs, error_code, address); no_context(regs, error_code, address);
#ifdef CONFIG_X86_32 #ifdef CONFIG_X86_32
/* User space => ok to do another page fault */ /* User space => ok to do another page fault: */
if (is_prefetch(regs, error_code, address)) if (is_prefetch(regs, error_code, address))
return; return;
#endif #endif
tsk->thread.cr2 = address; tsk->thread.cr2 = address;
tsk->thread.error_code = error_code; tsk->thread.error_code = error_code;
tsk->thread.trap_no = 14; tsk->thread.trap_no = 14;
force_sig_info_fault(SIGBUS, BUS_ADRERR, address, tsk); force_sig_info_fault(SIGBUS, BUS_ADRERR, address, tsk);
} }
static noinline void mm_fault_error(struct pt_regs *regs, static noinline void
unsigned long error_code, unsigned long address, unsigned int fault) mm_fault_error(struct pt_regs *regs, unsigned long error_code,
unsigned long address, unsigned int fault)
{ {
if (fault & VM_FAULT_OOM) if (fault & VM_FAULT_OOM) {
out_of_memory(regs, error_code, address); out_of_memory(regs, error_code, address);
else if (fault & VM_FAULT_SIGBUS) } else {
if (fault & VM_FAULT_SIGBUS)
do_sigbus(regs, error_code, address); do_sigbus(regs, error_code, address);
else else
BUG(); BUG();
}
} }
static int spurious_fault_check(unsigned long error_code, pte_t *pte) static int spurious_fault_check(unsigned long error_code, pte_t *pte)
{ {
if ((error_code & PF_WRITE) && !pte_write(*pte)) if ((error_code & PF_WRITE) && !pte_write(*pte))
return 0; return 0;
if ((error_code & PF_INSTR) && !pte_exec(*pte)) if ((error_code & PF_INSTR) && !pte_exec(*pte))
return 0; return 0;
...@@ -630,16 +713,19 @@ static int spurious_fault_check(unsigned long error_code, pte_t *pte) ...@@ -630,16 +713,19 @@ static int spurious_fault_check(unsigned long error_code, pte_t *pte)
} }
/* /*
* Handle a spurious fault caused by a stale TLB entry. This allows * Handle a spurious fault caused by a stale TLB entry.
* us to lazily refresh the TLB when increasing the permissions of a *
* kernel page (RO -> RW or NX -> X). Doing it eagerly is very * This allows us to lazily refresh the TLB when increasing the
* expensive since that implies doing a full cross-processor TLB * permissions of a kernel page (RO -> RW or NX -> X). Doing it
* flush, even if no stale TLB entries exist on other processors. * eagerly is very expensive since that implies doing a full
* cross-processor TLB flush, even if no stale TLB entries exist
* on other processors.
*
* There are no security implications to leaving a stale TLB when * There are no security implications to leaving a stale TLB when
* increasing the permissions on a page. * increasing the permissions on a page.
*/ */
static noinline int spurious_fault(unsigned long error_code, static noinline int
unsigned long address) spurious_fault(unsigned long error_code, unsigned long address)
{ {
pgd_t *pgd; pgd_t *pgd;
pud_t *pud; pud_t *pud;
...@@ -678,19 +764,22 @@ static noinline int spurious_fault(unsigned long error_code, ...@@ -678,19 +764,22 @@ static noinline int spurious_fault(unsigned long error_code,
return 0; return 0;
/* /*
* Make sure we have permissions in PMD * Make sure we have permissions in PMD.
* If not, then there's a bug in the page tables. * If not, then there's a bug in the page tables:
*/ */
ret = spurious_fault_check(error_code, (pte_t *) pmd); ret = spurious_fault_check(error_code, (pte_t *) pmd);
WARN_ONCE(!ret, "PMD has incorrect permission bits\n"); WARN_ONCE(!ret, "PMD has incorrect permission bits\n");
return ret; return ret;
} }
/* /*
* X86_32 * 32-bit:
*
* Handle a fault on the vmalloc or module mapping area * Handle a fault on the vmalloc or module mapping area
* *
* X86_64 * 64-bit:
*
* Handle a fault on the vmalloc area * Handle a fault on the vmalloc area
* *
* This assumes no large pages in there. * This assumes no large pages in there.
...@@ -702,7 +791,7 @@ static noinline int vmalloc_fault(unsigned long address) ...@@ -702,7 +791,7 @@ static noinline int vmalloc_fault(unsigned long address)
pmd_t *pmd_k; pmd_t *pmd_k;
pte_t *pte_k; pte_t *pte_k;
/* Make sure we are in vmalloc area */ /* Make sure we are in vmalloc area: */
if (!(address >= VMALLOC_START && address < VMALLOC_END)) if (!(address >= VMALLOC_START && address < VMALLOC_END))
return -1; return -1;
...@@ -717,9 +806,11 @@ static noinline int vmalloc_fault(unsigned long address) ...@@ -717,9 +806,11 @@ static noinline int vmalloc_fault(unsigned long address)
pmd_k = vmalloc_sync_one(__va(pgd_paddr), address); pmd_k = vmalloc_sync_one(__va(pgd_paddr), address);
if (!pmd_k) if (!pmd_k)
return -1; return -1;
pte_k = pte_offset_kernel(pmd_k, address); pte_k = pte_offset_kernel(pmd_k, address);
if (!pte_present(*pte_k)) if (!pte_present(*pte_k))
return -1; return -1;
return 0; return 0;
#else #else
pgd_t *pgd, *pgd_ref; pgd_t *pgd, *pgd_ref;
...@@ -727,68 +818,83 @@ static noinline int vmalloc_fault(unsigned long address) ...@@ -727,68 +818,83 @@ static noinline int vmalloc_fault(unsigned long address)
pmd_t *pmd, *pmd_ref; pmd_t *pmd, *pmd_ref;
pte_t *pte, *pte_ref; pte_t *pte, *pte_ref;
/* Make sure we are in vmalloc area */ /* Make sure we are in vmalloc area: */
if (!(address >= VMALLOC_START && address < VMALLOC_END)) if (!(address >= VMALLOC_START && address < VMALLOC_END))
return -1; return -1;
/* Copy kernel mappings over when needed. This can also /*
happen within a race in page table update. In the later * Copy kernel mappings over when needed. This can also
case just flush. */ * happen within a race in page table update. In the later
* case just flush:
*/
pgd = pgd_offset(current->active_mm, address); pgd = pgd_offset(current->active_mm, address);
pgd_ref = pgd_offset_k(address); pgd_ref = pgd_offset_k(address);
if (pgd_none(*pgd_ref)) if (pgd_none(*pgd_ref))
return -1; return -1;
if (pgd_none(*pgd)) if (pgd_none(*pgd))
set_pgd(pgd, *pgd_ref); set_pgd(pgd, *pgd_ref);
else else
BUG_ON(pgd_page_vaddr(*pgd) != pgd_page_vaddr(*pgd_ref)); BUG_ON(pgd_page_vaddr(*pgd) != pgd_page_vaddr(*pgd_ref));
/* Below here mismatches are bugs because these lower tables /*
are shared */ * Below here mismatches are bugs because these lower tables
* are shared:
*/
pud = pud_offset(pgd, address); pud = pud_offset(pgd, address);
pud_ref = pud_offset(pgd_ref, address); pud_ref = pud_offset(pgd_ref, address);
if (pud_none(*pud_ref)) if (pud_none(*pud_ref))
return -1; return -1;
if (pud_none(*pud) || pud_page_vaddr(*pud) != pud_page_vaddr(*pud_ref)) if (pud_none(*pud) || pud_page_vaddr(*pud) != pud_page_vaddr(*pud_ref))
BUG(); BUG();
pmd = pmd_offset(pud, address); pmd = pmd_offset(pud, address);
pmd_ref = pmd_offset(pud_ref, address); pmd_ref = pmd_offset(pud_ref, address);
if (pmd_none(*pmd_ref)) if (pmd_none(*pmd_ref))
return -1; return -1;
if (pmd_none(*pmd) || pmd_page(*pmd) != pmd_page(*pmd_ref)) if (pmd_none(*pmd) || pmd_page(*pmd) != pmd_page(*pmd_ref))
BUG(); BUG();
pte_ref = pte_offset_kernel(pmd_ref, address); pte_ref = pte_offset_kernel(pmd_ref, address);
if (!pte_present(*pte_ref)) if (!pte_present(*pte_ref))
return -1; return -1;
pte = pte_offset_kernel(pmd, address); pte = pte_offset_kernel(pmd, address);
/* Don't use pte_page here, because the mappings can point
outside mem_map, and the NUMA hash lookup cannot handle /*
that. */ * Don't use pte_page here, because the mappings can point
* outside mem_map, and the NUMA hash lookup cannot handle
* that:
*/
if (!pte_present(*pte) || pte_pfn(*pte) != pte_pfn(*pte_ref)) if (!pte_present(*pte) || pte_pfn(*pte) != pte_pfn(*pte_ref))
BUG(); BUG();
return 0; return 0;
#endif #endif
} }
int show_unhandled_signals = 1; int show_unhandled_signals = 1;
static inline int access_error(unsigned long error_code, int write, static inline int
struct vm_area_struct *vma) access_error(unsigned long error_code, int write, struct vm_area_struct *vma)
{ {
if (write) { if (write) {
/* write, present and write, not present */ /* write, present and write, not present: */
if (unlikely(!(vma->vm_flags & VM_WRITE))) if (unlikely(!(vma->vm_flags & VM_WRITE)))
return 1; return 1;
} else if (unlikely(error_code & PF_PROT)) { return 0;
/* read, present */ }
/* read, present: */
if (unlikely(error_code & PF_PROT))
return 1; return 1;
} else {
/* read, not present */ /* read, not present: */
if (unlikely(!(vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE)))) if (unlikely(!(vma->vm_flags & (VM_READ | VM_EXEC | VM_WRITE))))
return 1; return 1;
}
return 0; return 0;
} }
...@@ -797,9 +903,9 @@ static int fault_in_kernel_space(unsigned long address) ...@@ -797,9 +903,9 @@ static int fault_in_kernel_space(unsigned long address)
{ {
#ifdef CONFIG_X86_32 #ifdef CONFIG_X86_32
return address >= TASK_SIZE; return address >= TASK_SIZE;
#else /* !CONFIG_X86_32 */ #else
return address >= TASK_SIZE64; return address >= TASK_SIZE64;
#endif /* CONFIG_X86_32 */ #endif
} }
/* /*
...@@ -812,18 +918,19 @@ asmlinkage ...@@ -812,18 +918,19 @@ asmlinkage
#endif #endif
void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
{ {
unsigned long address; struct vm_area_struct *vma;
struct task_struct *tsk; struct task_struct *tsk;
unsigned long address;
struct mm_struct *mm; struct mm_struct *mm;
struct vm_area_struct *vma;
int write; int write;
int fault; int fault;
tsk = current; tsk = current;
mm = tsk->mm; mm = tsk->mm;
prefetchw(&mm->mmap_sem); prefetchw(&mm->mmap_sem);
/* get the address */ /* Get the faulting address: */
address = read_cr2(); address = read_cr2();
if (unlikely(kmmio_fault(regs, address))) if (unlikely(kmmio_fault(regs, address)))
...@@ -847,22 +954,23 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) ...@@ -847,22 +954,23 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
vmalloc_fault(address) >= 0) vmalloc_fault(address) >= 0)
return; return;
/* Can handle a stale RO->RW TLB */ /* Can handle a stale RO->RW TLB: */
if (spurious_fault(error_code, address)) if (spurious_fault(error_code, address))
return; return;
/* kprobes don't want to hook the spurious faults. */ /* kprobes don't want to hook the spurious faults: */
if (notify_page_fault(regs)) if (notify_page_fault(regs))
return; return;
/* /*
* Don't take the mm semaphore here. If we fixup a prefetch * Don't take the mm semaphore here. If we fixup a prefetch
* fault we could otherwise deadlock. * fault we could otherwise deadlock:
*/ */
bad_area_nosemaphore(regs, error_code, address); bad_area_nosemaphore(regs, error_code, address);
return; return;
} }
/* kprobes don't want to hook the spurious faults. */ /* kprobes don't want to hook the spurious faults: */
if (unlikely(notify_page_fault(regs))) if (unlikely(notify_page_fault(regs)))
return; return;
/* /*
...@@ -870,13 +978,15 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) ...@@ -870,13 +978,15 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
* vmalloc fault has been handled. * vmalloc fault has been handled.
* *
* User-mode registers count as a user access even for any * User-mode registers count as a user access even for any
* potential system fault or CPU buglet. * potential system fault or CPU buglet:
*/ */
if (user_mode_vm(regs)) { if (user_mode_vm(regs)) {
local_irq_enable(); local_irq_enable();
error_code |= PF_USER; error_code |= PF_USER;
} else if (regs->flags & X86_EFLAGS_IF) } else {
if (regs->flags & X86_EFLAGS_IF)
local_irq_enable(); local_irq_enable();
}
#ifdef CONFIG_X86_64 #ifdef CONFIG_X86_64
if (unlikely(error_code & PF_RSVD)) if (unlikely(error_code & PF_RSVD))
...@@ -884,8 +994,8 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) ...@@ -884,8 +994,8 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
#endif #endif
/* /*
* If we're in an interrupt, have no user context or are running in an * If we're in an interrupt, have no user context or are running
* atomic region then we must not take the fault. * in an atomic region then we must not take the fault:
*/ */
if (unlikely(in_atomic() || !mm)) { if (unlikely(in_atomic() || !mm)) {
bad_area_nosemaphore(regs, error_code, address); bad_area_nosemaphore(regs, error_code, address);
...@@ -894,19 +1004,19 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) ...@@ -894,19 +1004,19 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
/* /*
* When running in the kernel we expect faults to occur only to * When running in the kernel we expect faults to occur only to
* addresses in user space. All other faults represent errors in the * addresses in user space. All other faults represent errors in
* kernel and should generate an OOPS. Unfortunately, in the case of an * the kernel and should generate an OOPS. Unfortunately, in the
* erroneous fault occurring in a code path which already holds mmap_sem * case of an erroneous fault occurring in a code path which already
* we will deadlock attempting to validate the fault against the * holds mmap_sem we will deadlock attempting to validate the fault
* address space. Luckily the kernel only validly references user * against the address space. Luckily the kernel only validly
* space from well defined areas of code, which are listed in the * references user space from well defined areas of code, which are
* exceptions table. * listed in the exceptions table.
* *
* As the vast majority of faults will be valid we will only perform * As the vast majority of faults will be valid we will only perform
* the source reference check when there is a possibility of a deadlock. * the source reference check when there is a possibility of a
* Attempt to lock the address space, if we cannot we then validate the * deadlock. Attempt to lock the address space, if we cannot we then
* source. If this is invalid we can skip the address space check, * validate the source. If this is invalid we can skip the address
* thus avoiding the deadlock. * space check, thus avoiding the deadlock:
*/ */
if (unlikely(!down_read_trylock(&mm->mmap_sem))) { if (unlikely(!down_read_trylock(&mm->mmap_sem))) {
if ((error_code & PF_USER) == 0 && if ((error_code & PF_USER) == 0 &&
...@@ -917,8 +1027,9 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) ...@@ -917,8 +1027,9 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
down_read(&mm->mmap_sem); down_read(&mm->mmap_sem);
} else { } else {
/* /*
* The above down_read_trylock() might have succeeded in which * The above down_read_trylock() might have succeeded in
* case we'll have missed the might_sleep() from down_read(). * which case we'll have missed the might_sleep() from
* down_read():
*/ */
might_sleep(); might_sleep();
} }
...@@ -938,7 +1049,7 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) ...@@ -938,7 +1049,7 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
/* /*
* Accessing the stack below %sp is always a bug. * Accessing the stack below %sp is always a bug.
* The large cushion allows instructions like enter * The large cushion allows instructions like enter
* and pusha to work. ("enter $65535,$31" pushes * and pusha to work. ("enter $65535, $31" pushes
* 32 pointers and then decrements %sp by 65535.) * 32 pointers and then decrements %sp by 65535.)
*/ */
if (unlikely(address + 65536 + 32 * sizeof(unsigned long) < regs->sp)) { if (unlikely(address + 65536 + 32 * sizeof(unsigned long) < regs->sp)) {
...@@ -957,6 +1068,7 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) ...@@ -957,6 +1068,7 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
*/ */
good_area: good_area:
write = error_code & PF_WRITE; write = error_code & PF_WRITE;
if (unlikely(access_error(error_code, write, vma))) { if (unlikely(access_error(error_code, write, vma))) {
bad_area_access_error(regs, error_code, address); bad_area_access_error(regs, error_code, address);
return; return;
...@@ -965,13 +1077,15 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code) ...@@ -965,13 +1077,15 @@ void __kprobes do_page_fault(struct pt_regs *regs, unsigned long error_code)
/* /*
* If for any reason at all we couldn't handle the fault, * If for any reason at all we couldn't handle the fault,
* make sure we exit gracefully rather than endlessly redo * make sure we exit gracefully rather than endlessly redo
* the fault. * the fault:
*/ */
fault = handle_mm_fault(mm, vma, address, write); fault = handle_mm_fault(mm, vma, address, write);
if (unlikely(fault & VM_FAULT_ERROR)) { if (unlikely(fault & VM_FAULT_ERROR)) {
mm_fault_error(regs, error_code, address, fault); mm_fault_error(regs, error_code, address, fault);
return; return;
} }
if (fault & VM_FAULT_MAJOR) if (fault & VM_FAULT_MAJOR)
tsk->maj_flt++; tsk->maj_flt++;
else else
...@@ -1004,13 +1118,13 @@ void vmalloc_sync_all(void) ...@@ -1004,13 +1118,13 @@ void vmalloc_sync_all(void)
for (address = VMALLOC_START & PMD_MASK; for (address = VMALLOC_START & PMD_MASK;
address >= TASK_SIZE && address < FIXADDR_TOP; address >= TASK_SIZE && address < FIXADDR_TOP;
address += PMD_SIZE) { address += PMD_SIZE) {
unsigned long flags; unsigned long flags;
struct page *page; struct page *page;
spin_lock_irqsave(&pgd_lock, flags); spin_lock_irqsave(&pgd_lock, flags);
list_for_each_entry(page, &pgd_list, lru) { list_for_each_entry(page, &pgd_list, lru) {
if (!vmalloc_sync_one(page_address(page), if (!vmalloc_sync_one(page_address(page), address))
address))
break; break;
} }
spin_unlock_irqrestore(&pgd_lock, flags); spin_unlock_irqrestore(&pgd_lock, flags);
...@@ -1018,12 +1132,14 @@ void vmalloc_sync_all(void) ...@@ -1018,12 +1132,14 @@ void vmalloc_sync_all(void)
#else /* CONFIG_X86_64 */ #else /* CONFIG_X86_64 */
for (address = VMALLOC_START & PGDIR_MASK; address <= VMALLOC_END; for (address = VMALLOC_START & PGDIR_MASK; address <= VMALLOC_END;
address += PGDIR_SIZE) { address += PGDIR_SIZE) {
const pgd_t *pgd_ref = pgd_offset_k(address); const pgd_t *pgd_ref = pgd_offset_k(address);
unsigned long flags; unsigned long flags;
struct page *page; struct page *page;
if (pgd_none(*pgd_ref)) if (pgd_none(*pgd_ref))
continue; continue;
spin_lock_irqsave(&pgd_lock, flags); spin_lock_irqsave(&pgd_lock, flags);
list_for_each_entry(page, &pgd_list, lru) { list_for_each_entry(page, &pgd_list, lru) {
pgd_t *pgd; pgd_t *pgd;
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7