* Re: [patch 16/16] Add hardware breakpoint support for i386
@ 2005-08-30 8:26 Prasanna S Panchamukhi
2005-08-30 8:28 ` Prasanna S Panchamukhi
0 siblings, 1 reply; 6+ messages in thread
From: Prasanna S Panchamukhi @ 2005-08-30 8:26 UTC (permalink / raw)
To: Tom Rini, Andi Kleen, kaos; +Cc: linux-kernel
Hi,
> > This adds hardware breakpoint support for i386. This is not as well tested
> as
> > software breakpoints, but in some minimal testing appears to be
> functional.
>
> This really would need so coordination with user space using
> them. Otherwise it'll be quite unreliable because any user program
> can break it.
>
> Long ago (in 2.4 time frame) there used to be a IBM patch floating
> around to reserve them globally and user space to use specific ones. I
> guess something like that would be needed again.
>
Yes, to add hardware breakpoint support for i386, there are two patches.
1. Provides hardware debug register allocation mechanism.
2. Provides light weight interface for kernel-space watchpoint probes
These patches have been posted & reviewd on lkml and systemtap mailing lists.
Your comments are welcome.
Thanks
Prasanna
This patch provides debug register allocation mechanism.
Useful for debuggers like IOW, kgdb, kdb, kernel watchpoint.
---
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
---
---
linux-2.6.13-prasanna/arch/i386/Kconfig.debug | 8
linux-2.6.13-prasanna/arch/i386/kernel/Makefile | 1
linux-2.6.13-prasanna/arch/i386/kernel/debugreg.c | 281 ++++++++++++++++++++++
linux-2.6.13-prasanna/arch/i386/kernel/process.c | 31 +-
linux-2.6.13-prasanna/arch/i386/kernel/ptrace.c | 5
linux-2.6.13-prasanna/arch/i386/kernel/signal.c | 3
linux-2.6.13-prasanna/arch/i386/kernel/traps.c | 2
linux-2.6.13-prasanna/include/asm-i386/debugreg.h | 189 ++++++++++++++
8 files changed, 511 insertions(+), 9 deletions(-)
diff -puN arch/i386/Kconfig.debug~kprobes-debug-regs arch/i386/Kconfig.debug
--- linux-2.6.13/arch/i386/Kconfig.debug~kprobes-debug-regs 2005-08-30 11:43:49.369626152 +0530
+++ linux-2.6.13-prasanna/arch/i386/Kconfig.debug 2005-08-30 11:43:49.442615056 +0530
@@ -32,6 +32,14 @@ config KPROBES
for kernel debugging, non-intrusive instrumentation and testing.
If in doubt, say "N".
+config DEBUGREG
+ bool "Global Debug Registers"
+ depends on DEBUG_KERNEL
+ default off
+ help
+ Global debug register allocation mechanism is useful for debuggers
+ IOW, Kgdb, Kdb, Kernel Watchpoint probes. If in doubt say "N"
+
config DEBUG_STACK_USAGE
bool "Stack utilization instrumentation"
depends on DEBUG_KERNEL
diff -puN /dev/null arch/i386/kernel/debugreg.c
--- /dev/null 2005-08-30 16:04:24.253093808 +0530
+++ linux-2.6.13-prasanna/arch/i386/kernel/debugreg.c 2005-08-30 11:43:49.444614752 +0530
@@ -0,0 +1,281 @@
+/*
+ * Debug register
+ * arch/i386/kernel/debugreg.c
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright (C) IBM Corporation, 2002, 2004
+ *
+ * 2002-Oct Created by Vamsi Krishna S <vamsi_krishna@in.ibm.com> and
+ * Bharata Rao <bharata@in.ibm.com> to provide debug register
+ * allocation mechanism.
+ * 2004-Oct Updated by Prasanna S Panchamukhi <prasanna@in.ibm.com> with
+ * idr_allocations mechanism as suggested by Andi Kleen.
+ */
+/*
+ * This provides a debug register allocation mechanism, to be
+ * used by all debuggers, which need debug registers.
+ *
+ */
+#include <linux/kernel.h>
+#include <linux/spinlock.h>
+#include <linux/module.h>
+#include <linux/idr.h>
+#include <asm/system.h>
+#include <asm/debugreg.h>
+
+struct debugreg dr_list[DR_MAX];
+unsigned long dr7_global_mask = 0;
+static spinlock_t dr_lock = SPIN_LOCK_UNLOCKED;
+static DEFINE_IDR(debugreg_idr);
+static DECLARE_MUTEX(debugreg_idr_mutex);
+static spinlock_t debugreg_idr_lock = SPIN_LOCK_UNLOCKED;
+
+static unsigned long dr7_global_bits[] = {
+ DR7_DR0_BITS, DR7_DR1_BITS, DR7_DR2_BITS, DR7_DR3_BITS
+};
+
+static inline void set_dr7_global_mask(int regnum)
+{
+ if (DR_IS_ADDR(regnum))
+ dr7_global_mask |= dr7_global_bits[regnum];
+}
+
+static inline void clear_dr7_global_mask(int regnum)
+{
+ if (DR_IS_ADDR(regnum))
+ dr7_global_mask |= ~dr7_global_bits[regnum];
+}
+
+/*
+ * See if specific debug register is free.
+ */
+static int specific_debugreg(unsigned int regnum)
+{
+ int r, n;
+
+ if (regnum >= DR_MAX)
+ return -EINVAL;
+
+ down(&debugreg_idr_mutex);
+ r = idr_pre_get(&debugreg_idr, GFP_KERNEL);
+ up(&debugreg_idr_mutex);
+ if (!r)
+ return -ENOMEM;
+
+ spin_lock(&debugreg_idr_lock);
+
+ if (idr_find(&debugreg_idr, regnum)) {
+ r = -EBUSY;
+ goto out;
+ }
+
+ r = idr_get_new_above(&debugreg_idr, specific_debugreg, regnum, &n);
+ if (r)
+ goto out;
+
+ if (n != regnum) {
+ idr_remove(&debugreg_idr, n);
+ r = -EBUSY;
+ goto out;
+ }
+
+ out:
+ spin_unlock(&debugreg_idr_lock);
+ return r;
+}
+
+static int next_free_debugreg(unsigned int *regnum)
+{
+ int r;
+ unsigned int n;
+
+ down(&debugreg_idr_mutex);
+ r = idr_pre_get(&debugreg_idr, GFP_KERNEL);
+ up(&debugreg_idr_mutex);
+ if (!r)
+ return -ENOMEM;
+
+ spin_lock(&debugreg_idr_lock);
+
+ r = idr_get_new(&debugreg_idr, next_free_debugreg, &n);
+ if (r)
+ goto out;
+
+ if (n >= DR_MAX) {
+ idr_remove(&debugreg_idr, n);
+ r = -ENOSPC;
+ goto out;
+ }
+
+ *regnum = n;
+ out:
+ spin_unlock(&debugreg_idr_lock);
+ return r;
+}
+
+static void free_debugreg(int regnum)
+{
+ spin_lock(&debugreg_idr_lock);
+ idr_remove(&debugreg_idr, regnum);
+ spin_unlock(&debugreg_idr_lock);
+}
+
+static int get_dr(int regnum, int flag)
+{
+ int ret;
+
+ ret = specific_debugreg(regnum);
+
+ if ((flag == DR_ALLOC_GLOBAL) && (ret >= 0)) {
+ dr_list[regnum].flag = DR_GLOBAL;
+ set_dr7_global_mask(regnum);
+ return regnum;
+ } else if (dr_list[regnum].flag != DR_GLOBAL) {
+ dr_list[regnum].use_count++;
+ dr_list[regnum].flag = DR_LOCAL;
+ return regnum;
+ }
+ return ret;
+}
+
+static int get_any_dr(int flag)
+{
+ int i, ret = 0;
+
+ if (flag == DR_ALLOC_LOCAL) {
+ for (i = 0; i < DR_MAX; i++) {
+ if ((idr_find(&debugreg_idr, i)) && (dr_list[i].flag == DR_LOCAL)) {
+ dr_list[i].use_count++;
+ return i;
+ }
+ }
+ if ((ret = next_free_debugreg(&i)) >= 0) {
+ dr_list[i].flag = DR_LOCAL;
+ dr_list[i].use_count = 1;
+ return i;
+ }
+ } else {
+ if ((ret = next_free_debugreg(&i)) >= 0) {
+ dr_list[i].flag = DR_GLOBAL;
+ set_dr7_global_mask(i);
+ return i;
+ }
+ }
+ return ret;
+}
+
+static inline void dr_free_local(int regnum)
+{
+ if (!(--dr_list[regnum].use_count)) {
+ free_debugreg(regnum);
+ dr_list[regnum].flag = DR_UNUSED;
+ }
+}
+
+static inline void dr_free_global(int regnum)
+{
+ free_debugreg(regnum);
+ dr_list[regnum].flag = DR_UNUSED;
+ dr_list[regnum].use_count = 0;
+ clear_dr7_global_mask(regnum);
+}
+
+int dr_alloc(int regnum, int flag)
+{
+ int ret = 0;
+
+ spin_lock(&dr_lock);
+ if (regnum == DR_ANY)
+ ret = get_any_dr(flag);
+ else if (regnum >= DR_MAX)
+ ret = -1;
+ else
+ ret = get_dr(regnum, flag);
+ spin_unlock(&dr_lock);
+ if (ret < 0)
+ printk("dr_alloc:Cannot allocate debug register %d\n", regnum);
+ return ret;
+}
+
+int dr_free(int regnum)
+{
+ spin_lock(&dr_lock);
+ if (regnum >= DR_MAX || (!idr_find(&debugreg_idr, regnum))) {
+ spin_unlock(&dr_lock);
+ printk("dr_free:Cannot free debug register %d\n", regnum);
+ return -1;
+ }
+ if (dr_list[regnum].flag == DR_LOCAL)
+ dr_free_local(regnum);
+ else
+ dr_free_global(regnum);
+ spin_unlock(&dr_lock);
+ return 0;
+}
+
+void dr_inc_use_count(unsigned long debugreg)
+{
+ int i;
+
+ spin_lock(&dr_lock);
+ for (i = 0; i < DR_MAX; i++) {
+ if ((idr_find(&debugreg_idr, i)) && (DR_IS_LOCAL(debugreg, i)))
+ dr_list[i].use_count++;
+ }
+ spin_unlock(&dr_lock);
+}
+
+void dr_dec_use_count(unsigned long debugreg)
+{
+ int i;
+
+ spin_lock(&dr_lock);
+ for (i = 0; i < DR_MAX; i++) {
+ if ((idr_find(&debugreg_idr, i)) && (DR_IS_LOCAL(debugreg, i)))
+ dr_free_local(i);
+ }
+ spin_unlock(&dr_lock);
+}
+
+/*
+ * This routine decides if the ptrace request is for enabling or disabling
+ * a debug reg, and accordingly calls dr_alloc() or dr_free().
+ *
+ * gdb uses ptrace to write to debug registers. It assumes that writing to
+ * debug register always succeds and it doesn't check the return value of
+ * ptrace. Now with this new global debug register allocation/freeing,
+ * ptrace request for a local debug register can fail, if the required debug
+ * register is already globally allocated. Since gdb fails to notice this
+ * failure, it sometimes tries to free a debug register, which is not
+ * allocated for it.
+ */
+int enable_debugreg(unsigned long old_dr7, unsigned long new_dr7)
+{
+ int i, dr_shift = 1UL;
+ for (i = 0; i < DR_MAX; i++, dr_shift <<= 2) {
+ if ((old_dr7 ^ new_dr7) & dr_shift) {
+ if (new_dr7 & dr_shift)
+ dr_alloc(i, DR_ALLOC_LOCAL);
+ else
+ dr_free(i);
+ return 0;
+ }
+ }
+ return -1;
+}
+
+EXPORT_SYMBOL(dr_alloc);
+EXPORT_SYMBOL(dr_free);
diff -puN arch/i386/kernel/Makefile~kprobes-debug-regs arch/i386/kernel/Makefile
--- linux-2.6.13/arch/i386/kernel/Makefile~kprobes-debug-regs 2005-08-30 11:43:49.374625392 +0530
+++ linux-2.6.13-prasanna/arch/i386/kernel/Makefile 2005-08-30 11:43:49.444614752 +0530
@@ -34,6 +34,7 @@ obj-$(CONFIG_ACPI_SRAT) += srat.o
obj-$(CONFIG_HPET_TIMER) += time_hpet.o
obj-$(CONFIG_EFI) += efi.o efi_stub.o
obj-$(CONFIG_EARLY_PRINTK) += early_printk.o
+obj-$(CONFIG_DEBUGREG) += debugreg.o
EXTRA_AFLAGS := -traditional
diff -puN arch/i386/kernel/process.c~kprobes-debug-regs arch/i386/kernel/process.c
--- linux-2.6.13/arch/i386/kernel/process.c~kprobes-debug-regs 2005-08-30 11:43:49.377624936 +0530
+++ linux-2.6.13-prasanna/arch/i386/kernel/process.c 2005-08-30 11:43:49.447614296 +0530
@@ -52,6 +52,7 @@
#ifdef CONFIG_MATH_EMULATION
#include <asm/math_emu.h>
#endif
+#include <asm/debugreg.h>
#include <linux/irq.h>
#include <linux/err.h>
@@ -399,6 +400,8 @@ void exit_thread(void)
tss->io_bitmap_base = INVALID_IO_BITMAP_OFFSET;
put_cpu();
}
+ if (tsk->thread.debugreg[7])
+ dr_dec_use_count(tsk->thread.debugreg[7]);
}
void flush_thread(void)
@@ -412,6 +415,8 @@ void flush_thread(void)
*/
kprobe_flush_task(tsk);
+ if (tsk->thread.debugreg[7])
+ dr_dec_use_count(tsk->thread.debugreg[7]);
memset(tsk->thread.debugreg, 0, sizeof(unsigned long)*8);
memset(tsk->thread.tls_array, 0, sizeof(tsk->thread.tls_array));
/*
@@ -513,6 +518,9 @@ int copy_thread(int nr, unsigned long cl
desc->b = LDT_entry_b(&info);
}
+ if (tsk->thread.debugreg[7])
+ dr_inc_use_count(tsk->thread.debugreg[7]);
+
err = 0;
out:
if (err && p->thread.io_bitmap_ptr) {
@@ -676,6 +684,7 @@ struct task_struct fastcall * __switch_t
*next = &next_p->thread;
int cpu = smp_processor_id();
struct tss_struct *tss = &per_cpu(init_tss, cpu);
+ unsigned long next_dr7 = next->debugreg[7];
/* never put a printk in __switch_to... printk() calls wake_up*() indirectly */
@@ -713,14 +722,22 @@ struct task_struct fastcall * __switch_t
/*
* Now maybe reload the debug registers
*/
- if (unlikely(next->debugreg[7])) {
- set_debugreg(next->debugreg[0], 0);
- set_debugreg(next->debugreg[1], 1);
- set_debugreg(next->debugreg[2], 2);
- set_debugreg(next->debugreg[3], 3);
+ /*
+ * Don't reload global debug registers. Don't touch the global debug
+ * register settings in dr7.
+ */
+ if (unlikely(next_dr7)) {
+ if (prev->debugreg[0] != next->debugreg[0])
+ set_debugreg(current->thread.debugreg[0], 0);
+ if (prev->debugreg[1] != next->debugreg[1])
+ set_debugreg(current->thread.debugreg[1], 1);
+ if (prev->debugreg[2] != next->debugreg[2])
+ set_debugreg(current->thread.debugreg[2], 2);
+ if (prev->debugreg[3] != next->debugreg[3])
+ set_debugreg(current->thread.debugreg[3], 3);
/* no 4 and 5 */
- set_debugreg(next->debugreg[6], 6);
- set_debugreg(next->debugreg[7], 7);
+ set_debugreg(current->thread.debugreg[6], 6);
+ load_process_dr7(next_dr7);
}
if (unlikely(prev->io_bitmap_ptr || next->io_bitmap_ptr))
diff -puN arch/i386/kernel/ptrace.c~kprobes-debug-regs arch/i386/kernel/ptrace.c
--- linux-2.6.13/arch/i386/kernel/ptrace.c~kprobes-debug-regs 2005-08-30 11:43:49.380624480 +0530
+++ linux-2.6.13-prasanna/arch/i386/kernel/ptrace.c 2005-08-30 11:43:49.449613992 +0530
@@ -504,6 +504,11 @@ asmlinkage int sys_ptrace(long request,
addr -= (long) &dummy->u_debugreg;
addr = addr >> 2;
+
+ if (addr == 7 && (enable_debugreg(child->thread.debugreg[addr], data)) < 0) {
+ ret = -EBUSY;
+ break;
+ }
child->thread.debugreg[addr] = data;
ret = 0;
}
diff -puN arch/i386/kernel/signal.c~kprobes-debug-regs arch/i386/kernel/signal.c
--- linux-2.6.13/arch/i386/kernel/signal.c~kprobes-debug-regs 2005-08-30 11:43:49.430616880 +0530
+++ linux-2.6.13-prasanna/arch/i386/kernel/signal.c 2005-08-30 11:43:49.450613840 +0530
@@ -25,6 +25,7 @@
#include <asm/ucontext.h>
#include <asm/uaccess.h>
#include <asm/i387.h>
+#include <asm/debugreg.h>
#include "sigframe.h"
#define DEBUG_SIG 0
@@ -622,7 +623,7 @@ int fastcall do_signal(struct pt_regs *r
* inside the kernel.
*/
if (unlikely(current->thread.debugreg[7])) {
- set_debugreg(current->thread.debugreg[7], 7);
+ load_process_dr7(current->thread.debugreg[7]);
}
/* Whee! Actually deliver the signal. */
diff -puN arch/i386/kernel/traps.c~kprobes-debug-regs arch/i386/kernel/traps.c
--- linux-2.6.13/arch/i386/kernel/traps.c~kprobes-debug-regs 2005-08-30 11:43:49.434616272 +0530
+++ linux-2.6.13-prasanna/arch/i386/kernel/traps.c 2005-08-30 11:43:49.452613536 +0530
@@ -756,7 +756,7 @@ fastcall void do_debug(struct pt_regs *
* the signal is delivered.
*/
clear_dr7:
- set_debugreg(0, 7);
+ load_process_dr7(0);
return;
debug_vm86:
diff -puN include/asm-i386/debugreg.h~kprobes-debug-regs include/asm-i386/debugreg.h
--- linux-2.6.13/include/asm-i386/debugreg.h~kprobes-debug-regs 2005-08-30 11:43:49.437615816 +0530
+++ linux-2.6.13-prasanna/include/asm-i386/debugreg.h 2005-08-30 11:43:49.453613384 +0530
@@ -61,4 +61,193 @@
#define DR_LOCAL_SLOWDOWN (0x100) /* Local slow the pipeline */
#define DR_GLOBAL_SLOWDOWN (0x200) /* Global slow the pipeline */
+struct debugreg {
+ unsigned long flag;
+ unsigned long use_count;
+};
+
+/* debugreg flags */
+#define DR_UNUSED 0
+#define DR_LOCAL 1
+#define DR_GLOBAL 2
+
+#define DR_MAX 4
+#define DR_ANY DR_MAX + 1
+
+/* global or local allocation requests */
+#define DR_ALLOC_GLOBAL 0
+#define DR_ALLOC_LOCAL 1
+
+#define DR7_RW_SET(dr7, regnum, rw) do { \
+ (dr7) &= ~(0x3 << (16 + (4 * (regnum)))); \
+ (dr7) |= (((rw) & 0x3) << (16 + (4 * (regnum)))); \
+ } while (0)
+
+#define DR7_RW_VAL(dr7, regnum) \
+ (((dr7) >> (16 + (4 * (regnum)))) & 0x3)
+
+#define DR7_LEN_SET(dr7, regnum, len) do { \
+ (dr7) &= ~(0x3 << (18 + (4 * (regnum)))); \
+ (dr7) |= (((len-1) & 0x3) << (18 + (4 * (regnum)))); \
+ } while (0)
+
+#define DR7_LEN_VAL(dr7, regnum) \
+ (((dr7) >> (18 + (4 * (regnum)))) & 0x3)
+
+#define DR7_L0(dr0) (((dr0))&0x1)
+#define DR7_L1(dr1) (((dr1)>>2)&0x1)
+#define DR7_L2(dr2) (((dr2)>>4)&0x1)
+#define DR7_L3(dr3) (((dr3)>>6)&0x1)
+
+/* Check if local breakpoint is enabled */
+#define DR_IS_LOCAL(dr7, regnum) ((dr7) & (1UL << (regnum <<1)))
+
+/* Set the rw, len and global flag in dr7 for a debug register */
+#define SET_DR7(dr7, regnum, access, len) do { \
+ DR7_RW_SET(dr7, regnum, access); \
+ DR7_LEN_SET(dr7, regnum, len); \
+ dr7 |= (2UL << regnum*2); \
+ } while (0)
+
+/* Disable a debug register by clearing the global/local flag in dr7 */
+#define RESET_DR7(dr7, regnum) dr7 &= ~(3UL << regnum*2)
+
+#define DR_IS_ADDR(regnum) (0xf & (1 << (regnum)))
+
+#define DR7_DR0_BITS 0x000F0003
+#define DR7_DR1_BITS 0x00F0000C
+#define DR7_DR2_BITS 0x0F000030
+#define DR7_DR3_BITS 0xF00000C0
+
+#define DR_TRAP_MASK 0xF
+
+#define DR_TYPE_EXECUTE 0x0
+#define DR_TYPE_WRITE 0x1
+#define DR_TYPE_IO 0x2
+#define DR_TYPE_RW 0x3
+
+#define get_dr(regnum, val) get_debugreg(val, regnum)
+
+static inline unsigned long read_dr(int regnum)
+{
+ unsigned long val = 0;
+ switch (regnum) {
+ case 0:
+ get_dr(0, val);
+ break;
+ case 1:
+ get_dr(1, val);
+ break;
+ case 2:
+ get_dr(2, val);
+ break;
+ case 3:
+ get_dr(3, val);
+ break;
+ case 6:
+ get_dr(6, val);
+ break;
+ case 7:
+ get_dr(7, val);
+ break;
+ }
+ return val;
+}
+
+#undef get_dr
+
+static inline void write_dr(int regnum, unsigned long val)
+{
+ switch (regnum) {
+ case 0:
+ set_debugreg(val, 0);
+ break;
+ case 1:
+ set_debugreg(val, 1);
+ break;
+ case 2:
+ set_debugreg(val, 2);
+ break;
+ case 3:
+ set_debugreg(val, 3);
+ break;
+ case 7:
+ set_debugreg(val, 7);
+ break;
+ }
+}
+
+#ifdef CONFIG_DEBUGREG
+/*
+ * Given the debug status register, returns the debug register number
+ * which caused the debug trap.
+ */
+static inline int dr_trap(unsigned int condition)
+{
+ int i, reg_shift = 1UL;
+ for (i = 0; i < DR_MAX; i++, reg_shift <<= 1)
+ if ((condition & reg_shift))
+ return i;
+ return -1;
+}
+
+/*
+ * Given the debug status register, returns the address due to which
+ * the debug trap occured.
+ */
+static inline unsigned long dr_trap_addr(unsigned int condition)
+{
+ int regnum = dr_trap(condition);
+
+ if (regnum == -1)
+ return -1;
+ return read_dr(regnum);
+}
+
+/*
+ * Given the debug status register, returns the type of debug trap:
+ * execute, read/write, write or io.
+ */
+static inline int dr_trap_type(unsigned int condition)
+{
+ int regnum = dr_trap(condition);
+
+ if (regnum == -1)
+ return -1;
+ return DR7_RW_VAL(read_dr(7), regnum);
+}
+
+/* Function declarations */
+
+extern int dr_alloc(int regnum, int flag);
+extern int dr_free(int regnum);
+extern void dr_inc_use_count(unsigned long mask);
+extern void dr_dec_use_count(unsigned long mask);
+extern struct debugreg dr_list[DR_MAX];
+extern unsigned long dr7_global_mask;
+extern int enable_debugreg(unsigned long old_dr7, unsigned long new_dr7);
+
+static inline void load_process_dr7(unsigned long curr_dr7)
+{
+ write_dr(7, (read_dr(7) & dr7_global_mask) | curr_dr7);
+}
+#else
+static inline int enable_debugreg(unsigned long old_dr7, unsigned long new_dr7)
+{
+ return 0;
+}
+static inline void load_process_dr7(unsigned long curr_dr7)
+{
+ write_dr(7, curr_dr7);
+}
+
+static void dr_inc_use_count(unsigned long mask)
+{
+}
+
+static void dr_dec_use_count(unsigned long mask)
+{
+}
+
+#endif /* CONFIG_DEBUGREG */
#endif
_
--
Prasanna S Panchamukhi
Linux Technology Center
India Software Labs, IBM Bangalore
Ph: 91-80-25044636
<prasanna@in.ibm.com>
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [patch 16/16] Add hardware breakpoint support for i386
2005-08-30 8:26 [patch 16/16] Add hardware breakpoint support for i386 Prasanna S Panchamukhi
@ 2005-08-30 8:28 ` Prasanna S Panchamukhi
0 siblings, 0 replies; 6+ messages in thread
From: Prasanna S Panchamukhi @ 2005-08-30 8:28 UTC (permalink / raw)
To: Tom Rini, Andi Kleen, kaos; +Cc: linux-kernel
This patch provides a simple interface for kernel-space watchpoints
using processor's debug registers. Using Kwatch interface users can
monitor kernel global variables and dump the debugging information such
as kernel stack, global variables, processor registers.
int register_kwatch(unsigned long addr, u8 length, u8 type,
kwatch_handler_t handler)
-length of the breakpoint can be 1,2 or 4 bytes long.
-type can be read, write, execute.
0 Break on instruction execution only.
1 Break on data writes only.
3 Break on data reads or writes but not instruction fetches.
-return value is the debug register number allocated/used
for setting up this watch point.
Sample code:
This sample code sets a watchpoint on the pid_max
and registers a call back function if any writes
happen to pid_max.
struct kwatch kp;
void kwatch_handler(struct kwatch *p, struct pt_regs *regs)
{
.......<do-any-thing>........
}
Register watchpoint probe from init_module:
static int debug_regs_num;
int init_module(void)
{
..........<do-any-thing>............
debug_regs_num = register_kwatch(kallsyms_lookup_name(pid_max),
4, 1, kwatch_handler);
..........<do-any-thing>............
}
Test this by changing the value of pid_max in
/proc/sys/kernel/pid_max
echo 1000 > /proc/sys/kernel/pid_max
You see the call back function being called.
Unregister the watchpoint from cleanup_module:
void cleanup_module(void)
{
..........<do-any-thing>............
unregister_kwatch(debug_regs_num);
..........<do-any-thing>............
}
Signed-off-by: Prasanna S Panchamukhi <prasanna@in.ibm.com>
---
---
linux-2.6.13-prasanna/arch/i386/Kconfig.debug | 8 +
linux-2.6.13-prasanna/arch/i386/kernel/Makefile | 1
linux-2.6.13-prasanna/arch/i386/kernel/kwatch.c | 189 ++++++++++++++++++++++++
linux-2.6.13-prasanna/include/asm-i386/kwatch.h | 60 +++++++
4 files changed, 258 insertions(+)
diff -puN arch/i386/Kconfig.debug~kernel-watchpoint arch/i386/Kconfig.debug
--- linux-2.6.13/arch/i386/Kconfig.debug~kernel-watchpoint 2005-08-30 11:44:25.921069488 +0530
+++ linux-2.6.13-prasanna/arch/i386/Kconfig.debug 2005-08-30 11:44:25.932067816 +0530
@@ -32,6 +32,14 @@ config KPROBES
for kernel debugging, non-intrusive instrumentation and testing.
If in doubt, say "N".
+config KWATCH
+ bool "Kwatch points"
+ depends on DEBUG_KERNEL
+ select DEBUGREG
+ help
+ This enables kernel-space watchpoints using processor's debug
+ registers. If in doubt, say "N".
+
config DEBUGREG
bool "Global Debug Registers"
depends on DEBUG_KERNEL
diff -puN /dev/null arch/i386/kernel/kwatch.c
--- /dev/null 2005-08-30 16:04:24.253093808 +0530
+++ linux-2.6.13-prasanna/arch/i386/kernel/kwatch.c 2005-08-30 11:44:25.933067664 +0530
@@ -0,0 +1,189 @@
+/*
+ * Kernel Watchpoint interface.
+ * arch/i386/kernel/kwatch.c
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright (C) IBM Corporation, 2002, 2004
+ *
+ * 2002-Oct Created by Vamsi Krishna S <vamsi_krishna@in.ibm.com> for
+ * Kernel Watchpoint implementation.
+ * 2004-Oct Updated by Prasanna S Panchamukhi <prasanna@in.ibm.com> to
+ * to make use of notifiers.
+ */
+#include <linux/config.h>
+#include <linux/kprobes.h>
+#include <linux/ptrace.h>
+#include <linux/spinlock.h>
+#include <linux/module.h>
+#include <linux/init.h>
+#include <asm/kwatch.h>
+#include <asm/kdebug.h>
+#include <asm/debugreg.h>
+#include <asm/bitops.h>
+
+static struct kwatch kwatch_list[DR_MAX];
+static spinlock_t kwatch_lock = SPIN_LOCK_UNLOCKED;
+static unsigned long kwatch_in_progress; /* currently being handled */
+
+struct dr_info {
+ int debugreg;
+ unsigned long addr;
+ int type;
+};
+
+static inline void write_smp_dr(void *info)
+{
+ struct dr_info *dr = (struct dr_info *)info;
+
+ if (cpu_has_de && dr->type == DR_TYPE_IO)
+ set_in_cr4(X86_CR4_DE);
+ write_dr(dr->debugreg, dr->addr);
+}
+
+/* Update the debug register on all CPUs */
+static void sync_dr(int debugreg, unsigned long addr, int type)
+{
+ struct dr_info dr;
+ dr.debugreg = debugreg;
+ dr.addr = addr;
+ dr.type = type;
+ smp_call_function(write_smp_dr, &dr, 0, 0);
+}
+
+/*
+ * Interrupts are disabled on entry as trap1 is an interrupt gate and they
+ * remain disabled thorough out this function.
+ */
+int kwatch_handler(unsigned long condition, struct pt_regs *regs)
+{
+ int debugreg = dr_trap(condition);
+ unsigned long addr = dr_trap_addr(condition);
+ int retval = 0;
+
+ if (!(condition & (DR_TRAP0 | DR_TRAP1 | DR_TRAP2 | DR_TRAP3)))
+ return 0;
+
+ /* We're in an interrupt, but this is clear and BUG()-safe. */
+ preempt_disable();
+
+ /* If we are recursing, we already hold the lock. */
+ if (kwatch_in_progress)
+ goto recursed;
+
+ set_bit(debugreg, &kwatch_in_progress);
+
+ spin_lock(&kwatch_lock);
+ if (kwatch_list[debugreg].addr != addr)
+ goto out;
+
+ if (kwatch_list[debugreg].handler)
+ kwatch_list[debugreg].handler(&kwatch_list[debugreg], regs);
+
+ if (kwatch_list[debugreg].type == DR_TYPE_EXECUTE)
+ regs->eflags |= RF_MASK;
+ out:
+ clear_bit(debugreg, &kwatch_in_progress);
+ spin_unlock(&kwatch_lock);
+ preempt_enable_no_resched();
+ return retval;
+
+ recursed:
+ if (kwatch_list[debugreg].type == DR_TYPE_EXECUTE)
+ regs->eflags |= RF_MASK;
+ preempt_enable_no_resched();
+ return 1;
+}
+
+int register_kwatch(unsigned long addr, u8 length, u8 type,
+ kwatch_handler_t handler)
+{
+ int debugreg;
+ unsigned long dr7, flags;
+
+ debugreg = dr_alloc(DR_ANY, DR_ALLOC_GLOBAL);
+ if (debugreg < 0)
+ return -1;
+
+ spin_lock_irqsave(&kwatch_lock, flags);
+ kwatch_list[debugreg].addr = addr;
+ kwatch_list[debugreg].length = length;
+ kwatch_list[debugreg].type = type;
+ kwatch_list[debugreg].handler = handler;
+ spin_unlock_irqrestore(&kwatch_lock, flags);
+
+ write_dr(debugreg, (unsigned long)addr);
+ sync_dr(debugreg, (unsigned long)addr, type);
+ if (cpu_has_de && type == DR_TYPE_IO)
+ set_in_cr4(X86_CR4_DE);
+
+ dr7 = read_dr(7);
+ SET_DR7(dr7, debugreg, type, length);
+ write_dr(7, dr7);
+ sync_dr(7, dr7, 0);
+ return debugreg;
+}
+
+void unregister_kwatch(int debugreg)
+{
+ unsigned long flags;
+ unsigned long dr7 = read_dr(7);
+
+ RESET_DR7(dr7, debugreg);
+ write_dr(7, dr7);
+ sync_dr(7, dr7, 0);
+ dr_free(debugreg);
+
+ spin_lock_irqsave(&kwatch_lock, flags);
+ kwatch_list[debugreg].addr = 0;
+ kwatch_list[debugreg].handler = NULL;
+ spin_unlock_irqrestore(&kwatch_lock, flags);
+}
+
+/*
+ * Wrapper routine to for handling exceptions.
+ */
+int kwatch_exceptions_notify(struct notifier_block *self, unsigned long val,
+ void *data)
+{
+ struct die_args *args = (struct die_args *)data;
+ switch (val) {
+ case DIE_DEBUG:
+ if (kwatch_handler(args->err, args->regs))
+ return NOTIFY_STOP;
+ break;
+ default:
+ break;
+ }
+ return NOTIFY_DONE;
+}
+
+static struct notifier_block kwatch_exceptions_nb = {
+ .notifier_call = kwatch_exceptions_notify,
+ .priority = 0x7ffffffe /* we need to notified second */
+};
+
+static int __init init_kwatch(void)
+{
+ int err = 0;
+
+ err = register_die_notifier(&kwatch_exceptions_nb);
+ return err;
+}
+
+__initcall(init_kwatch);
+
+EXPORT_SYMBOL_GPL(register_kwatch);
+EXPORT_SYMBOL_GPL(unregister_kwatch);
diff -puN arch/i386/kernel/Makefile~kernel-watchpoint arch/i386/kernel/Makefile
--- linux-2.6.13/arch/i386/kernel/Makefile~kernel-watchpoint 2005-08-30 11:44:25.926068728 +0530
+++ linux-2.6.13-prasanna/arch/i386/kernel/Makefile 2005-08-30 11:44:25.934067512 +0530
@@ -35,6 +35,7 @@ obj-$(CONFIG_HPET_TIMER) += time_hpet.o
obj-$(CONFIG_EFI) += efi.o efi_stub.o
obj-$(CONFIG_EARLY_PRINTK) += early_printk.o
obj-$(CONFIG_DEBUGREG) += debugreg.o
+obj-$(CONFIG_KWATCH) += kwatch.o
EXTRA_AFLAGS := -traditional
diff -puN /dev/null include/asm-i386/kwatch.h
--- /dev/null 2005-08-30 16:04:24.253093808 +0530
+++ linux-2.6.13-prasanna/include/asm-i386/kwatch.h 2005-08-30 11:44:25.935067360 +0530
@@ -0,0 +1,60 @@
+#ifndef _ASM_KWATCH_H
+#define _ASM_KWATCH_H
+/*
+ * Kernel Watchpoint interface.
+ * include/asm-i386/kwatch.h
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License as published by
+ * the Free Software Foundation; either version 2 of the License, or
+ * (at your option) any later version.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write to the Free Software
+ * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
+ *
+ * Copyright (C) IBM Corporation, 2002, 2004
+ *
+ * 2002-Oct Created by Vamsi Krishna S <vamsi_krishna@in.ibm.com> for
+ * Kernel Watchpoint implementation.
+ */
+#include <linux/types.h>
+#include <linux/ptrace.h>
+
+struct kwatch;
+typedef void (*kwatch_handler_t) (struct kwatch *, struct pt_regs *);
+
+struct kwatch {
+ unsigned long addr; /* location of watchpoint */
+ u8 length; /* range of address */
+ u8 type; /* type of watchpoint */
+ kwatch_handler_t handler;
+};
+
+#define RF_MASK 0x00010000
+
+#ifdef CONFIG_KWATCH
+extern int register_kwatch(unsigned long addr, u8 length, u8 type,
+ kwatch_handler_t handler);
+extern void unregister_kwatch(int debugreg);
+extern int kwatch_handler(unsigned long condition, struct pt_regs *regs);
+#else
+static inline int register_kwatch(unsigned long addr, u8 length, u8 type,
+ kwatch_handler_t handler)
+{
+ return -ENOSYS;
+}
+static inline void unregister_kwatch(int debugreg)
+{
+}
+static inline int kwatch_handler(unsigned long condition, struct pt_regs *regs)
+{
+ return 0;
+}
+#endif
+#endif /* _ASM_KWATCH_H */
_
--
Prasanna S Panchamukhi
Linux Technology Center
India Software Labs, IBM Bangalore
Ph: 91-80-25044636
<prasanna@in.ibm.com>
^ permalink raw reply [flat|nested] 6+ messages in thread
* [patch 15/16] Allow KGDB to work well with loaded modules
[not found] ` <1.2982005.trini@kernel.crashing.org>
@ 2005-08-29 16:11 Tom Rini
[not found] ` <1.2982005.trini@kernel.crashing.org>
0 siblings, 1 reply; 6+ messages in thread
From: Tom Rini @ 2005-08-29 16:11 UTC (permalink / raw)
To: akpm; +Cc: linux-kernel, trini
This allows for KGDB to better deal with autoloaded modules. The way this
works, requires a patch to GDB. This patch can be found at
ftp://source.mvista.com/pub/kgdb/gdb-6.3-kgdb-module-notification.patch
The way this works is that the solib-search-path must contain the location of
any module to be debugged, and then when a module is loaded GDB acts like a
shared library has been loaded now, and can be used that way.
---
linux-2.6.13-trini/include/linux/module.h | 16 ++++++++
linux-2.6.13-trini/kernel/module.c | 56 ++++++++++++++++++++++++++++++
2 files changed, 72 insertions(+)
diff -puN include/linux/module.h~module include/linux/module.h
--- linux-2.6.13/include/linux/module.h~module 2005-08-29 09:03:42.000000000 -0700
+++ linux-2.6.13-trini/include/linux/module.h 2005-08-29 09:03:42.000000000 -0700
@@ -210,8 +210,17 @@ enum module_state
MODULE_STATE_LIVE,
MODULE_STATE_COMING,
MODULE_STATE_GOING,
+ MODULE_STATE_GONE,
};
+#ifdef CONFIG_KGDB
+#define MAX_SECTNAME 31
+struct mod_section {
+ void *address;
+ char name[MAX_SECTNAME + 1];
+};
+#endif
+
/* Similar stuff for section attributes. */
#define MODULE_SECT_NAME_LEN 32
struct module_sect_attr
@@ -239,6 +248,13 @@ struct module
/* Unique handle for this module */
char name[MODULE_NAME_LEN];
+#ifdef CONFIG_KGDB
+ /* keep kgdb info at the begining so that gdb doesn't have a chance to
+ * miss out any fields */
+ unsigned long num_sections;
+ struct mod_section *mod_sections;
+#endif
+
/* Sysfs stuff. */
struct module_kobject mkobj;
struct module_param_attrs *param_attrs;
diff -puN kernel/module.c~module kernel/module.c
--- linux-2.6.13/kernel/module.c~module 2005-08-29 09:03:42.000000000 -0700
+++ linux-2.6.13-trini/kernel/module.c 2005-08-29 09:03:42.000000000 -0700
@@ -623,6 +623,12 @@ sys_delete_module(const char __user *nam
if (ret != 0)
goto out;
+ down(¬ify_mutex);
+ notifier_call_chain(&module_notify_list, MODULE_STATE_GOING,
+ mod);
+ up(¬ify_mutex);
+
+
/* Never wait if forced. */
if (!forced && module_refcount(mod) != 0)
wait_for_zero_refcount(mod);
@@ -635,6 +641,11 @@ sys_delete_module(const char __user *nam
}
free_module(mod);
+ down(¬ify_mutex);
+ notifier_call_chain(&module_notify_list, MODULE_STATE_GONE,
+ NULL);
+ up(¬ify_mutex);
+
out:
up(&module_mutex);
return ret;
@@ -1173,6 +1184,11 @@ static void free_module(struct module *m
/* Arch-specific cleanup. */
module_arch_cleanup(mod);
+#ifdef CONFIG_KGDB
+ /* kgdb info */
+ vfree(mod->mod_sections);
+#endif
+
/* Module unload stuff */
module_unload_free(mod);
@@ -1407,6 +1423,31 @@ static void setup_modinfo(struct module
}
#endif
+#ifdef CONFIG_KGDB
+int add_modsects (struct module *mod, Elf_Ehdr *hdr, Elf_Shdr *sechdrs, const
+ char *secstrings)
+{
+ int i;
+
+ mod->num_sections = hdr->e_shnum - 1;
+ mod->mod_sections = vmalloc((hdr->e_shnum - 1)*
+ sizeof (struct mod_section));
+
+ if (mod->mod_sections == NULL) {
+ return -ENOMEM;
+ }
+
+ for (i = 1; i < hdr->e_shnum; i++) {
+ mod->mod_sections[i - 1].address = (void *)sechdrs[i].sh_addr;
+ strncpy(mod->mod_sections[i - 1].name, secstrings +
+ sechdrs[i].sh_name, MAX_SECTNAME);
+ mod->mod_sections[i - 1].name[MAX_SECTNAME] = '\0';
+ }
+
+ return 0;
+}
+#endif
+
#ifdef CONFIG_KALLSYMS
int is_exported(const char *name, const struct module *mod)
{
@@ -1775,6 +1816,12 @@ static struct module *load_module(void _
add_kallsyms(mod, sechdrs, symindex, strindex, secstrings);
+#ifdef CONFIG_KGDB
+ if ((err = add_modsects(mod, hdr, sechdrs, secstrings)) < 0) {
+ goto nomodsectinfo;
+ }
+#endif
+
err = module_finalize(hdr, sechdrs, mod);
if (err < 0)
goto cleanup;
@@ -1822,6 +1869,11 @@ static struct module *load_module(void _
arch_cleanup:
module_arch_cleanup(mod);
cleanup:
+
+#ifdef CONFIG_KGDB
+nomodsectinfo:
+ vfree(mod->mod_sections);
+#endif
module_unload_free(mod);
module_free(mod, mod->module_init);
free_core:
@@ -1909,6 +1961,10 @@ sys_init_module(void __user *umod,
/* Init routine failed: abort. Try to protect us from
buggy refcounters. */
mod->state = MODULE_STATE_GOING;
+ down(¬ify_mutex);
+ notifier_call_chain(&module_notify_list, MODULE_STATE_GOING,
+ mod);
+ up(¬ify_mutex);
synchronize_sched();
if (mod->unsafe)
printk(KERN_ERR "%s: module is now stuck!\n",
_
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2005-08-31 14:39 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2005-08-30 8:26 [patch 16/16] Add hardware breakpoint support for i386 Prasanna S Panchamukhi
2005-08-30 8:28 ` Prasanna S Panchamukhi
-- strict thread matches above, loose matches on Subject: below --
2005-08-29 16:11 [patch 15/16] Allow KGDB to work well with loaded modules Tom Rini
[not found] ` <1.2982005.trini@kernel.crashing.org>
2005-08-29 16:12 ` [patch 16/16] Add hardware breakpoint support for i386 Tom Rini
2005-08-29 21:23 ` Andi Kleen
2005-08-31 14:39 ` Tom Rini
2005-08-30 1:06 ` Keith Owens
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox
Powered by JetHome