[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <20170703122226.97483c8a761da860721519bd@kernel.org>
Date: Mon, 3 Jul 2017 12:22:26 +0900
From: Masami Hiramatsu <mhiramat@...nel.org>
To: "Naveen N. Rao" <naveen.n.rao@...ux.vnet.ibm.com>
Cc: Ingo Molnar <mingo@...nel.org>,
Ananth N Mavinakayanahalli <ananth@...ux.vnet.ibm.com>,
linux-kernel@...r.kernel.org
Subject: Re: [RFC PATCH 1/2] kprobes: Simplify kprobe blacklist management
Hi Naveen,
On Thu, 29 Jun 2017 19:38:37 +0530
"Naveen N. Rao" <naveen.n.rao@...ux.vnet.ibm.com> wrote:
> Currently, the kprobe blacklist file in debugfs (usually
> /sys/kernel/debug/kprobes/blacklist) does not capture all address
> ranges that are blacklisted. For instance, functions annotated with
> __kprobes (between __kprobes_text_start and __kprobes_text_end) are
> blacklisted on all architectures. Certain architectures have additional
> address ranges that are blacklisted.
Ah, I see. The problem is that the blacklist is not populated for
arch-dependent areas.
>
> These address ranges are currently prevented from being kprobe'd by
> arch_within_kprobe_blacklist(). Instead, add the relevant address ranges
> to the main kprobe blacklist allowing such address ranges to be
> discovered by userspace.
>
> With this patch, on powerpc64le, we see 2 additional entries at the end
> of the blacklist:
> root@...ntu:/sys/kernel/debug# tail kprobes/blacklist
> 0xc00000000022d500-0xc00000000022d580 print_type_s64
> 0xc00000000022d480-0xc00000000022d500 print_type_s32
> 0xc00000000022d400-0xc00000000022d480 print_type_s16
> 0xc00000000022d380-0xc00000000022d400 print_type_s8
> 0xc00000000022d300-0xc00000000022d380 print_type_u64
> 0xc00000000022d280-0xc00000000022d300 print_type_u32
> 0xc00000000022d200-0xc00000000022d280 print_type_u16
> 0xc00000000022d180-0xc00000000022d200 print_type_u8
> 0xc0000000009ed778-0xc0000000009ed778 __irqentry_text_end
> 0xc000000000000000-0xc000000000008000 start_first_256B
>
> __irqentry_text_end() happens to be at __kprobes_text_start and hence
> figures there. Ditto with start_first_256B(). A subsequent patch helps
> provide meaningful names for such address ranges.
>
OK, looks good to me :)
Acked-by: Masami Hiramatsu <mhiramat@...nel.org>
Thank you!
> Signed-off-by: Naveen N. Rao <naveen.n.rao@...ux.vnet.ibm.com>
> ---
> arch/arm64/kernel/probes/kprobes.c | 31 +++++++++---------
> arch/powerpc/kernel/kprobes.c | 10 +++---
> arch/x86/kernel/kprobes/core.c | 10 +++---
> include/linux/kprobes.h | 2 ++
> kernel/kprobes.c | 64 ++++++++++++++++++++++----------------
> 5 files changed, 66 insertions(+), 51 deletions(-)
>
> diff --git a/arch/arm64/kernel/probes/kprobes.c b/arch/arm64/kernel/probes/kprobes.c
> index c5c45942fb6e..7f453a71b5a8 100644
> --- a/arch/arm64/kernel/probes/kprobes.c
> +++ b/arch/arm64/kernel/probes/kprobes.c
> @@ -535,24 +535,27 @@ int __kprobes longjmp_break_handler(struct kprobe *p, struct pt_regs *regs)
>
> bool arch_within_kprobe_blacklist(unsigned long addr)
> {
> - if ((addr >= (unsigned long)__kprobes_text_start &&
> - addr < (unsigned long)__kprobes_text_end) ||
> - (addr >= (unsigned long)__entry_text_start &&
> - addr < (unsigned long)__entry_text_end) ||
> - (addr >= (unsigned long)__idmap_text_start &&
> - addr < (unsigned long)__idmap_text_end) ||
> - !!search_exception_tables(addr))
> + if (!!search_exception_tables(addr))
> return true;
>
> + return false;
> +}
> +
> +void __init arch_populate_kprobe_blacklist(void)
> +{
> + insert_kprobe_blacklist((unsigned long)__kprobes_text_start,
> + (unsigned long)__kprobes_text_end);
> + insert_kprobe_blacklist((unsigned long)__entry_text_start,
> + (unsigned long)__entry_text_end);
> + insert_kprobe_blacklist((unsigned long)__idmap_text_start,
> + (unsigned long)__idmap_text_end);
> +
> if (!is_kernel_in_hyp_mode()) {
> - if ((addr >= (unsigned long)__hyp_text_start &&
> - addr < (unsigned long)__hyp_text_end) ||
> - (addr >= (unsigned long)__hyp_idmap_text_start &&
> - addr < (unsigned long)__hyp_idmap_text_end))
> - return true;
> + insert_kprobe_blacklist((unsigned long)__hyp_text_start,
> + (unsigned long)__hyp_text_end);
> + insert_kprobe_blacklist((unsigned long)__hyp_idmap_text_start,
> + (unsigned long)__hyp_idmap_text_end);
> }
> -
> - return false;
> }
>
> void __kprobes __used *trampoline_probe_handler(struct pt_regs *regs)
> diff --git a/arch/powerpc/kernel/kprobes.c b/arch/powerpc/kernel/kprobes.c
> index 01addfb0ed0a..f707e42b3482 100644
> --- a/arch/powerpc/kernel/kprobes.c
> +++ b/arch/powerpc/kernel/kprobes.c
> @@ -49,12 +49,12 @@ int is_current_kprobe_addr(unsigned long addr)
> return (p && (unsigned long)p->addr == addr) ? 1 : 0;
> }
>
> -bool arch_within_kprobe_blacklist(unsigned long addr)
> +void __init arch_populate_kprobe_blacklist(void)
> {
> - return (addr >= (unsigned long)__kprobes_text_start &&
> - addr < (unsigned long)__kprobes_text_end) ||
> - (addr >= (unsigned long)_stext &&
> - addr < (unsigned long)__head_end);
> + insert_kprobe_blacklist((unsigned long)__kprobes_text_start,
> + (unsigned long)__kprobes_text_end);
> + insert_kprobe_blacklist((unsigned long)_stext,
> + (unsigned long)__head_end);
> }
>
> kprobe_opcode_t *kprobe_lookup_name(const char *name, unsigned int offset)
> diff --git a/arch/x86/kernel/kprobes/core.c b/arch/x86/kernel/kprobes/core.c
> index 6b877807598b..4a5fc04b3e92 100644
> --- a/arch/x86/kernel/kprobes/core.c
> +++ b/arch/x86/kernel/kprobes/core.c
> @@ -1141,12 +1141,12 @@ int longjmp_break_handler(struct kprobe *p, struct pt_regs *regs)
> }
> NOKPROBE_SYMBOL(longjmp_break_handler);
>
> -bool arch_within_kprobe_blacklist(unsigned long addr)
> +void __init arch_populate_kprobe_blacklist(void)
> {
> - return (addr >= (unsigned long)__kprobes_text_start &&
> - addr < (unsigned long)__kprobes_text_end) ||
> - (addr >= (unsigned long)__entry_text_start &&
> - addr < (unsigned long)__entry_text_end);
> + insert_kprobe_blacklist((unsigned long)__kprobes_text_start,
> + (unsigned long)__kprobes_text_end);
> + insert_kprobe_blacklist((unsigned long)__entry_text_start,
> + (unsigned long)__entry_text_end);
> }
>
> int __init arch_init_kprobes(void)
> diff --git a/include/linux/kprobes.h b/include/linux/kprobes.h
> index 541df0b5b815..6d931ea0604e 100644
> --- a/include/linux/kprobes.h
> +++ b/include/linux/kprobes.h
> @@ -271,6 +271,8 @@ extern bool arch_function_offset_within_entry(unsigned long offset);
> extern bool function_offset_within_entry(kprobe_opcode_t *addr, const char *sym, unsigned long offset);
>
> extern bool within_kprobe_blacklist(unsigned long addr);
> +extern void insert_kprobe_blacklist(unsigned long start, unsigned long end);
> +extern void arch_populate_kprobe_blacklist(void);
>
> struct kprobe_insn_cache {
> struct mutex mutex;
> diff --git a/kernel/kprobes.c b/kernel/kprobes.c
> index adfe3b4cfe05..8ddbee01ce01 100644
> --- a/kernel/kprobes.c
> +++ b/kernel/kprobes.c
> @@ -1364,12 +1364,7 @@ static int register_aggr_kprobe(struct kprobe *orig_p, struct kprobe *p)
> return ret;
> }
>
> -bool __weak arch_within_kprobe_blacklist(unsigned long addr)
> -{
> - /* The __kprobes marked functions and entry code must not be probed */
> - return addr >= (unsigned long)__kprobes_text_start &&
> - addr < (unsigned long)__kprobes_text_end;
> -}
> +bool __weak arch_within_kprobe_blacklist(unsigned long addr) { return false; }
>
> bool within_kprobe_blacklist(unsigned long addr)
> {
> @@ -2125,12 +2120,40 @@ NOKPROBE_SYMBOL(dump_kprobe);
> * since a kprobe need not necessarily be at the beginning
> * of a function.
> */
> -static int __init populate_kprobe_blacklist(unsigned long *start,
> - unsigned long *end)
> +
> +/* Markers of _kprobe_blacklist section */
> +extern unsigned long __start_kprobe_blacklist[];
> +extern unsigned long __stop_kprobe_blacklist[];
> +
> +void __init insert_kprobe_blacklist(unsigned long start, unsigned long end)
> {
> - unsigned long *iter;
> struct kprobe_blacklist_entry *ent;
> +
> + ent = kmalloc(sizeof(*ent), GFP_KERNEL);
> + if (!ent) {
> + pr_err_once("kprobes: failed to populate blacklist: %d\n", -ENOMEM);
> + pr_err_once("Please take care of using kprobes.\n");
> + }
> +
> + ent->start_addr = start;
> + ent->end_addr = end;
> + INIT_LIST_HEAD(&ent->list);
> + list_add_tail(&ent->list, &kprobe_blacklist);
> +}
> +
> +void __weak __init arch_populate_kprobe_blacklist(void)
> +{
> + /* The __kprobes marked functions and entry code must not be probed */
> + insert_kprobe_blacklist((unsigned long)__kprobes_text_start,
> + (unsigned long)__kprobes_text_end);
> +}
> +
> +static void __init populate_kprobe_blacklist(void)
> +{
> + unsigned long *iter;
> unsigned long entry, offset = 0, size = 0;
> + unsigned long *start = __start_kprobe_blacklist;
> + unsigned long *end = __stop_kprobe_blacklist;
>
> for (iter = start; iter < end; iter++) {
> entry = arch_deref_entry_point((void *)*iter);
> @@ -2142,15 +2165,11 @@ static int __init populate_kprobe_blacklist(unsigned long *start,
> continue;
> }
>
> - ent = kmalloc(sizeof(*ent), GFP_KERNEL);
> - if (!ent)
> - return -ENOMEM;
> - ent->start_addr = entry;
> - ent->end_addr = entry + size;
> - INIT_LIST_HEAD(&ent->list);
> - list_add_tail(&ent->list, &kprobe_blacklist);
> + insert_kprobe_blacklist(entry, entry + size);
> }
> - return 0;
> +
> + /* Let architectures add their own entries as well */
> + arch_populate_kprobe_blacklist();
> }
>
> /* Module notifier call back, checking kprobes on the module */
> @@ -2202,10 +2221,6 @@ static struct notifier_block kprobe_module_nb = {
> .priority = 0
> };
>
> -/* Markers of _kprobe_blacklist section */
> -extern unsigned long __start_kprobe_blacklist[];
> -extern unsigned long __stop_kprobe_blacklist[];
> -
> static int __init init_kprobes(void)
> {
> int i, err = 0;
> @@ -2218,12 +2233,7 @@ static int __init init_kprobes(void)
> raw_spin_lock_init(&(kretprobe_table_locks[i].lock));
> }
>
> - err = populate_kprobe_blacklist(__start_kprobe_blacklist,
> - __stop_kprobe_blacklist);
> - if (err) {
> - pr_err("kprobes: failed to populate blacklist: %d\n", err);
> - pr_err("Please take care of using kprobes.\n");
> - }
> + populate_kprobe_blacklist();
>
> if (kretprobe_blacklist_size) {
> /* lookup the function address from its name */
> --
> 2.13.1
>
--
Masami Hiramatsu <mhiramat@...nel.org>
Powered by blists - more mailing lists