lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:   Sat, 15 Jan 2022 19:13:38 +0900
From:   Masami Hiramatsu <mhiramat@...nel.org>
To:     kernel test robot <lkp@...el.com>
Cc:     stable@...r.kernel.org,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        kbuild-all@...ts.01.org, linux-kernel@...r.kernel.org,
        Ingo Molnar <mingo@...nel.org>, Sasha Levin <sashal@...nel.org>
Subject: Re: [PATCH 5.4.y] Revert
 "ia64: kprobes: Use generic kretprobe trampoline handler"

On Sat, 15 Jan 2022 12:58:46 +0800
kernel test robot <lkp@...el.com> wrote:

> Hi Masami,
> 
> I love your patch! Perhaps something to improve:
> 
> [auto build test WARNING on stable/linux-5.4.y]
> 
> url:    https://github.com/0day-ci/linux/commits/Masami-Hiramatsu/Revert-ia64-kprobes-Use-generic-kretprobe-trampoline-handler/20220114-182111
> base:   https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git linux-5.4.y
> config: ia64-allmodconfig (https://download.01.org/0day-ci/archive/20220115/202201151231.g2sW8oWt-lkp@intel.com/config)
> compiler: ia64-linux-gcc (GCC) 11.2.0
> reproduce (this is a W=1 build):
>         wget https://raw.githubusercontent.com/intel/lkp-tests/master/sbin/make.cross -O ~/bin/make.cross
>         chmod +x ~/bin/make.cross
>         # https://github.com/0day-ci/linux/commit/514059de80b018e0edcf434519ff6bf41b4a519b
>         git remote add linux-review https://github.com/0day-ci/linux
>         git fetch --no-tags linux-review Masami-Hiramatsu/Revert-ia64-kprobes-Use-generic-kretprobe-trampoline-handler/20220114-182111
>         git checkout 514059de80b018e0edcf434519ff6bf41b4a519b
>         # save the config file to linux build tree
>         mkdir build_dir
>         COMPILER_INSTALL_PATH=$HOME/0day COMPILER=gcc-11.2.0 make.cross O=build_dir ARCH=ia64 SHELL=/bin/bash
> 
> If you fix the issue, kindly add following tag as appropriate
> Reported-by: kernel test robot <lkp@...el.com>
> 
> All warnings (new ones prefixed by >>):
> 
>    arch/ia64/kernel/kprobes.c: In function 'get_kprobe_inst':
>    arch/ia64/kernel/kprobes.c:325:22: warning: variable 'template' set but not used [-Wunused-but-set-variable]
>      325 |         unsigned int template;
>          |                      ^~~~~~~~

This one is fixed by recent patch.

https://lore.kernel.org/all/164208575387.1590449.8278421820882450166.stgit@devnote2/T/#u

>    arch/ia64/kernel/kprobes.c: At top level:
>    arch/ia64/kernel/kprobes.c:407:15: warning: no previous prototype for 'trampoline_probe_handler' [-Wmissing-prototypes]
>      407 | int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs)
>          |               ^~~~~~~~~~~~~~~~~~~~~~~~
>    arch/ia64/kernel/kprobes.c: In function 'trampoline_probe_handler':
> >> arch/ia64/kernel/kprobes.c:414:17: warning: initialization of 'long unsigned int' from 'void *' makes integer from pointer without a cast [-Wint-conversion]
>      414 |                 dereference_function_descriptor(kretprobe_trampoline);
>          |                 ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

Oops. I forgot to cast it. I'll resend the patch.

Thank you,


> 
> Kconfig warnings: (for reference only)
>    WARNING: unmet direct dependencies detected for FRAME_POINTER
>    Depends on DEBUG_KERNEL && (M68K || UML || SUPERH) || ARCH_WANT_FRAME_POINTERS
>    Selected by
>    - FAULT_INJECTION_STACKTRACE_FILTER && FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT && !X86_64 && !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM && !ARC && !X86
> 
> 
> vim +414 arch/ia64/kernel/kprobes.c
> 
>    320	
>    321	static void __kprobes get_kprobe_inst(bundle_t *bundle, uint slot,
>    322		       	unsigned long *kprobe_inst, uint *major_opcode)
>    323	{
>    324		unsigned long kprobe_inst_p0, kprobe_inst_p1;
>  > 325		unsigned int template;
>    326	
>    327		template = bundle->quad0.template;
>    328	
>    329		switch (slot) {
>    330		  case 0:
>    331			*major_opcode = (bundle->quad0.slot0 >> SLOT0_OPCODE_SHIFT);
>    332			*kprobe_inst = bundle->quad0.slot0;
>    333			  break;
>    334		  case 1:
>    335			*major_opcode = (bundle->quad1.slot1_p1 >> SLOT1_p1_OPCODE_SHIFT);
>    336			kprobe_inst_p0 = bundle->quad0.slot1_p0;
>    337			kprobe_inst_p1 = bundle->quad1.slot1_p1;
>    338			*kprobe_inst = kprobe_inst_p0 | (kprobe_inst_p1 << (64-46));
>    339			break;
>    340		  case 2:
>    341			*major_opcode = (bundle->quad1.slot2 >> SLOT2_OPCODE_SHIFT);
>    342			*kprobe_inst = bundle->quad1.slot2;
>    343			break;
>    344		}
>    345	}
>    346	
>    347	/* Returns non-zero if the addr is in the Interrupt Vector Table */
>    348	static int __kprobes in_ivt_functions(unsigned long addr)
>    349	{
>    350		return (addr >= (unsigned long)__start_ivt_text
>    351			&& addr < (unsigned long)__end_ivt_text);
>    352	}
>    353	
>    354	static int __kprobes valid_kprobe_addr(int template, int slot,
>    355					       unsigned long addr)
>    356	{
>    357		if ((slot > 2) || ((bundle_encoding[template][1] == L) && slot > 1)) {
>    358			printk(KERN_WARNING "Attempting to insert unaligned kprobe "
>    359					"at 0x%lx\n", addr);
>    360			return -EINVAL;
>    361		}
>    362	
>    363		if (in_ivt_functions(addr)) {
>    364			printk(KERN_WARNING "Kprobes can't be inserted inside "
>    365					"IVT functions at 0x%lx\n", addr);
>    366			return -EINVAL;
>    367		}
>    368	
>    369		return 0;
>    370	}
>    371	
>    372	static void __kprobes save_previous_kprobe(struct kprobe_ctlblk *kcb)
>    373	{
>    374		unsigned int i;
>    375		i = atomic_add_return(1, &kcb->prev_kprobe_index);
>    376		kcb->prev_kprobe[i-1].kp = kprobe_running();
>    377		kcb->prev_kprobe[i-1].status = kcb->kprobe_status;
>    378	}
>    379	
>    380	static void __kprobes restore_previous_kprobe(struct kprobe_ctlblk *kcb)
>    381	{
>    382		unsigned int i;
>    383		i = atomic_read(&kcb->prev_kprobe_index);
>    384		__this_cpu_write(current_kprobe, kcb->prev_kprobe[i-1].kp);
>    385		kcb->kprobe_status = kcb->prev_kprobe[i-1].status;
>    386		atomic_sub(1, &kcb->prev_kprobe_index);
>    387	}
>    388	
>    389	static void __kprobes set_current_kprobe(struct kprobe *p,
>    390				struct kprobe_ctlblk *kcb)
>    391	{
>    392		__this_cpu_write(current_kprobe, p);
>    393	}
>    394	
>    395	static void kretprobe_trampoline(void)
>    396	{
>    397	}
>    398	
>    399	/*
>    400	 * At this point the target function has been tricked into
>    401	 * returning into our trampoline.  Lookup the associated instance
>    402	 * and then:
>    403	 *    - call the handler function
>    404	 *    - cleanup by marking the instance as unused
>    405	 *    - long jump back to the original return address
>    406	 */
>    407	int __kprobes trampoline_probe_handler(struct kprobe *p, struct pt_regs *regs)
>    408	{
>    409		struct kretprobe_instance *ri = NULL;
>    410		struct hlist_head *head, empty_rp;
>    411		struct hlist_node *tmp;
>    412		unsigned long flags, orig_ret_address = 0;
>    413		unsigned long trampoline_address =
>  > 414			dereference_function_descriptor(kretprobe_trampoline);
>    415	
>    416		INIT_HLIST_HEAD(&empty_rp);
>    417		kretprobe_hash_lock(current, &head, &flags);
>    418	
>    419		/*
>    420		 * It is possible to have multiple instances associated with a given
>    421		 * task either because an multiple functions in the call path
>    422		 * have a return probe installed on them, and/or more than one return
>    423		 * return probe was registered for a target function.
>    424		 *
>    425		 * We can handle this because:
>    426		 *     - instances are always inserted at the head of the list
>    427		 *     - when multiple return probes are registered for the same
>    428		 *       function, the first instance's ret_addr will point to the
>    429		 *       real return address, and all the rest will point to
>    430		 *       kretprobe_trampoline
>    431		 */
>    432		hlist_for_each_entry_safe(ri, tmp, head, hlist) {
>    433			if (ri->task != current)
>    434				/* another task is sharing our hash bucket */
>    435				continue;
>    436	
>    437			orig_ret_address = (unsigned long)ri->ret_addr;
>    438			if (orig_ret_address != trampoline_address)
>    439				/*
>    440				 * This is the real return address. Any other
>    441				 * instances associated with this task are for
>    442				 * other calls deeper on the call stack
>    443				 */
>    444				break;
>    445		}
>    446	
>    447		regs->cr_iip = orig_ret_address;
>    448	
>    449		hlist_for_each_entry_safe(ri, tmp, head, hlist) {
>    450			if (ri->task != current)
>    451				/* another task is sharing our hash bucket */
>    452				continue;
>    453	
>    454			if (ri->rp && ri->rp->handler)
>    455				ri->rp->handler(ri, regs);
>    456	
>    457			orig_ret_address = (unsigned long)ri->ret_addr;
>    458			recycle_rp_inst(ri, &empty_rp);
>    459	
>    460			if (orig_ret_address != trampoline_address)
>    461				/*
>    462				 * This is the real return address. Any other
>    463				 * instances associated with this task are for
>    464				 * other calls deeper on the call stack
>    465				 */
>    466				break;
>    467		}
>    468		kretprobe_assert(ri, orig_ret_address, trampoline_address);
>    469	
>    470		kretprobe_hash_unlock(current, &flags);
>    471	
>    472		hlist_for_each_entry_safe(ri, tmp, &empty_rp, hlist) {
>    473			hlist_del(&ri->hlist);
>    474			kfree(ri);
>    475		}
>    476		/*
>    477		 * By returning a non-zero value, we are telling
>    478		 * kprobe_handler() that we don't want the post_handler
>    479		 * to run (and have re-enabled preemption)
>    480		 */
>    481		return 1;
>    482	}
>    483	
> 
> ---
> 0-DAY CI Kernel Test Service, Intel Corporation
> https://lists.01.org/hyperkitty/list/kbuild-all@lists.01.org


-- 
Masami Hiramatsu <mhiramat@...nel.org>

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ