Re: [PATCH v3 06/11] kselftest: arm64: fake_sigreturn_bad_magic

[Cristian Marussi <cristian.marussi@arm.com>]

Hi

On 13/08/2019 17:25, Dave Martin wrote:
> On Fri, Aug 02, 2019 at 06:02:55PM +0100, Cristian Marussi wrote:
>> Added a simple fake_sigreturn testcase which builds a ucontext_t
> 
> Add

Ok
> 
>> with a bad magic header and place it onto the stack.
>> Expects a SIGSEGV on test PASS.
>> This commit also introduces a common utility assembly function to
>> invoke a sigreturn using a fake provided sigframe.
>>
>> Signed-off-by: Cristian Marussi <cristian.marussi@arm.com>
>> ---
>>  tools/testing/selftests/arm64/signal/Makefile |  2 +-
>>  .../testing/selftests/arm64/signal/signals.S  | 64 +++++++++++++++++++
>>  .../arm64/signal/test_signals_utils.h         |  1 +
>>  .../arm64/signal/testcases/.gitignore         |  1 +
>>  .../testcases/fake_sigreturn_bad_magic.c      | 63 ++++++++++++++++++
>>  5 files changed, 130 insertions(+), 1 deletion(-)
>>  create mode 100644 tools/testing/selftests/arm64/signal/signals.S
>>  create mode 100644 tools/testing/selftests/arm64/signal/testcases/fake_sigreturn_bad_magic.c
>>
>> diff --git a/tools/testing/selftests/arm64/signal/Makefile b/tools/testing/selftests/arm64/signal/Makefile
>> index 8c8d08be4b0d..b3dcf315b5a4 100644
>> --- a/tools/testing/selftests/arm64/signal/Makefile
>> +++ b/tools/testing/selftests/arm64/signal/Makefile
>> @@ -80,7 +80,7 @@ endif
>>  # Common test-unit targets to build common-layout test-cases executables
>>  # Needs secondary expansion to properly include the testcase c-file in pre-reqs
>>  .SECONDEXPANSION:
>> -$(PROGS): test_signals.c test_signals_utils.c testcases/testcases.c $$@.c test_signals.h test_signals_utils.h testcases/testcases.h
>> +$(PROGS): test_signals.c test_signals_utils.c testcases/testcases.c signals.S $$@.c test_signals.h test_signals_utils.h testcases/testcases.h
>>  	@if [ ! -d $(khdr_dir) ]; then \
>>  		echo -n "\n!!! WARNING: $(khdr_dir) NOT FOUND."; \
>>  		echo "===>  Are you sure Kernel Headers have been installed properly ?\n"; \
>> diff --git a/tools/testing/selftests/arm64/signal/signals.S b/tools/testing/selftests/arm64/signal/signals.S
>> new file mode 100644
>> index 000000000000..6262b877400b
>> --- /dev/null
>> +++ b/tools/testing/selftests/arm64/signal/signals.S
>> @@ -0,0 +1,64 @@
>> +/*
>> + * SPDX-License-Identifier: GPL-2.0
>> + * Copyright (C) 2019 ARM Limited
>> + */
>> +
>> +#include <asm/unistd.h>
>> +
>> +.section        ".rodata", "a"
> 
> The section name is not usually quoted in .section (though I guess it
> works).
> 
Ok

>> +call_fmt:
>> +	.asciz "Calling sigreturn with fake sigframe sized:%zd at calculated SP @%08lX\n"
>> +
>> +.text
>> +
>> +.globl fake_sigreturn
>> +
>> +/*	fake_sigreturn	x0:&sigframe,  x1:sigframe_size,  x2:alignment_SP */
>> +fake_sigreturn:
>> +	mov x20, x0
>> +	mov x21, x1
>> +	mov x22, x2
>> +	mov x23, sp
>> +
>> +	/* create space on the stack for fake sigframe..."x22"-aligned */
>> +	mov x0, #0
> 
> Why to we move 0 into x0?  We clobber x0 again in the next instruction:
Yes unneeded.

>> +	add x0, x21, x22
>> +	sub x22, x22, #1
>> +	bic x0, x0, x22
>> +	sub x23, x23, x0
> 
> Can you explain the logic here?  I'm not sure I understand what this is
> trying to do exactly.

The attempt is to make space on the stack for a x22-bytes-aligned sigframe
of at least x21-bytes-size (assuming x22 is power-of-two):

- calculate the needed aligned space (sigframe_size + x22) & ~(x22 - 1)
- calculate new SP in x23: x3 = sp - needed_aligned_space
- later down... move SP to the calculated sp in x23
- copy the provided sigframe on such new SP

> I notice you further modify this in patch 11 -- see my comments there,
> where I also suggest an alternative way of specifying a misaligned frame
> that might be simpler.

I've seen your comments on patch 11. I think I'll simplify that as you suggested:
patch 11 was meant to add the capability to misalign the sigframe  on the SP
(which is needed in testcase contained in 11) but in fact is better if I move all
of the simplified logic here.

>> +
>> +	ldr x0, =call_fmt
>> +	mov x1, x21
>> +	mov x2, x23
>> +	bl printf
>> +
>> +	mov sp, x23
>> +
>> +	/* now fill it with the provided content... */
>> +	mov x0, sp
>> +	mov x1, x20
>> +	mov x2, x21
>> +	bl memcpy
>> +
>> +	/*
>> +	 * Here saving a last minute SP to current->token acts as a marker:
>> +	 * if we got here, we are successfully faking a sigreturn; in other
>> +	 * words we are sure no bad fatal signal has been raised till now
>> +	 * for unrelated reasons, so we should consider the possibl observed
> 
> typo: possible (or possibly?)
> 
possibly

>> +	 * fatal signal like SEGV coming from Kernel restore_sigframe() and
>> +	 * triggered as expected from our test-case.
>> +	 * For simplicity this assumes that current field 'token' is laid out
>> +	 * as first in struct tdescr
>> +	 */
>> +	ldr x0, current
>> +	str x23, [x0]
>> +	/* SP is already pointing back to the just built fake sigframe here */
>> +	mov x8, #__NR_rt_sigreturn
>> +	svc #0
>> +
>> +	/*
>> +	 * Above sigreturn should not return...looping here leads to a timeout
>> +	 * and ensure proper and clean test failure, instead of jumping around
>> +	 * on a potentially corrupted stack.
>> +	 */
>> +	b .
>> diff --git a/tools/testing/selftests/arm64/signal/test_signals_utils.h b/tools/testing/selftests/arm64/signal/test_signals_utils.h
>> index ce35be8ebc8e..2a71da7e6695 100644
>> --- a/tools/testing/selftests/arm64/signal/test_signals_utils.h
>> +++ b/tools/testing/selftests/arm64/signal/test_signals_utils.h
>> @@ -12,4 +12,5 @@ int test_run(struct tdescr *td);
>>  void test_result(struct tdescr *td);
>>  
>>  bool get_current_context(struct tdescr *td, ucontext_t *dest_uc);
>> +int fake_sigreturn(void *sigframe, size_t sz, int alignment);
>>  #endif
>> diff --git a/tools/testing/selftests/arm64/signal/testcases/.gitignore b/tools/testing/selftests/arm64/signal/testcases/.gitignore
>> index a48a118b1a1a..0ea6fdc3765c 100644
>> --- a/tools/testing/selftests/arm64/signal/testcases/.gitignore
>> +++ b/tools/testing/selftests/arm64/signal/testcases/.gitignore
>> @@ -4,3 +4,4 @@ mangle_pstate_invalid_mode_el1
>>  mangle_pstate_invalid_mode_el2
>>  mangle_pstate_invalid_mode_el3
>>  mangle_pstate_ssbs_regs
>> +fake_sigreturn_bad_magic
>> diff --git a/tools/testing/selftests/arm64/signal/testcases/fake_sigreturn_bad_magic.c b/tools/testing/selftests/arm64/signal/testcases/fake_sigreturn_bad_magic.c
>> new file mode 100644
>> index 000000000000..b4c063e02a7a
>> --- /dev/null
>> +++ b/tools/testing/selftests/arm64/signal/testcases/fake_sigreturn_bad_magic.c
>> @@ -0,0 +1,63 @@
>> +/* SPDX-License-Identifier: GPL-2.0 */
>> +/* Copyright (C) 2019 ARM Limited */
>> +
> 
> We should probably also include <signal.h> here, since we're using
> siginfo_t etc.

Yes indeed. It was included via some other headers only.
> 
>> +#include <stdio.h>
>> +#include <ucontext.h>
>> +
>> +#include "test_signals_utils.h"
>> +#include "testcases.h"
>> +
>> +struct fake_sigframe sf;
>> +
>> +static int fake_sigreturn_bad_magic_run(struct tdescr *td,
>> +					siginfo_t *si, ucontext_t *uc)
>> +{
>> +	size_t resv_sz, offset;
>> +	struct _aarch64_ctx *shead = GET_SF_RESV_HEAD(sf), *head;
>> +
>> +	/* just to fill the ucontext_t with something real */
>> +	if (!get_current_context(td, &sf.uc))
>> +		return 1;
>> +
>> +	resv_sz = GET_SF_RESV_SIZE(sf);
>> +	/*
>> +	 * find the terminator, preserving existing headers
>> +	 * and verify amount of spare room in __reserved area.
>> +	 */
>> +	head = get_terminator(shead, resv_sz, &offset);
>> +	/*
>> +	 * try stripping extra_context header when low on space:
>> +	 * we need at least 2*HDR_SZ space ... one for the KSFT_BAD_MAGIC
>> +	 * and the other for the usual terminator.
>> +	 */
>> +	if (head && resv_sz - offset < HDR_SZ * 2) {
> 
> Can we factor out this logic for finding space in the signal frame?
> 
> We do pretty much the same thing in all the fake_sigreturn tests...

Ok
> 
>> +		fprintf(stderr, "Low on space:%zd. Discarding extra_context.\n",
>> +			resv_sz - offset);
>> +		head = get_header(shead, EXTRA_MAGIC, resv_sz, &offset);
>> +	}
>> +	/* just give up and timeout if still not enough space */
> 
> Do we actually time out?  I don't see where we actually wait, so doesn't
> test_run() just fail immediately?
> 
> The same applies to all the other fake_sigreturn tests too.
> 
Right. It is probably a leftover.

SIGALRM is used as an extreme measure to kill tests gone bad, but this
can happen only once the fake sigframe has been effectively placed on the stack
and sigreturned.


>> +	if (head && resv_sz - offset >= HDR_SZ) {
> 
> Should this be HDR_SZ * 2 again?  We need space for the face header and
> space to write a terminator after it.

I'll fix in the new factored out code.

> 
>> +		fprintf(stderr, "Mangling template header. Spare space:%zd\n",
>> +			resv_sz - offset);
>> +		/*
>> +		 * use a well known NON existent bad magic...something
>> +		 * we should pretty sure won't be ever defined in Kernel
>> +		 */
>> +		head->magic = KSFT_BAD_MAGIC;
>> +		head->size = HDR_SZ;
>> +		write_terminator_record(GET_RESV_NEXT_HEAD(head));
>> +
>> +		ASSERT_BAD_CONTEXT(&sf.uc);
>> +		fake_sigreturn(&sf, sizeof(sf), 16);
>> +	}
>> +
>> +	return 1;
>> +}
> 
> [...]
> 
> Cheers
> ---Dave
> 

Cheers

Cristian