| 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
| ||
|
Message-ID: <66615de5-4acb-8d85-6d69-ddd0b9609348@infradead.org>
Date: Mon, 23 Aug 2021 18:24:44 -0700
From: Randy Dunlap <rdunlap@...radead.org>
To: Stephen Rothwell <sfr@...b.auug.org.au>
Cc: Kees Cook <keescook@...omium.org>,
Andrew Morton <akpm@...ux-foundation.org>,
Linux Next Mailing List <linux-next@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
Nathan Chancellor <nathan@...nel.org>
Subject: Re: linux-next: Tree for Aug 20 (Wno-alloc-size-larger-than)
On 8/23/21 3:37 AM, Stephen Rothwell wrote:
> Hi all,
>
> On Sat, 21 Aug 2021 12:09:48 -0700 Randy Dunlap <rdunlap@...radead.org> wrote:
>>
>> On 8/20/21 10:48 PM, Kees Cook wrote:
>>> On Fri, Aug 20, 2021 at 02:54:05PM -0700, Randy Dunlap wrote:
>>>> On 8/20/21 2:26 AM, Stephen Rothwell wrote:
>>>>> Hi all,
>>>>>
>>>>> Changes since 20210819:
>>>>>
>>>>
>>>> Both linux-next and mmotm have many of these warnings when using
>>>> gcc (SUSE Linux) 7.5.0:
>>>>
>>>> cc1: warning: unrecognized command line option '-Wno-alloc-size-larger-than'
>>>
>>> Ew. Thanks for letting me know. I thought I'd verified this existed in
>>> gcc going back to 4.9, but it looks like I did something wrong in that
>>> test.
>>>
>>> I think this should fix it:
>>>
>>> diff --git a/Makefile b/Makefile
>>> index b0fafc41b686..e33ffa05899e 100644
>>> --- a/Makefile
>>> +++ b/Makefile
>>> @@ -1097,7 +1097,7 @@ endif
>>> ifdef CONFIG_CC_IS_GCC
>>> # The allocators already balk at large sizes, so silence the compiler
>>> # warnings for bounds checks involving those possible values.
>>> -KBUILD_CFLAGS += -Wno-alloc-size-larger-than
>>> +KBUILD_CFLAGS += $(call cc-option, -Wno-alloc-size-larger-than)
>>> endif
>>> > # disable invalid "can't wrap" optimizations for signed / pointers
>>>
>>
>> Well. That didn't help. This is very weird.
>>
>> This -Wno... option works (is accepted, no warning) on most files
>> that are being built, but a few files report an error with it:
>>
>> ../drivers/gpu/drm/radeon/radeon_object.c: At top level:
>> cc1: warning: unrecognized command line option '-Wno-alloc-size-larger-than'
>>
>> ../drivers/gpu/drm/amd/amdgpu/amdgpu_object.c: At top level:
>> cc1: warning: unrecognized command line option '-Wno-alloc-size-larger-than'
>>
>> ../drivers/hwmon/dell-smm-hwmon.c: At top level:
>> cc1: warning: unrecognized command line option '-Wno-alloc-size-larger-than'
>>
>> CC arch/x86/kernel/cpu/proc.o
>> cc1: warning: unrecognized command line option '-Wno-alloc-size-larger-than'
>>
>> ../arch/x86/kvm/mmu/mmu.c: At top level:
>> cc1: error: unrecognized command line option '-Wno-alloc-size-larger-than' [-Werror]
>>
>> ../drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_chardev.c: At top level:
>> cc1: warning: unrecognized command line option '-Wno-alloc-size-larger-than'
>>
>> ../kernel/trace/trace_osnoise.c: At top level:
>> cc1: warning: unrecognized command line option '-Wno-alloc-size-larger-than'
>>
>> CC kernel/dma/mapping.o
>> cc1: warning: unrecognized command line option '-Wno-alloc-size-larger-than'
>>
>>
>> It seems like it might be related to some .config option.
>>
>> I did a couple of partial builds with V=1 but that info didn't help me any.
>>
>>
>> If I am the only person seeing (reporting) this build warning, it could
>> just be (another) SUSE GCC-ism. (had one just last week with -Wmain and
>> kernel/trace/trace_osnoise.c)
>
> Today, I am also seeing thsese, but only with my sparc{,64} defconfig
> cross builds. This is with gcc 7.3.1 built from sources. I also just
> get a few of them.
>
> Also, I have the above "fix" patch applied (Andrew added it to mmots
> today).
>
> So something weird is happening.
This is just weird. What I am seeing is that for every source file
where gcc emits a warning: it then follows that up with this
>> cc1: warning: unrecognized command line option '-Wno-alloc-size-larger-than'
E.g.:
../kernel/printk/index.c:146:13: warning: ‘pi_remove_file’ defined but not used [-Wunused-function]
static void pi_remove_file(struct module *mod)
^~~~~~~~~~~~~~
cc1: warning: unrecognized command line option ‘-Wno-alloc-size-larger-than’
../fs/remap_range.c: In function ‘remap_verify_area’:
../fs/remap_range.c:102:16: warning: unused variable ‘inode’ [-Wunused-variable]
struct inode *inode = file_inode(file);
^~~~~
../fs/remap_range.c: At top level:
cc1: warning: unrecognized command line option ‘-Wno-alloc-size-larger-than’
../fs/locks.c: In function ‘fcntl_setlk64’:
../fs/locks.c:2509:16: warning: unused variable ‘inode’ [-Wunused-variable]
struct inode *inode = locks_inode(filp);
^~~~~
../fs/locks.c: At top level:
cc1: warning: unrecognized command line option ‘-Wno-alloc-size-larger-than’
../drivers/platform/chrome/./cros_ec_trace.h: In function ‘trace_raw_output_cros_ec_sensorhub_filter’:
../drivers/platform/chrome/./cros_ec_trace.h:156:12: warning: format ‘%d’ expects argument of type ‘int’, but argument 7 has type ‘s64 {aka long long int}’ [-Wformat=]
TP_printk("dx: %12lld. dy: %12lld median_m: %12lld median_error: %12lld len: %d x: %12lld y: %12lld",
^
../include/trace/trace_events.h:399:27: note: in definition of macro ‘DECLARE_EVENT_CLASS’
trace_event_printf(iter, print); \
^~~~~
../include/trace/trace_events.h:80:9: note: in expansion of macro ‘PARAMS’
PARAMS(print)); \
^~~~~~
../drivers/platform/chrome/./cros_ec_trace.h:135:1: note: in expansion of macro ‘TRACE_EVENT’
TRACE_EVENT(cros_ec_sensorhub_filter,
^~~~~~~~~~~
../drivers/platform/chrome/./cros_ec_trace.h:156:2: note: in expansion of macro ‘TP_printk’
TP_printk("dx: %12lld. dy: %12lld median_m: %12lld median_error: %12lld len: %d x: %12lld y: %12lld",
^~~~~~~~~
In file included from ../include/trace/trace_events.h:433:0,
from ../include/trace/define_trace.h:102,
from ../drivers/platform/chrome/cros_ec_trace.h:178,
from ../drivers/platform/chrome/cros_ec_trace.c:215:
../drivers/platform/chrome/./cros_ec_trace.h:156:80: note: format string is defined here
TP_printk("dx: %12lld. dy: %12lld median_m: %12lld median_error: %12lld len: %d x: %12lld y: %12lld",
~^
%lld
../drivers/platform/chrome/cros_ec_trace.c: At top level:
cc1: warning: unrecognized command line option ‘-Wno-alloc-size-larger-than’
warning: unsafe strscpy() usage lacked '__write_overflow' warning in ../lib/test_fortify/write_overflow-strscpy.c
CC drivers/base/driver.o
In file included from ../include/linux/string.h:253:0,
from ../arch/x86/include/asm/page_32.h:35,
from ../arch/x86/include/asm/page.h:14,
from ../arch/x86/include/asm/thread_info.h:12,
from ../include/linux/thread_info.h:60,
from ../arch/x86/include/asm/preempt.h:7,
from ../include/linux/preempt.h:78,
from ../include/linux/spinlock.h:55,
from ../include/linux/mmzone.h:8,
from ../include/linux/gfp.h:6,
from ../include/linux/slab.h:15,
from ../lib/test_fortify/test_fortify.h:4,
from ../lib/test_fortify/write_overflow-strscpy.c:5:
In function ‘strscpy’,
inlined from ‘do_fortify_tests’ at ../lib/test_fortify/test_fortify.h:34:2:
../include/linux/fortify-string.h:154:3: error: call to ‘__write_overflow’ declared with attribute error: detected write beyond size of object (1st parameter)
__write_overflow();
^~~~~~~~~~~~~~~~~~
../lib/test_fortify/write_overflow-strscpy.c: At top level:
cc1: error: unrecognized command line option ‘-Wno-alloc-size-larger-than’ [-Werror]
Smells like a gcc bug to me.
--
~Randy
Powered by blists - more mailing lists