Skip to contentSkip to conversation

uninitialized variables bugs

DISCUSSION Dan Carpenter 7 participants 7 messages May 6, 2022 – May 22, 2022

lore
Share & tools
lore
#0 lore raw

Ever since commit 78a5255ffb6a ("Stop the ad-hoc games with -Wno-maybe-initialized"), GCC's uninitialized variable warnings have been disabled by default. Now, you have to turn on W=1 or W=2 to see the warnings which nobody except Arnd does.

Disabling that has lead to a bunch of embarrassing bugs where variables are *never* initialized. Very unsubtle bugs. The bugs doesn't reach users because Nathan Chancellor and I review Clang and Smatch warnings respectively. Also the kbuild-bot reports uninitialized variables.

It's a lot to deal with. Uninitialized variable bugs are probably the most common bug I have to deal with.

It's frustrating. Sometimes the false positives are hard to analyse because I have to read through multiple functions. A lot of times when I write a patch and a commit message Nathan has already fixed it so it's just a waste of time.

It's risky as well. The Smatch check for uninitialized variables was broken for most of 2021. Nathan sometimes goes on vacation.

I guess I would hope that one day we can turn on the GCC uninitialized variable warnings again. That would mean silencing false positives which a lot of people don't want to do... Maybe Clang has fewer false positives than GCC?

The Smatch check for uninitialized variable was deliberately written to be more strict than GCC because GCC was missing bugs. So I think leaving Smatch false positives is fine. There is a trade off between fewer false positives and missing bugs and Smatch is meant to err on the side of finding bugs but with the cost of false positives.

Most of the Smatch uninitialized false positives are caused by loops:

	int i, ret;

	for (i = 0; i < bytes; i++) { // <-- what if bytes is zero?
		if (...)
			continue; // <-- can every iteration hit continue?
		ret = frob();
	}

	return ret;

There is also stuff like this which is harmless:

	uint val;

	ret = read(&val);
	*p = val;  // <-- uninitialized variable if read() fails
	return ret;

Btw, here is how to run Smatch on your code:
https://staticthinking.wordpress.com/2022/04/25/how-to-run-smatch-on-your-code/

regards,
dan carpenter

#1 lore raw
Quoted 46 lines · Dan Carpenter

Ever since commit 78a5255ffb6a ("Stop the ad-hoc games with -Wno-maybe-initialized"), GCC's uninitialized variable warnings have been disabled by default. Now, you have to turn on W=1 or W=2 to see the warnings which nobody except Arnd does.

Disabling that has lead to a bunch of embarrassing bugs where variables are *never* initialized. Very unsubtle bugs. The bugs doesn't reach users because Nathan Chancellor and I review Clang and Smatch warnings respectively. Also the kbuild-bot reports uninitialized variables.

It's a lot to deal with. Uninitialized variable bugs are probably the most common bug I have to deal with.

It's frustrating. Sometimes the false positives are hard to analyse because I have to read through multiple functions. A lot of times when I write a patch and a commit message Nathan has already fixed it so it's just a waste of time.

It's risky as well. The Smatch check for uninitialized variables was broken for most of 2021. Nathan sometimes goes on vacation.

I guess I would hope that one day we can turn on the GCC uninitialized variable warnings again. That would mean silencing false positives which a lot of people don't want to do... Maybe Clang has fewer false positives than GCC?

The Smatch check for uninitialized variable was deliberately written to be more strict than GCC because GCC was missing bugs. So I think leaving Smatch false positives is fine. There is a trade off between fewer false positives and missing bugs and Smatch is meant to err on the side of finding bugs but with the cost of false positives.

Most of the Smatch uninitialized false positives are caused by loops:

	int i, ret;

	for (i = 0; i < bytes; i++) { // <-- what if bytes is zero?
		if (...)
			continue; // <-- can every iteration hit continue?
		ret = frob();
	}

	return ret;

There is also stuff like this which is harmless:

Quoted 8 lines
	uint val;

	ret = read(&val);
	*p = val;  // <-- uninitialized variable if read() fails
	return ret;

Btw, here is how to run Smatch on your code:
https://staticthinking.wordpress.com/2022/04/25/how-to-run-smatch-on-your-code/

Could smatch inform the user that some results are likely false positives, or even order the results according to their likely true positiveness?

julia

#2 lore raw
In reply to Dan Carpenter
Quoted 5 lines · Dan Carpenter

It's frustrating. Sometimes the false positives are hard to analyse because I have to read through multiple functions. A lot of times when I write a patch and a commit message Nathan has already fixed it so it's just a waste of time.

Agreed. I'm not actually checking for those warnings on gcc any more, but just the clang warnings point to a bigger problem.

Quoted 7 lines

It's risky as well. The Smatch check for uninitialized variables was broken for most of 2021. Nathan sometimes goes on vacation.

I guess I would hope that one day we can turn on the GCC uninitialized variable warnings again. That would mean silencing false positives which a lot of people don't want to do... Maybe Clang has fewer false positives than GCC?

I think for the gcc warnings to become useful again, we may have to wait for a future compiler release. I have not checked gcc-12 for this, but it's a very old topic.

Fundamentally, it's impossible for any compiler to do this correctly, because of the halting problem. gcc apparently has some heuristics that worked well enough in the past, but it misses some obvious cases and causes false positives in unexpected places, often depending on optimization flags.

Recent gcc versions are much worse than older ones, since the inlining changed in a way that caused a ton of false-positives.

clang is generally better at catching the simple cases reliably, and it does this independent of optimization flags. However, it stops at the function boundary, so it never catches some of the cases that gcc was good at.

The gcc static analyzer apparently gained an option[1] that works similarly to what you have in smatch. I have not tried using this, but this may be something we can do in CI systems that may not want to run smatch for some reason.

Arnd

[1] https://gcc.gnu.org/onlinedocs/gcc-12.1.0/gcc/Static-Analyzer-Options.html#index-Wanalyzer-use-of-uninitialized-value

#3 lore raw
Quoted 19 lines · Arnd Bergmann
Quoted 5 lines · Dan Carpenter

It's frustrating. Sometimes the false positives are hard to analyse because I have to read through multiple functions. A lot of times when I write a patch and a commit message Nathan has already fixed it so it's just a waste of time.

Agreed. I'm not actually checking for those warnings on gcc any more, but just the clang warnings point to a bigger problem.

Quoted 7 lines

It's risky as well. The Smatch check for uninitialized variables was broken for most of 2021. Nathan sometimes goes on vacation.

I guess I would hope that one day we can turn on the GCC uninitialized variable warnings again. That would mean silencing false positives which a lot of people don't want to do... Maybe Clang has fewer false positives than GCC?

I would like to throw resource leak bugs in the mix. I am finding cppcheck has been effective in finding them.

I am seeing a lot of file pointer leaks in error legs in kselftest code error paths. I have a few fixes in the works to send out.

We could discuss this topic at the LPC Kernel Testing and Dependability mini-conf as well.

thanks,
-- Shuah

#4 lore raw
In reply to Dan Carpenter

Hi Dan,

Quoted 4 lines · Dan Carpenter

Ever since commit 78a5255ffb6a ("Stop the ad-hoc games with -Wno-maybe-initialized"), GCC's uninitialized variable warnings have been disabled by default. Now, you have to turn on W=1 or W=2 to see the warnings which nobody except Arnd does.

Thank you a lot for bringing this up; the situation does leave much to be desired from my side, as I am having to fix quite a number of these issues because people just are not seeing them and they break our builds because of CONFIG_WERROR (as they should).

Quoted 4 lines

Disabling that has lead to a bunch of embarrassing bugs where variables are *never* initialized. Very unsubtle bugs. The bugs doesn't reach users because Nathan Chancellor and I review Clang and Smatch warnings respectively. Also the kbuild-bot reports uninitialized variables.

Thankfully, I believe the situation is a lot less worse than it could be because the kbuild test robot tests with clang and finds these before they make it into any tree:

https://lore.kernel.org/llvm/?q=f%3Alkp%40intel.com+Wuninitialized
https://lore.kernel.org/llvm/?q=f%3Alkp%40intel.com+Wsometimes-uninitialized

Quoted 2 lines

It's a lot to deal with. Uninitialized variable bugs are probably the most common bug I have to deal with.

Agreed.

Quoted 4 lines

It's frustrating. Sometimes the false positives are hard to analyse because I have to read through multiple functions. A lot of times when I write a patch and a commit message Nathan has already fixed it so it's just a waste of time.

Sorry :( I should be better about either cc'ing you directly or adding the kernel-janitors mailing list, as there are others who would benefit from seeing these patches fly by. I know that isn't really the point of the email but I'll try to make your life easier in the future.

Quoted 7 lines

It's risky as well. The Smatch check for uninitialized variables was broken for most of 2021. Nathan sometimes goes on vacation.

I guess I would hope that one day we can turn on the GCC uninitialized variable warnings again. That would mean silencing false positives which a lot of people don't want to do... Maybe Clang has fewer false positives than GCC?

Yes, clang does have fewer false positives than GCC for a couple of reasons:

1. As Arnd touched on, Clang's -Wuninitialized and -Wsometimes-uninitialized do not check for initializations across function boundaries. In your example below with 'read(&val)', clang will assume that read() initializes val. While that does mean that there is slightly less coverage, it does drives the false positive rate way down, almost to zero. There are occasionally times where clang fails to figure out certain conditions which will avoid an uninitialized use but I believe that means the code is not as clear as it could be. For example, commit 118de6106735 ("net: ethernet: rmnet: Restructure if checks to avoid uninitialized warning").

2. clang used to only have these warnings under -Wconditional-uninitialized, which suffers from the same issue as -Wmaybe-uninitialized ("maybe it is uninitialized?"). -Wsometimes-uninitialized was split off from that warning back in 2011 to be more assertive ("this IS uninitialized if these conditions hold"):

https://github.com/llvm/llvm-project/commit/4323bf8e2e5135c49f814940b2b546298c01ecbc

Perhaps GCC could consider something to this?

Clang's static analyzer, which Tom regularly runs, will check variables across function boundaries. I am not sure what the false positive rate on that check is but it does turn up issues like smatch does.

Quoted 5 lines

The Smatch check for uninitialized variable was deliberately written to be more strict than GCC because GCC was missing bugs. So I think leaving Smatch false positives is fine. There is a trade off between fewer false positives and missing bugs and Smatch is meant to err on the side of finding bugs but with the cost of false positives.

I would agree with this too.

Cheers,
Nathan

#5 lore raw
Quoted 7 lines · Nathan Chancellor

Hi Dan,

Quoted 4 lines · Dan Carpenter

Ever since commit 78a5255ffb6a ("Stop the ad-hoc games with -Wno-maybe-initialized"), GCC's uninitialized variable warnings have been disabled by default. Now, you have to turn on W=1 or W=2 to see the warnings which nobody except Arnd does.

...

Quoted 3 lines

Clang's static analyzer, which Tom regularly runs, will check variables across function boundaries. I am not sure what the false positive rate on that check is but it does turn up issues like smatch does.

Clang's static analyzer is pretty go wrt uninitialized variables.

But the issues do not turn up in the report, the show up as errors and is why

I post a fix for a build break every couple of weeks.

Tom

#6 lore raw
In reply to Dan Carpenter
Quoted 10 lines · Dan Carpenter

There is also stuff like this which is harmless:

	uint val;

	ret = read(&val);
	*p = val;  // <-- uninitialized variable if read() fails
	return ret;

Btw, here is how to run Smatch on your code:
https://staticthinking.wordpress.com/2022/04/25/how-to-run-smatch-on-your-code/

In the topic of suppressing false positives we also have several "fixes", sometimes pointed out incorrectly by Coverity, for missing check for of_device_get_match_data().

Compare:
https://elixir.bootlin.com/linux/v5.18-rc7/source/drivers/clk/clk-aspeed.c#L415
https://elixir.bootlin.com/linux/v5.18-rc7/source/drivers/clk/clk-oxnas.c#L216

Although in theory the of_device_get_match_data() can return NULL, in practice it is not possible because driver matches via OF thus there will be always of_device_id->driver data.

Coverity screams about it, people fix it by adding checks for NULL, which is pointless. Half of drivers add the !NULL check, half do not...

Best regards,
Krzysztof