linux-kernel - Re: [PATCH 1/1] arm64: Add initial set of stack unwinder self tests

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]

Message-ID: <Yrx2xwbYMfcl8Qok@sirena.org.uk>
Date:   Wed, 29 Jun 2022 16:59:03 +0100
From:   Mark Brown <broonie@...nel.org>
To:     "Ivan T. Ivanov" <iivanov@...e.de>
Cc:     Catalin Marinas <catalin.marinas@....com>,
        Will Deacon <will@...nel.org>,
        Mark Rutland <mark.rutland@....com>,
        "Madhavan T . Venkataraman" <madvenka@...ux.microsoft.com>,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/1] arm64: Add initial set of stack unwinder self tests

On Fri, Jun 24, 2022 at 05:10:00PM +0300, Ivan T. Ivanov wrote:
> Add kunit tests for obvious cases where stack unwind could be needed.
> Like these:
> 
>  * Unwind a separate task
>  * Unwind starting from caller
>  * Unwind from irq context
>  * Unwind from kprobe handler called via ftrace
>  * Unwind from ftrace handler
>  * Unwind through kretprobed function
>  * Unwind from kretprobe handler
> 
> Tests are completely based on code used in s390 unwinder tests.
> Cases which where not relevant to aarch64 where removed and
> some places where adjusted to address aarch64 specifics.

I think this would be a bit easier to digest if it were a series which
builds things up with the test cases in individual patches, or at least
things like ftrace and kprobes split out a bit more, rather than every
single test all at once.  I've got a few *very* superficial comments
below, I think the code is fine but there's several moving pieces to
check.

> +/*
> + * Calls test_arch_stack_walk() which is handy wrapper of aarch64 unwind
> + * functionality, and verifies that the result contains unwindme_func2
> + *followed by unwindme_func1.

Missing space.

> +	ret = register_ftrace_function(fops);
> +	if (!ret) {
> +		ret = test_unwind_ftraced_func(u);
> +		unregister_ftrace_function(fops);
> +	} else {
> +		kunit_err(current_test,
> +			  "failed to register ftrace handler (%d)\n", ret);
> +	}

Shouldn't we return an error here?

Download attachment "signature.asc" of type "application/pgp-signature" (489 bytes)