Re: [PATCH v2 2/6] usage.c: add a non-fatal bug() function to go with BUG()

From: Glen Choo <chooglen@google.com>
Date: May 31, 2022, 18:18 UTC
Message-ID: <kl6lr149efbf.fsf@chooglen-macbookpro.roam.corp.google.com>
In-Reply-To: <patch-v2-2.6-2d0527f86dc-20220531T164806Z-avarab@gmail.com>

Ævar Arnfjörð Bjarmason <avarab@gmail.com> writes:

Show 9 quoted lines

> +static void check_bug_if_BUG(void)
> +{
> +	if (!bug_called_must_BUG)
> +		return;
> +
> +	/* BUG_vfl() calls exit(), which calls us again */
> +	bug_called_must_BUG = 0;
> +	BUG("on exit(): had bug() call(s) in this process without explicit BUG_if_bug()");
> +}

Hm thinking out loud here, what if we set bug_called_must_BUG = 0 inside of BUG()? Then we'd be guaranteed that BUG() will never infinitely recurse.

Another benefit is that we can invoke bug() and follow up with BUG() to unconditionally exit without a bogus warning "had bug() call(s) without explicit BUG_if_bug()". This pattern could be useful for 6/6 [1], where we no longer exit after printing the "BUG:" messages.

Just an idea :)

[1] https://lore.kernel.org/git/patch-v2-6.6-cbbe0276966-20220531T164806Z-avarab@gmail.com