linux-kernel - Re: [PATCH] staging: greybus: fix stack size warning with UBSAN

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: <d9c341c9-9005-b425-9dd8-e797869bbcb0@ieee.org>
Date:   Sun, 3 Jan 2021 20:18:47 -0600
From:   Alex Elder <elder@...e.org>
To:     Arnd Bergmann <arnd@...nel.org>,
        Vaibhav Agarwal <vaibhav.sr@...il.com>,
        Mark Greer <mgreer@...malcreek.com>,
        Johan Hovold <johan@...nel.org>, Alex Elder <elder@...nel.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Nathan Chancellor <natechancellor@...il.com>,
        Nick Desaulniers <ndesaulniers@...gle.com>
Cc:     Arnd Bergmann <arnd@...db.de>,
        Dan Carpenter <dan.carpenter@...cle.com>,
        Coiby Xu <coiby.xu@...il.com>, greybus-dev@...ts.linaro.org,
        devel@...verdev.osuosl.org, linux-kernel@...r.kernel.org,
        clang-built-linux@...glegroups.com
Subject: Re: [PATCH] staging: greybus: fix stack size warning with UBSAN

On 1/3/21 4:35 PM, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@...db.de>
> 
> clang warns about excessive stack usage in this driver when
> UBSAN is enabled:
> 
> drivers/staging/greybus/audio_topology.c:977:12: error: stack frame size of 1836 bytes in function 'gbaudio_tplg_create_widget' [-Werror,-Wframe-larger-than=]
> 
> Rework this code to no longer use compound literals for
> initializing the structure in each case, but instead keep
> the common bits in a preallocated constant array and copy
> them as needed.

This is good, but I have a few comments.

You took out the default case, and it seems you are using
a w->type value bigger than the initialization array to
determine validity.  But there are more values defined in
the snd_soc_dapm_type enumerated type than are explicitly
listed as cases in the switch statement.  So the switch
statement no longer treats some types as invalid (such
as snd_soc_dapm_demux).  Am I missing something?

You are setting explicit names, such as "spk", "hp",
"mic", etc. in the initialization array.  But you
update the name after (struct) assigning from the
array.  I have no real objection I guess, but why
bother?  Why not just use null pointers in the
initialization array?

You change a semicolon to a comma in one spot, and you
should not have.  I'll point that out below.

I like that you got rid of the type casts, which were
apparently unnecessary.

You dropped the break in the final case in the switch
statement, but in an earlier discussion I think we
concluded that wasn't a problem.

I guess the main thing is the first thing mentioned.


Thanks.

					-Alex

> Signed-off-by: Arnd Bergmann <arnd@...db.de>
> ---
>   drivers/staging/greybus/audio_topology.c | 106 ++++++++++-------------
>   1 file changed, 47 insertions(+), 59 deletions(-)
> 
> diff --git a/drivers/staging/greybus/audio_topology.c b/drivers/staging/greybus/audio_topology.c
> index 96b8b29fe899..c03873915c20 100644
> --- a/drivers/staging/greybus/audio_topology.c
> +++ b/drivers/staging/greybus/audio_topology.c
> @@ -974,6 +974,44 @@ static int gbaudio_widget_event(struct snd_soc_dapm_widget *w,
>   	return ret;
>   }
>   
> +static const struct snd_soc_dapm_widget gbaudio_widgets[] = {
> +	[snd_soc_dapm_spk]	= SND_SOC_DAPM_SPK("spk", gbcodec_event_spk),
> +	[snd_soc_dapm_hp]	= SND_SOC_DAPM_HP("hp", gbcodec_event_hp),
> +	[snd_soc_dapm_mic]	= SND_SOC_DAPM_MIC("mic", gbcodec_event_int_mic),

. . .

> @@ -1050,78 +1088,28 @@ static int gbaudio_tplg_create_widget(struct gbaudio_module_info *module,
>   	strlcpy(temp_name, w->name, NAME_SIZE);
>   	snprintf(w->name, NAME_SIZE, "GB %d %s", module->dev_id, temp_name);
>   
> +	if (w->type > ARRAY_SIZE(gbaudio_widgets)) {
> +		ret = -EINVAL;
> +		goto error;
> +	}
> +	*dw = gbaudio_widgets[w->type];
> +	dw->name = w->name;
> +
>   	switch (w->type) {
>   	case snd_soc_dapm_spk:
> -		*dw = (struct snd_soc_dapm_widget)
> -			SND_SOC_DAPM_SPK(w->name, gbcodec_event_spk);
>   		module->op_devices |= GBAUDIO_DEVICE_OUT_SPEAKER;
>   		break;
>   	case snd_soc_dapm_hp:
> -		*dw = (struct snd_soc_dapm_widget)
> -			SND_SOC_DAPM_HP(w->name, gbcodec_event_hp);
>   		module->op_devices |= (GBAUDIO_DEVICE_OUT_WIRED_HEADSET
> -					| GBAUDIO_DEVICE_OUT_WIRED_HEADPHONE);
> +					| GBAUDIO_DEVICE_OUT_WIRED_HEADPHONE),

Please fix this (above) to preserve the original semicolon.

>   		module->ip_devices |= GBAUDIO_DEVICE_IN_WIRED_HEADSET;
>   		break;
>   	case snd_soc_dapm_mic:
> -		*dw = (struct snd_soc_dapm_widget)
> -			SND_SOC_DAPM_MIC(w->name, gbcodec_event_int_mic);
>   		module->ip_devices |= GBAUDIO_DEVICE_IN_BUILTIN_MIC;
>   		break;

. . .