Message-Id: <20250601232937.3510379-89-sashal@kernel.org>
Date: Sun,  1 Jun 2025 19:29:21 -0400
From: Sasha Levin <sashal@...nel.org>
To: patches@...ts.linux.dev,
	stable@...r.kernel.org
Cc: I Hsin Cheng <richard120310@...il.com>,
	Mark Brown <broonie@...nel.org>,
	Sasha Levin <sashal@...nel.org>,
	lgirdwood@...il.com,
	perex@...ex.cz,
	tiwai@...e.com,
	yung-chuan.liao@...ux.intel.com,
	pierre-louis.bossart@...ux.dev,
	Vijendar.Mukunda@....com,
	gregkh@...uxfoundation.org,
	peterz@...radead.org,
	linux-sound@...r.kernel.org,
	linux-kernel@...r.kernel.org
Subject: [PATCH AUTOSEL 6.14 089/102] ASoC: intel/sdw_utils: Assign initial value in asoc_sdw_rt_amp_spk_rtd_init()

From: I Hsin Cheng <richard120310@...il.com>

[ Upstream commit 5fb3878216aece471af030b33a9fbef3babd8617 ]

Initialize "ret" with "-EINVAL" to handle cases where "strstr()" for
"codec_dai->component->name_prefix" doesn't find "-1" nor "-2". In that
case "name_prefix" is invalid because for current implementation it's
expected to have either "-1" or "-2" in it. (Maybe "-3", "-4" and so on
in the future.)

Link: https://scan5.scan.coverity.com/#/project-view/36179/10063?selectedIssue=1627120
Signed-off-by: I Hsin Cheng <richard120310@...il.com>
Link: https://patch.msgid.link/20250505185423.680608-1-richard120310@gmail.com
Signed-off-by: Mark Brown <broonie@...nel.org>
Signed-off-by: Sasha Levin <sashal@...nel.org>
---

Based on my comprehensive analysis of this commit, I can now provide my
backport recommendation: **YES** This commit should be backported to
stable kernel trees. Here's my extensive analysis: ## Bug Description
and Analysis The commit fixes a critical bug in the
`asoc_sdw_rt_amp_spk_rtd_init()` function in
`sound/soc/sdw_utils/soc_sdw_rt_amp.c`. The issue is that the `ret`
variable is declared but never initialized, and in certain execution
paths, it may be returned uninitialized. ### Specific Problem Details 1.
**Uninitialized Variable Issue**: The `ret` variable is declared on line
193 but never given an initial value. 2. **Problematic Code Path**: In
the `for_each_rtd_codec_dais()` loop (lines 204-209), `ret` is only
assigned a value if `strstr()` finds either "-1" or "-2" in the
`codec_dai->component->name_prefix`. However, if neither condition is
met, `ret` remains uninitialized. 3. **Return of Garbage Value**: When
the function reaches `return ret;` on line 211, if neither condition in
the loop was satisfied, the function returns an uninitialized value,
which could be any random memory content. ### Why This Is a Serious Bug
1. **Unpredictable Behavior**: The function may return success (0) or
failure (non-zero) randomly depending on stack contents, leading to
inconsistent system behavior. 2. **Audio Subsystem Impact**: The
function is used as an `rtd_init` callback for multiple Realtek audio
amplifier codecs (rt1308, rt1316, rt1318, rt1320) as evidenced by the
codec info list in `soc_sdw_utils.c`. 3. **Error Propagation**: The
return value is checked in `asoc_sdw_rtd_init()` and propagated upward -
a garbage positive value would cause audio initialization to fail
unexpectedly. ### Evidence from Code Analysis The function is called via
the `rtd_init` callback mechanism: ```c if
(codec_info->dais[dai_index].rtd_init) { ret =
codec_info->dais[dai_index].rtd_init(rtd, dai); if (ret) return ret; //
Failure propagated upward } ``` The fix correctly initializes `ret =
-EINVAL` to handle the case where the `name_prefix` doesn't contain the
expected "-1" or "-2" suffixes, which indicates an invalid configuration
that should properly return an error. ### Characteristics Supporting
Backport 1. **Clear Bug Fix**: This is a straightforward bug fix with no
new features or architectural changes. 2. **Minimal Risk**: The change
is one line adding initialization - extremely low risk of introducing
regressions. 3. **Wide Hardware Impact**: Affects multiple Realtek audio
amplifier codecs used in Intel SoundWire configurations. 4. **Stable
Tree Compatible**: Bug was introduced in v6.10 (commit 84e0a19adb73d),
making it applicable to stable kernels from v6.10 onward. 5. **Static
Analysis Found**: The bug was found by Coverity static analysis,
indicating it's a real issue that could manifest in practice. 6. **User
Impact**: Audio hardware initialization failures can significantly
impact user experience, especially on laptops and embedded systems using
these codecs. This commit perfectly fits the stable kernel criteria: it
fixes an important bug, has minimal risk, is contained to a specific
subsystem, and addresses undefined behavior that could cause random
failures in audio hardware initialization.

 sound/soc/sdw_utils/soc_sdw_rt_amp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sound/soc/sdw_utils/soc_sdw_rt_amp.c b/sound/soc/sdw_utils/soc_sdw_rt_amp.c
index 0538c252ba69b..83c2368170cb5 100644
--- a/sound/soc/sdw_utils/soc_sdw_rt_amp.c
+++ b/sound/soc/sdw_utils/soc_sdw_rt_amp.c
@@ -190,7 +190,7 @@ int asoc_sdw_rt_amp_spk_rtd_init(struct snd_soc_pcm_runtime *rtd, struct snd_soc
 	const struct snd_soc_dapm_route *rt_amp_map;
 	char codec_name[CODEC_NAME_SIZE];
 	struct snd_soc_dai *codec_dai;
-	int ret;
+	int ret = -EINVAL;
 	int i;
 
 	rt_amp_map = get_codec_name_and_route(dai, codec_name);
-- 
2.39.5

Powered by blists - more mailing lists