From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.remote-ip=100.103.45.18 ARC-Seal: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; cv=none; t=1794901734; b=oa20Z7pTiWxpHXzg2RGSIrypDEiG7pn1aXXTH8tJ0Nzcpcz8WfbRuuM9f+lcDED4QNlg 2asUqMxxa+YmxSdnM2GGt+Yp/fV4SN0UgoUOzdEDypfZrvhHyRDCMnxrdI9yNfawMEX+b JO1TuEwTfwSXfFMSn5buSjXIrSQjXZhj6joywpgRimXv3Cc8Rdg1yEHMjt/2SP25RIXey KUsRpnJGa+oSmUDmtfKUWKE5mdWz09kh6+pZArb3LMd/jp5EibyJc9YsYFaR9x9wbqc9r gpduOdrWYqJg8PdfXy6rODiS1lLm/oh+mbJM4KH7OEbHIIx8nRcK2uFePEjICM36sJQ== ARC-Message-Signature: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; c=relaxed/relaxed; t=1794901734; h=Received:DKIM-Signature:Message-ID:From:Subject:Reply-To:To:Cc: In-Reply-To:References:Content-Type:Content-Transfer-Encoding:Date; bh=FXO/KnGMUHe1lx1XKTfuGNvFYaS7k5EuhcB6VaSsrNU=; b=3xAXs8qxz2aSagmTvhZVqT9SdPIao7r6DiXtzMcsoPUlRZ17H1BxaygvZSolBa/WlLBU T3F7P9GAtlZEga5MuB1T6S7fjkHobqRNHkSSMdD2GByJSUHHo89zJNs3V4MbfJeBMp6Ln qPlRrMhVYXNj9Oayh4JJT5LcVgw3zCnbrvU/f0YBezAfHAj05xXsWy/hzhp6w4bx9Qz4U E1vv5jbGr9deDAbrgAI16KDxFtEbnfAfCNVN/NiAGWYo/jOIRISn+XjJRaHBPYncL6TJw zJ6X+mdRjMo5CTxrGi7ykb0diDtiNK+sGKc2yYKNF72FpcnVx0zQ4IPLjwJlolmOVDw== ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass header.d=kernel.org header.i=@kernel.org header.a=rsa-sha256 header.s=k20260515 header.b=dXVW2lhJ; arc=none smtp.remote-ip=100.103.45.18 Received: from smtp.kernel.org (quasi.space.kernel.org [100.103.45.18]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (prime256v1)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id C4999606A31C for ; Tue, 17 Nov 2026 07:48:54 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.a=rsa-sha256 header.s=k20260515 header.b=dXVW2lhJ Received: by smtp.kernel.org (Postfix) with ESMTPSA id 76EC61F000C4; Tue, 17 Nov 2026 07:48:54 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel.org; s=k20260515; t=1794901734; bh=FXO/KnGMUHe1lx1XKTfuGNvFYaS7k5EuhcB6VaSsrNU=; h=From:Subject:Reply-To:To:Cc:In-Reply-To:References:Date; b=dXVW2lhJMdjYorG6Zl6e3c06qKqpmbjfFVZdywqgZJ4n7yA4cws6+/91p206YrWUJ muPtuEtot7OaXPt4PKrn2CaFLEpfEYnL629qz/cytZucPGFTO4TU8+xQ54hehj8Dh6 NdoZI1u+skGJuqqGEgSVQxEW1Li0gfvNBySTGEEzBHJ1E0VBFzKS5JWlCZhfKdUfA5 dlnwofkoj+apVVoR0+0b4480mZilVsdC/SY0DxIwipcsEzWu76UGG9g9slTNAux186 wOCqIcD0WJnjtN5brHbJSw0bfKXc50xNON/yI30dyYuNyLOv17FksOwEcVilSOpxkj pXqsTJF0izVow== Message-ID: From: sashiko-bot@kernel.org Subject: Re: [PATCH v6 20/23] soundwire: amd: fix interrupt gate and work drain ordering in PM ops Reply-To: sashiko-reviews@lists.linux.dev To: "Vijendar Mukunda" Cc: linux-sound@vger.kernel.org In-Reply-To: <20261117070906.803139-21-Vijendar.Mukunda@amd.com> References: <20261117070906.803139-1-Vijendar.Mukunda@amd.com> <20261117070906.803139-21-Vijendar.Mukunda@amd.com> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Date: Tue, 17 Nov 2026 07:48:54 +0000 X-Mailing-List: linux-sound@vger.kernel.org List-Id: List-Post: List-Help: List-Subscribe: List-Unsubscribe: List-Owner: Precedence: list Thank you for your contribution! Sashiko AI review found 2 potential issues to consider: - [High] soundwire: amd: runtime PM deadlock in amd_suspend_runtime() work drain - [Medium] soundwire: amd: stale status bits and early IRQ enable in amd_suspend() -- commit 9fb00434cce0e9e4cd0523448b2e4bde7b333b08 Author: Vijendar Mukunda soundwire: amd: fix interrupt gate and work drain ordering in PM ops This commit reorders interrupt masking, workqueue draining, and wake configuration across the CLK_STOP_MODE and POWER_OFF_MODE paths in amd_suspend() and amd_suspend_runtime(). > diff --git a/drivers/soundwire/amd_manager.c b/drivers/soundwire/amd_mana= ger.c > index 57238c81997a2..220abf73cb25e 100644 > --- a/drivers/soundwire/amd_manager.c > +++ b/drivers/soundwire/amd_manager.c > @@ -1430,17 +1430,20 @@ static int __maybe_unused amd_suspend(struct devi= ce *dev) > } > =20 > if (amd_manager->power_mode_mask & AMD_SDW_CLK_STOP_MODE) { > + amd_disable_sdw_interrupts(amd_manager); > + cancel_work_sync(&amd_manager->amd_sdw_irq_thread); > cancel_work_sync(&amd_manager->amd_sdw_work); > - amd_sdw_wake_enable(amd_manager, false); > ret =3D amd_sdw_host_wake_enable_by_rev(amd_manager, false); > if (ret) > return ret; > ret =3D amd_sdw_clock_stop(amd_manager); > if (ret) > return ret; > + amd_enable_sdw_interrupts(amd_manager); [Severity: Medium] Can calling amd_enable_sdw_interrupts() after amd_sdw_clock_stop() without clearing ACP_SW_STATE_CHANGE_STATUS_0TO7 and ACP_SW_STATE_CHANGE_STATUS_8TO11 cause amd_sdw_irq_thread() to process stale peripheral status bits while the SoundWire clock is stopped? When a SoundWire interrupt fires just before amd_disable_sdw_interrupts(), the parent ACP hard IRQ handler clears ACP_EXTERNAL_INTR_STAT and schedules amd_sdw_irq_thread, but the status registers are only cleared at the end of amd_sdw_irq_thread(): drivers/soundwire/amd_manager.c:amd_sdw_irq_thread() { if (!READ_ONCE(amd_manager->interrupt_enabled)) return; status_change_8to11 =3D readl(amd_manager->mmio + ACP_SW_STATE_CHANGE_STAT= US_8TO11); status_change_0to7 =3D readl(amd_manager->mmio + ACP_SW_STATE_CHANGE_STATU= S_0TO7); ... } Because amd_disable_sdw_interrupts() sets interrupt_enabled to false and cancels amd_sdw_irq_thread without clearing those registers, any pending status bits (or bits latched during amd_sdw_clock_stop()) remain set in hardware. In amd_suspend_runtime(), when a subsequent wake event triggers amd_sdw_irq_thread(), amd_sdw_process_wake_event() only queues an asynchronous pm_request_resume() before falling through to read the stale status_change_8to11 and status_change_0to7 bits and schedule amd_sdw_work while the link is still down, resulting in MIPI command timeouts (-ETIMEDOUT). In addition, the parent ACP IRQ handlers do not check interrupt_enabled before calling schedule_work(&amd_manager->amd_sdw_irq_thread), so an in-flight hard IRQ can queue amd_sdw_irq_thread after cancel_work_sync() returns and run once amd_enable_sdw_interrupts() sets interrupt_enabled back to true. Finally, amd_sdw_clock_stop() returns 0 on its error paths, so if clock stop fails, amd_enable_sdw_interrupts() re-enables state-change interrupts with the bus clock still running. [ ... ] > @@ -1491,18 +1493,24 @@ static int __maybe_unused amd_suspend_runtime(str= uct device *dev) > return 0; > } > if (amd_manager->power_mode_mask & AMD_SDW_CLK_STOP_MODE) { > - amd_sdw_wake_enable(amd_manager, true); > + amd_disable_sdw_interrupts(amd_manager); > + cancel_work_sync(&amd_manager->amd_sdw_irq_thread); > + cancel_work_sync(&amd_manager->amd_sdw_work); > ret =3D amd_sdw_host_wake_enable_by_rev(amd_manager, true); > if (ret) > return ret; > ret =3D amd_sdw_clock_stop(amd_manager); > if (ret) > return ret; > + amd_enable_sdw_interrupts(amd_manager); > + amd_sdw_wake_enable(amd_manager, true); > } else if (amd_manager->power_mode_mask & AMD_SDW_POWER_OFF_MODE) { > + amd_disable_sdw_interrupts(amd_manager); > + cancel_work_sync(&amd_manager->amd_sdw_irq_thread); > + cancel_work_sync(&amd_manager->amd_sdw_work); [Severity: High] Can cancel_work_sync(&amd_manager->amd_sdw_work) in amd_suspend_runtime() deadlock with amd_sdw_update_slave_status_work()? kworker/0 kworker/1 rpm_suspend(amd_manager->dev) runtime_status =3D RPM_SUSPENDING amd_suspend_runtime() amd_sdw_update_slave_status_work() sdw_handle_slave_status() sdw_handle_slave_alerts() pm_runtime_get_sync(&slave->dev) rpm_resume(bus->dev, 0) rpm_resume(amd_manager->dev, 0) /* waits on RPM_SUSPENDING */ cancel_work_sync(&amd_sdw_work) <-- deadlock In the SoundWire device hierarchy, slave->dev is a child of bus->dev, which is a child of amd_manager->dev, so both children are in RPM_SUSPENDED when amd_manager->dev enters rpm_suspend(). If a peripheral alert schedules amd_sdw_work just before amd_disable_sdw_interrupts() masks interrupts, sdw_handle_slave_alerts() calls pm_runtime_get_sync(&slave->dev), which synchronously resumes the parent chain up to amd_manager->dev and sleeps on amd_manager->dev->power.wait_queue waiting for amd_suspend_runtime() to finish, while amd_suspend_runtime() waits in cancel_work_sync() for amd_sdw_work to complete. --=20 Sashiko AI review =C2=B7 https://sashiko.dev/#/patchset/20261117070906.8031= 39-1-Vijendar.Mukunda@amd.com?part=3D20