From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-alma10-1.taild15c8.ts.net [100.103.45.18]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 5873B314D35 for ; Fri, 17 Jul 2026 18:44:37 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=100.103.45.18 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784313878; cv=none; b=HzqEBd31b4Q5zdlbAqxCgBYfY7y9SfLgDlk1VQU2GxmkFA2ntCx3ZCv4bS7FDmJPYeMjC4zL6QdXQfbJqVShJZ7+zZUeNpfm652xebzhD3MMXtC55gX5uLl66WQASE3QqJI2XwN8XdPYdFWVoRTXqGKUWdmBvT4lI4629imKao8= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1784313878; c=relaxed/simple; bh=GOV1ezgMbO1uXM8Gou7nJIVdWQIq5yqV2eXceQFoA9w=; h=From:Subject:To:Cc:In-Reply-To:References:Content-Type:Date: Message-Id; b=kqMroQFUTs8zjJlIfN8R+G65LGRrhlDDgTaUYT+dcwxPhu3PObZVNeHFjsvdUKAUt45z+DyYEyevS0A1yigDDVPc+cHuLbocfqxF9eS6x6VIAbj71J4B0+IG5Ty7AdbX8n7Hdpprrb4FEbm37MNFOtyPeKN4stOeCYc2JYk5IjY= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=BLAGbwee; arc=none smtp.client-ip=100.103.45.18 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="BLAGbwee" Received: by smtp.kernel.org (Postfix) with ESMTPSA id D6D211F000E9; Fri, 17 Jul 2026 18:44:36 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=kernel.org; s=k20260515; t=1784313877; bh=zGUMs6nCQ2z1AJ0q//aPA8GlHfy54BsA7Q94Klh5j44=; h=From:Subject:Reply-To:To:Cc:In-Reply-To:References:Date; b=BLAGbweeCUIAl/V/z0mrydWyaanFrN7Pu41Znkfj+qf68BqTXFaAFT6QlxPe07eTL VmNm9/tSp2BwJskvpeiHzKGoJfUMAqDmrAiLq1hVlssS1EobePDNPP4lPmpeLZv+vO OwSbdEuhB0hVj4iQATeY/trH8K6hQQuCEZ9HCZrv/tUQWPEgDYfa1F5vzYRaNaZVNJ PHUO3fbEjwfhQdESndOevaHZ4d/69EWCm2L/5Ybr8RX1T15WOpGOn1vw0QjDO0upmf yZehA9IZnCRgN2p+UcsikERT+4ViIAGVG/3HfiM9TAPNqWsSQB5MdsaKIwoabLAKsw M37JL63B4vt0A== From: sashiko-bot@kernel.org Subject: Re: [PATCH v3 06/10] ACPI: extlog: Defer CXL protocol error handling to avoid lock inversion Reply-To: sashiko-reviews@lists.linux.dev To: "Dave Jiang" Cc: linux-cxl@vger.kernel.org In-Reply-To: <20260717161647.1493259-7-dave.jiang@intel.com> References: <20260717161647.1493259-1-dave.jiang@intel.com> <20260717161647.1493259-7-dave.jiang@intel.com> Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Date: Fri, 17 Jul 2026 18:44:36 +0000 Message-Id: <20260717184436.D6D211F000E9@smtp.kernel.org> Precedence: bulk X-Mailing-List: linux-cxl@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: Thank you for your contribution! Sashiko AI review found 2 potential issue(= s) to consider: Pre-existing issues: - [High] Preexisting lock inversion (AA deadlock) between process context a= nd IRQ context on CPER work locks. - [Medium] extlog_print_pcie() is compiled out due to a misspelled preproce= ssor macro, rendering the function and its newly added bounds check dead co= de. -- --- Patch [6]: [PATCH v3 06/10] ACPI: extlog: Defer CXL protocol error hand= ling to avoid lock inversion --- commit 004b48e5ad5ad98c6645296077a38ad36a959b26 Author: Dave Jiang ACPI: extlog: Defer CXL protocol error handling to avoid lock inversion The commit exports cxl_cper_post_prot_err() from ghes.c and modifies extlog_print() to use it. This avoids an AB-BA deadlock by deferring CXL protocol error handling to a workqueue instead of calling it synchronously while holding the MCE notifier chain rwsem. Link: https://lore.kernel.org/linux-cxl/20260709165457.8BA181F000E9@smtp.ke= rnel.org/ > diff --git a/drivers/acpi/acpi_extlog.c b/drivers/acpi/acpi_extlog.c > index 0c440d75d9a78..ae79d090de33b 100644 > --- a/drivers/acpi/acpi_extlog.c > +++ b/drivers/acpi/acpi_extlog.c [ ... ] > @@ -244,9 +227,9 @@ static int extlog_print(struct notifier_block *nb, un= signed long val, > struct cxl_cper_sec_prot_err *prot_err =3D > acpi_hest_get_payload(gdata); > =20 > - extlog_cxl_cper_handle_prot_err(prot_err, > - gdata->error_severity, > - gdata->error_data_length); > + cxl_cper_post_prot_err(prot_err, > + gdata->error_severity, > + gdata->error_data_length); > } else if (guid_equal(sec_type, &CPER_SEC_PCIE)) { > struct cper_sec_pcie *pcie_err =3D acpi_hest_get_payload(gdata); > =20 [Severity: Medium] This isn't a bug introduced by this patch, but does the extlog_print_pcie() function called for this section type get entirely compiled out? Looking at extlog_print_pcie() in drivers/acpi/acpi_extlog.c: drivers/acpi/acpi_extlog.c:extlog_print_pcie() { #ifdef ACPI_APEI_PCIEAER ... } Kernel Kconfig macros must use the CONFIG_ prefix to evaluate to true. Since it checks for ACPI_APEI_PCIEAER instead of CONFIG_ACPI_APEI_PCIEAER, does t= his silently disable PCIe AER error logging via the extlog driver and render the newly added bounds check dead code? > diff --git a/drivers/acpi/apei/ghes.c b/drivers/acpi/apei/ghes.c > index 17e4ef5552925..b8dbd99da47ef 100644 > --- a/drivers/acpi/apei/ghes.c > +++ b/drivers/acpi/apei/ghes.c > @@ -752,8 +752,8 @@ static DEFINE_KFIFO(cxl_cper_prot_err_fifo, struct cx= l_cper_prot_err_work_data, > static DEFINE_SPINLOCK(cxl_cper_prot_err_work_lock); > struct work_struct *cxl_cper_prot_err_work; > =20 > -static void cxl_cper_post_prot_err(struct cxl_cper_sec_prot_err *prot_er= r, > - int severity, u32 len) > +void cxl_cper_post_prot_err(struct cxl_cper_sec_prot_err *prot_err, > + int severity, u32 len) > { [Severity: High] This is a pre-existing issue, but can this cause an AA deadlock? Looking at cxl_cper_register_prot_err_work() and cxl_cper_unregister_prot_err_work() in drivers/acpi/apei/ghes.c, they acqui= re the cxl_cper_prot_err_work_lock using guard(spinlock) without disabling interrupts since they execute in process context: drivers/acpi/apei/ghes.c:cxl_cper_register_prot_err_work() { ... guard(spinlock)(&cxl_cper_prot_err_work_lock); ... } If an asynchronous hardware error interrupt fires on the same CPU while this lock is held, the interrupt handler calls cxl_cper_post_prot_err() which attempts to acquire the exact same lock using guard(spinlock_irqsave): drivers/acpi/apei/ghes.c:cxl_cper_post_prot_err() { ... guard(spinlock_irqsave)(&cxl_cper_prot_err_work_lock); ... } Would this cause the interrupt handler to spin forever waiting for the lock that the interrupted process context currently holds? --=20 Sashiko AI review =C2=B7 https://sashiko.dev/#/patchset/20260717161647.1493= 259-1-dave.jiang@intel.com?part=3D6