From: Alexandru Gagniuc <mr.nuke.me@gmail.com> To: bhelgaas@google.com Cc: austin_bolen@dell.com, alex_gagniuc@dellteam.com, keith.busch@intel.com, Shyam_Iyer@Dell.com, lukas@wunner.de, okaya@kernel.org, scott.faasse@hpe.com, leo.duran@amd.com, Alexandru Gagniuc <mr.nuke.me@gmail.com>, "Rafael J. Wysocki" <rjw@rjwysocki.net>, Len Brown <lenb@kernel.org>, Russell Currey <ruscur@russell.cc>, Sam Bobroff <sbobroff@linux.ibm.com>, Oliver O'Halloran <oohall@gmail.com>, linux-pci@vger.kernel.org, linux-acpi@vger.kernel.org, linux-kernel@vger.kernel.org, linuxppc-dev@lists.ozlabs.org Subject: [PATCH v2 2/2] PCI/AER: Determine AER ownership based on _OSC instead of HEST Date: Tue, 26 Mar 2019 12:23:42 -0500 [thread overview] Message-ID: <20190326172343.28946-3-mr.nuke.me@gmail.com> (raw) In-Reply-To: <20190326172343.28946-1-mr.nuke.me@gmail.com> HEST is used to describe the meaning of errors received as part of ACPI Platform Error Interfaces (APEI), however the correct way to determine AER ownership is the _OSC method. The ACPI spec allows _OSC and HEST to say things that might not make sense on a specific hardware implementation. For example _OSC can say "OS has control" on the root bus, while HEST says "This PCIe device is firmware-first". This is fine when the platform can differentiate error sources at the root complex. On x86 platforms, AER errors can be routed to either to an MSI vector (native AER), or to SMM (firmware-first). As this is done at the root complex level, the example above would not make sense. Notice that in neither case is it correct to go to HEST to determine AER ownership. This is the conclusion of a six-months discussion in the ACPI Software Working Group. pci_dev->__aer_firmware_first is used to prevent modification of AER registers when firmware owns AER. This is synonymous with the AER ownership negotiated during _OSC. Thus _OSC is the correct way to use to set this flag, not HEST. Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com> --- drivers/pci/pcie/aer.c | 57 ++---------------------------------------- 1 file changed, 2 insertions(+), 55 deletions(-) diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c index d029979e61f6..08cee30b3ef3 100644 --- a/drivers/pci/pcie/aer.c +++ b/drivers/pci/pcie/aer.c @@ -237,66 +237,13 @@ static inline bool hest_match_type(struct acpi_hest_header *hest_hdr, return false; } -struct aer_hest_parse_info { - struct pci_dev *pci_dev; - int firmware_first; -}; - -static int hest_source_is_pcie_aer(struct acpi_hest_header *hest_hdr) -{ - if (hest_hdr->type == ACPI_HEST_TYPE_AER_ROOT_PORT || - hest_hdr->type == ACPI_HEST_TYPE_AER_ENDPOINT || - hest_hdr->type == ACPI_HEST_TYPE_AER_BRIDGE) - return 1; - return 0; -} - -static int aer_hest_parse(struct acpi_hest_header *hest_hdr, void *data) -{ - struct aer_hest_parse_info *info = data; - struct acpi_hest_aer_common *p; - int ff; - - if (!hest_source_is_pcie_aer(hest_hdr)) - return 0; - - p = (struct acpi_hest_aer_common *)(hest_hdr + 1); - ff = !!(p->flags & ACPI_HEST_FIRMWARE_FIRST); - - /* - * If no specific device is supplied, determine whether - * FIRMWARE_FIRST is set for *any* PCIe device. - */ - if (!info->pci_dev) { - info->firmware_first |= ff; - return 0; - } - /* Otherwise, check the specific device */ - if (p->flags & ACPI_HEST_GLOBAL) { - if (hest_match_type(hest_hdr, info->pci_dev)) - info->firmware_first = ff; - } else - if (hest_match_pci(p, info->pci_dev)) - info->firmware_first = ff; - - return 0; -} static void aer_set_firmware_first(struct pci_dev *pci_dev) { - int rc; - struct aer_hest_parse_info info = { - .pci_dev = pci_dev, - .firmware_first = 0, - }; + struct pci_host_bridge *host = pci_find_host_bridge(pci_dev->bus); - rc = apei_hest_parse(aer_hest_parse, &info); - - if (rc) - pci_dev->__aer_firmware_first = 0; - else - pci_dev->__aer_firmware_first = info.firmware_first; + pci_dev->__aer_firmware_first = !host->native_aer; pci_dev->__aer_firmware_first_valid = 1; } -- 2.19.2
WARNING: multiple messages have this Message-ID (diff)
From: Alexandru Gagniuc <mr.nuke.me@gmail.com> To: bhelgaas@google.com Cc: alex_gagniuc@dellteam.com, Sam Bobroff <sbobroff@linux.ibm.com>, "Rafael J. Wysocki" <rjw@rjwysocki.net>, Shyam_Iyer@Dell.com, okaya@kernel.org, linux-kernel@vger.kernel.org, linux-pci@vger.kernel.org, keith.busch@intel.com, linux-acpi@vger.kernel.org, lukas@wunner.de, leo.duran@amd.com, Alexandru Gagniuc <mr.nuke.me@gmail.com>, scott.faasse@hpe.com, Oliver O'Halloran <oohall@gmail.com>, austin_bolen@dell.com, linuxppc-dev@lists.ozlabs.org, Len Brown <lenb@kernel.org> Subject: [PATCH v2 2/2] PCI/AER: Determine AER ownership based on _OSC instead of HEST Date: Tue, 26 Mar 2019 12:23:42 -0500 [thread overview] Message-ID: <20190326172343.28946-3-mr.nuke.me@gmail.com> (raw) In-Reply-To: <20190326172343.28946-1-mr.nuke.me@gmail.com> HEST is used to describe the meaning of errors received as part of ACPI Platform Error Interfaces (APEI), however the correct way to determine AER ownership is the _OSC method. The ACPI spec allows _OSC and HEST to say things that might not make sense on a specific hardware implementation. For example _OSC can say "OS has control" on the root bus, while HEST says "This PCIe device is firmware-first". This is fine when the platform can differentiate error sources at the root complex. On x86 platforms, AER errors can be routed to either to an MSI vector (native AER), or to SMM (firmware-first). As this is done at the root complex level, the example above would not make sense. Notice that in neither case is it correct to go to HEST to determine AER ownership. This is the conclusion of a six-months discussion in the ACPI Software Working Group. pci_dev->__aer_firmware_first is used to prevent modification of AER registers when firmware owns AER. This is synonymous with the AER ownership negotiated during _OSC. Thus _OSC is the correct way to use to set this flag, not HEST. Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com> --- drivers/pci/pcie/aer.c | 57 ++---------------------------------------- 1 file changed, 2 insertions(+), 55 deletions(-) diff --git a/drivers/pci/pcie/aer.c b/drivers/pci/pcie/aer.c index d029979e61f6..08cee30b3ef3 100644 --- a/drivers/pci/pcie/aer.c +++ b/drivers/pci/pcie/aer.c @@ -237,66 +237,13 @@ static inline bool hest_match_type(struct acpi_hest_header *hest_hdr, return false; } -struct aer_hest_parse_info { - struct pci_dev *pci_dev; - int firmware_first; -}; - -static int hest_source_is_pcie_aer(struct acpi_hest_header *hest_hdr) -{ - if (hest_hdr->type == ACPI_HEST_TYPE_AER_ROOT_PORT || - hest_hdr->type == ACPI_HEST_TYPE_AER_ENDPOINT || - hest_hdr->type == ACPI_HEST_TYPE_AER_BRIDGE) - return 1; - return 0; -} - -static int aer_hest_parse(struct acpi_hest_header *hest_hdr, void *data) -{ - struct aer_hest_parse_info *info = data; - struct acpi_hest_aer_common *p; - int ff; - - if (!hest_source_is_pcie_aer(hest_hdr)) - return 0; - - p = (struct acpi_hest_aer_common *)(hest_hdr + 1); - ff = !!(p->flags & ACPI_HEST_FIRMWARE_FIRST); - - /* - * If no specific device is supplied, determine whether - * FIRMWARE_FIRST is set for *any* PCIe device. - */ - if (!info->pci_dev) { - info->firmware_first |= ff; - return 0; - } - /* Otherwise, check the specific device */ - if (p->flags & ACPI_HEST_GLOBAL) { - if (hest_match_type(hest_hdr, info->pci_dev)) - info->firmware_first = ff; - } else - if (hest_match_pci(p, info->pci_dev)) - info->firmware_first = ff; - - return 0; -} static void aer_set_firmware_first(struct pci_dev *pci_dev) { - int rc; - struct aer_hest_parse_info info = { - .pci_dev = pci_dev, - .firmware_first = 0, - }; + struct pci_host_bridge *host = pci_find_host_bridge(pci_dev->bus); - rc = apei_hest_parse(aer_hest_parse, &info); - - if (rc) - pci_dev->__aer_firmware_first = 0; - else - pci_dev->__aer_firmware_first = info.firmware_first; + pci_dev->__aer_firmware_first = !host->native_aer; pci_dev->__aer_firmware_first_valid = 1; } -- 2.19.2
next prev parent reply other threads:[~2019-03-26 17:23 UTC|newest] Thread overview: 6+ messages / expand[flat|nested] mbox.gz Atom feed top 2019-03-26 17:23 [PATCH v2 0/2] PCI/AER: Consistently use _OSC to determine who owns AER Alexandru Gagniuc 2019-03-26 17:23 ` Alexandru Gagniuc 2019-03-26 17:23 ` [PATCH v2 1/2] PCI/AER: Do not use APEI/HEST to disable AER services globally Alexandru Gagniuc 2019-03-26 17:23 ` Alexandru Gagniuc 2019-03-26 17:23 ` Alexandru Gagniuc [this message] 2019-03-26 17:23 ` [PATCH v2 2/2] PCI/AER: Determine AER ownership based on _OSC instead of HEST Alexandru Gagniuc
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=20190326172343.28946-3-mr.nuke.me@gmail.com \ --to=mr.nuke.me@gmail.com \ --cc=Shyam_Iyer@Dell.com \ --cc=alex_gagniuc@dellteam.com \ --cc=austin_bolen@dell.com \ --cc=bhelgaas@google.com \ --cc=keith.busch@intel.com \ --cc=lenb@kernel.org \ --cc=leo.duran@amd.com \ --cc=linux-acpi@vger.kernel.org \ --cc=linux-kernel@vger.kernel.org \ --cc=linux-pci@vger.kernel.org \ --cc=linuxppc-dev@lists.ozlabs.org \ --cc=lukas@wunner.de \ --cc=okaya@kernel.org \ --cc=oohall@gmail.com \ --cc=rjw@rjwysocki.net \ --cc=ruscur@russell.cc \ --cc=sbobroff@linux.ibm.com \ --cc=scott.faasse@hpe.com \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes, see mirroring instructions on how to clone and mirror all data and code used by this external index.