cxl: Remove ifdef blocks of CONFIG_PCIEAER_CXL from core/pci.c
Dave Jiang
maintainerPATCH v2 1/1 cxl: Remove ifdef blocks of CONFIG_PCIEAER_CXL from core/pci.c
Remove from core/pci_aer.c and move the CONFIG_PCIEAER_CXL ifdef to cxlpci.h to deal with the exported symbols as needed. There is enough AER handling code (and more to come) to move the AER code to its own C file.
---
v2:
- Address 0-day reprot of cxl_handle_rdport_errors symbol undefined
--- drivers/cxl/core/Makefile | 1 + drivers/cxl/core/core.h | 9 ++ drivers/cxl/core/pci.c | 171 +------------------------------------ drivers/cxl/core/pci_aer.c | 168 ++++++++++++++++++++++++++++++++++++ drivers/cxl/cxl.h | 8 -- drivers/cxl/cxlpci.h | 8 ++ tools/testing/cxl/Kbuild | 1 + 7 files changed, 189 insertions(+), 177 deletions(-) create mode 100644 drivers/cxl/core/pci_aer.c
drivers/cxl/core/Makefile +1 -0
@@ -21,3 +21,4 @@ cxl_core-$(CONFIG_CXL_REGION) += region.o cxl_core-$(CONFIG_CXL_MCE) += mce.o cxl_core-$(CONFIG_CXL_FEATURES) += features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += pci_aer.o
drivers/cxl/core/core.h +9 -0
@@ -136,4 +136,13 @@ int cxl_set_feature(struct cxl_mailbox *cxl_mbox, const uuid_t *feat_uuid, u16 *return_code);#endif +#ifdef CONFIG_PCIEAER_CXL +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds); +#else +static inline void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) {} +#endif + +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +#endif /* __CXL_CORE_H__ */
drivers/cxl/core/pci.c +2 -169
189 lines not on this page · raw message
drivers/cxl/core/pci_aer.c new file +168 -0
168 lines not on this page · raw message
drivers/cxl/cxl.h +0 -8
@@ -759,14 +759,6 @@ struct cxl_dport *devm_cxl_add_rch_dport(struct cxl_port *port, struct device *dport_dev, int port_id, resource_size_t rcrb); -#ifdef CONFIG_PCIEAER_CXL -void cxl_setup_parent_dport (struct device *host,struct cxl_dport *dport); -void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host); -#else -static inline void cxl_dport_init_ras_reporting (struct cxl_dport *dport, -struct device *host) { } -#endif -struct cxl_decoder *to_cxl_decoder (struct device *dev);struct cxl_root_decoder *to_cxl_root_decoder (struct device *dev);struct cxl_switch_decoder *to_cxl_switch_decoder (struct device *dev);
drivers/cxl/cxlpci.h +8 -0
@@ -135,4 +135,12 @@ void read_cdat_data(struct cxl_port *port);void cxl_cor_error_detected (struct pci_dev *pdev);pci_ers_result_t cxl_error_detected (struct pci_dev *pdev,pci_channel_state_t state); + +#ifdef CONFIG_PCIEAER_CXL +void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host); +#else +static inline void cxl_dport_init_ras_reporting (struct cxl_dport *dport, +struct device *host) { } +#endif +#endif /* __CXL_PCI_H__ */
tools/testing/cxl/Kbuild +1 -0
@@ -68,6 +68,7 @@ cxl_core-$(CONFIG_CXL_REGION) += $(CXL_CORE_SRC)/region.o cxl_core-$(CONFIG_CXL_MCE) += $(CXL_CORE_SRC)/mce.o cxl_core-$(CONFIG_CXL_FEATURES) += $(CXL_CORE_SRC)/features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += $(CXL_CORE_SRC)/edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += $(CXL_CORE_SRC)/pci_aer.o cxl_core-y += config_check.o cxl_core-y += cxl_core_test.o cxl_core-y += cxl_core_exports.o
base-commit: 347e9f5043c89695b01e66b3ed111755afcf1911
Signature · 1 line
2.50.1
Quoted 8 lines · On Mon, 21 Jul 2025 10:04:15 -0700 Dave Jiang <[email protected]> wrote:
Remove from
core/pci_aer.cand move theCONFIG_PCIEAER_CXLifdef tocxlpci.hto deal with the exported symbols as needed. There is enough AER handling code (and more to come) to move the AER code to its own C file.Cc Robert Richter <[email protected]>Cc Terry Bowman <[email protected]>Signed-off-by Dave Jiang <[email protected]>
Quoted 5 lines · On Mon, 21 Jul 2025 10:04:15 -0700 Dave Jiang <[email protected]> wrote:
Remove from
core/pci_aer.cand move theCONFIG_PCIEAER_CXLifdef tocxlpci.hto deal with the exported symbols as needed. There is enough AER handling code (and more to come) to move the AER code to its own C file.
Quoted 7 lines
Cc Robert Richter <[email protected]>Cc Terry Bowman <[email protected]>Signed-off-by Dave Jiang <[email protected]>---
v2:
- Address 0-day reprot ofcxl_handle_rdport_errorssymbol undefined
---
Hello Dave, thank you for the patch! This makes a lot of sense to me and it all looks good : -)
Feel free to add a
Have a great day!
Joshua
drivers/cxl/core/pci_aer.c
Quoted 250 lines · Dave Jiang
Remove from
core/pci_aer.cand move theCONFIG_PCIEAER_CXLifdef tocxlpci.hto deal with the exported symbols as needed. There is enough AER handling code (and more to come) to move the AER code to its own C file.Cc Robert Richter <[email protected]>Cc Terry Bowman <[email protected]>Signed-off-by Dave Jiang <[email protected]>---
v2:
- Address 0-day reprot ofcxl_handle_rdport_errorssymbol undefined--- drivers/cxl/core/Makefile | 1 + drivers/cxl/core/core.h | 9 ++ drivers/cxl/core/pci.c | 171 +------------------------------------ drivers/cxl/core/pci_aer.c | 168 ++++++++++++++++++++++++++++++++++++ drivers/cxl/cxl.h | 8 -- drivers/cxl/cxlpci.h | 8 ++ tools/testing/cxl/Kbuild | 1 + 7 files changed, 189 insertions(+), 177 deletions(-) create mode 100644 drivers/cxl/core/pci_aer.cdrivers/cxl/core/Makefile +1
-0@@ -21,3 +21,4 @@ cxl_core-$(CONFIG_CXL_REGION) += region.o cxl_core-$(CONFIG_CXL_MCE) += mce.o cxl_core-$(CONFIG_CXL_FEATURES) += features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += pci_aer.odrivers/cxl/core/core.h +9
-0@@ -136,4 +136,13 @@ int cxl_set_feature(struct cxl_mailbox *cxl_mbox, const uuid_t *feat_uuid, u16 *return_code);#endif +#ifdef CONFIG_PCIEAER_CXL +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds); +#else +static inline void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) {} +#endif + +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +#endif /* __CXL_CORE_H__ */ drivers/cxl/core/pci.c +2
-169@@ -664,8 +664,7 @@ void read_cdat_data(struct cxl_port *port) }EXPORT_SYMBOL_NS_GPL (read_cdat_data,"CXL" ); -static void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds, -void __iomem *ras_base) +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base) {void __iomem *addr;u32 status;@@ -707,8 +706,7 @@ static void header_log_copy(void __iomem *ras_base, u32 *log)* Log the state of the RAS status registers and prepare them to log the * next error status. Return 1 if reset needed. */ -static bool __cxl_handle_ras (struct cxl_dev_state *cxlds, -void __iomem *ras_base) +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base) {u32 hl[CXL_HEADERLOG_SIZE_U32 ];void __iomem *addr;@@ -746,171 +744,6 @@ static bool cxl_handle_endpoint_ras(struct cxl_dev_state *cxlds)return __cxl_handle_ras (cxlds, cxlds->regs.ras); } -#ifdef CONFIG_PCIEAER_CXL - -static void cxl_dport_map_rch_aer (struct cxl_dport *dport) -{ -resource_size_t aer_phys; -struct device *host; -u16 aer_cap; - - aer_cap =cxl_rcrb_to_aer (dport->dport_dev, dport->rcrb.base); -if (aer_cap) { - host = dport->reg_map.host; - aer_phys = aer_cap + dport->rcrb.base; - dport->regs.dport_aer =devm_cxl_iomap_block (host, aer_phys, -sizeof (struct aer_capability_regs )); - } -} - -static void cxl_dport_map_ras (struct cxl_dport *dport) -{ -struct cxl_register_map *map = &dport->reg_map; -struct device *dev = dport->dport_dev; - -if (!map->component_map.ras.valid) -dev_dbg (dev,"RAS registers not found\n" ); -else if (cxl_map_component_regs (map, &dport->regs.component, -BIT (CXL_CM_CAP_CAP_ID_RAS ))) -dev_dbg (dev,"Failed to map RAS capability.\n" ); -} - -static void cxl_disable_rch_root_ints (struct cxl_dport *dport) -{ -void __iomem *aer_base = dport->regs.dport_aer; -u32 aer_cmd_mask, aer_cmd; - -if (!aer_base) -return ; - -/* -* Disable RCH root port command interrupts. -* CXL 3.0 12.2.1.1 - RCH Downstream Port-detected Errors -* -* This sequence may not be necessary. CXL spec states disabling -* the root cmd register's interrupts is required. But, PCI spec -* shows these are disabled by default on reset. -*/ - aer_cmd_mask = (PCI_ERR_ROOT_CMD_COR_EN | -PCI_ERR_ROOT_CMD_NONFATAL_EN | -PCI_ERR_ROOT_CMD_FATAL_EN ); - aer_cmd =readl (aer_base +PCI_ERR_ROOT_COMMAND ); - aer_cmd &= ~aer_cmd_mask; -writel (aer_cmd, aer_base +PCI_ERR_ROOT_COMMAND ); -} - -/** -* cxl_dport_init_ras_reporting - Setup CXL RAS report on this dport -* @dport: the cxl_dport that needs to be initialized -* @host: host device for devm operations -*/ -void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host) -{ - dport->reg_map.host = host; -cxl_dport_map_ras (dport); - -if (dport->rch) { -struct pci_host_bridge *host_bridge =to_pci_host_bridge (dport->dport_dev); - -if (!host_bridge->native_aer) -return ; - -cxl_dport_map_rch_aer (dport); -cxl_disable_rch_root_ints (dport); - } -} -EXPORT_SYMBOL_NS_GPL (cxl_dport_init_ras_reporting,"CXL" ); - -static void cxl_handle_rdport_cor_ras (struct cxl_dev_state *cxlds, -struct cxl_dport *dport) -{ -return __cxl_handle_cor_ras (cxlds, dport->regs.ras); -} - -static bool cxl_handle_rdport_ras (struct cxl_dev_state *cxlds, -struct cxl_dport *dport) -{ -return __cxl_handle_ras (cxlds, dport->regs.ras); -} - -/* -* Copy the AER capability registers using 32 bit read accesses. -* This is necessary because RCRB AER capability is MMIO mapped. Clear the -* status after copying. -* -* @aer_base: base address of AER capability block in RCRB -* @aer_regs: destination for copying AER capability -*/ -static bool cxl_rch_get_aer_info (void __iomem *aer_base, -struct aer_capability_regs *aer_regs) -{ -int read_cnt =sizeof (struct aer_capability_regs ) /sizeof (u32); -u32 *aer_regs_buf = (u32 *)aer_regs; -int n; - -if (!aer_base) -return false; - -/* Use readl() to guarantee 32-bit accesses */ -for (n =0 ; n < read_cnt; n++) - aer_regs_buf[n] =readl (aer_base + n *sizeof (u32)); - -writel (aer_regs->uncor_status, aer_base +PCI_ERR_UNCOR_STATUS ); -writel (aer_regs->cor_status, aer_base +PCI_ERR_COR_STATUS ); - -return true; -} - -/* Get AER severity. Return false if there is no error. */ -static bool cxl_rch_get_aer_severity (struct aer_capability_regs *aer_regs, -int *severity) -{ -if (aer_regs->uncor_status & ~aer_regs->uncor_mask) { -if (aer_regs->uncor_status &PCI_ERR_ROOT_FATAL_RCV ) - *severity =AER_FATAL ; -else - *severity =AER_NONFATAL ; -return true; - } - -if (aer_regs->cor_status & ~aer_regs->cor_mask) { - *severity =AER_CORRECTABLE ; -return true; - } - -return false; -} - -static void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) -{ -struct pci_dev *pdev =to_pci_dev (cxlds->dev); -struct aer_capability_regs aer_regs; -struct cxl_dport *dport; -int severity; - - struct cxl_port *port __free(put_cxl_port) = -cxl_pci_find_port (pdev, &dport); -if (!port) -return ; - -if (!cxl_rch_get_aer_info (dport->regs.dport_aer, &aer_regs)) -return ; - -if (!cxl_rch_get_aer_severity (&aer_regs, &severity)) -return ; - -pci_print_aer (pdev, severity, &aer_regs); - -if (severity ==AER_CORRECTABLE ) -cxl_handle_rdport_cor_ras (cxlds, dport); -else -cxl_handle_rdport_ras (cxlds, dport); -} - -#else -static void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) { } -#endif -void cxl_cor_error_detected (struct pci_dev *pdev) {struct cxl_dev_state *cxlds =pci_get_drvdata (pdev);drivers/cxl/core/pci_aer.c new file +0
-0
Hi Dave,
While adding this into my series and looking closer, I believe that pci_aer.c filename does not accurately reflect its content. The only AER-related code pertains to CX1.1, which is now a corner case. The AER driver, not CXL driver, logs AER errors for all other scenarios. There is no AER processing in the CXL driver except for CXL1.1 dports.
Most of the code in this file is related to or will be directed towards CXL RAS. It's important to note that AER and RAS are distinct and separate registers, with CXL RAS only utilizing the AER interrupt notification and UIE/CIE.
I am concerned we are conflating AER and RAS, which may be misleading for others. We have other instances of this confusion, such as in the cxl_aer_correctable_error() and cxl_aer_uncorrectable_error() functions, which log RAS instead of AER.
My series aims to clarify the distinction between AER and CXL errors, categorizing CXL errors as a separate error class. At the least, can we add a documentation block at the top to indicate this source file provides functionality for CXL RAS?
-Terry
Quoted 219 lines
drivers/cxl/core/pci_aer.c new file +168
-0@@ -0,0 +1,168 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* Copyright(c) 2025 AMD Corporation. All rights reserved. */ +/* Copyright(c) 2025 Intel Corporation. All rights reserved. */ +#include <linux/pci.h> +#include <linux/aer.h> +#include <cxlpci.h> +#include <cxlmem.h> +#include <cxl.h> +#include "core.h" + +static void cxl_dport_map_rch_aer (struct cxl_dport *dport) +{ +resource_size_t aer_phys; +struct device *host; +u16 aer_cap; + + aer_cap =cxl_rcrb_to_aer (dport->dport_dev, dport->rcrb.base); +if (aer_cap) { + host = dport->reg_map.host; + aer_phys = aer_cap + dport->rcrb.base; + dport->regs.dport_aer =devm_cxl_iomap_block (host, aer_phys, +sizeof (struct aer_capability_regs )); + } +} + +static void cxl_dport_map_ras (struct cxl_dport *dport) +{ +struct cxl_register_map *map = &dport->reg_map; +struct device *dev = dport->dport_dev; + +if (!map->component_map.ras.valid) +dev_dbg (dev,"RAS registers not found\n" ); +else if (cxl_map_component_regs (map, &dport->regs.component, +BIT (CXL_CM_CAP_CAP_ID_RAS ))) +dev_dbg (dev,"Failed to map RAS capability.\n" ); +} + +static void cxl_disable_rch_root_ints (struct cxl_dport *dport) +{ +void __iomem *aer_base = dport->regs.dport_aer; +u32 aer_cmd_mask, aer_cmd; + +if (!aer_base) +return ; + +/* +* Disable RCH root port command interrupts. +* CXL 3.0 12.2.1.1 - RCH Downstream Port-detected Errors +* +* This sequence may not be necessary. CXL spec states disabling +* the root cmd register's interrupts is required. But, PCI spec +* shows these are disabled by default on reset. +*/ + aer_cmd_mask = (PCI_ERR_ROOT_CMD_COR_EN | +PCI_ERR_ROOT_CMD_NONFATAL_EN | +PCI_ERR_ROOT_CMD_FATAL_EN ); + aer_cmd =readl (aer_base +PCI_ERR_ROOT_COMMAND ); + aer_cmd &= ~aer_cmd_mask; +writel (aer_cmd, aer_base +PCI_ERR_ROOT_COMMAND ); +} + +/** +* cxl_dport_init_ras_reporting - Setup CXL RAS report on this dport +* @dport: the cxl_dport that needs to be initialized +* @host: host device for devm operations +*/ +void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host) +{ + dport->reg_map.host = host; +cxl_dport_map_ras (dport); + +if (dport->rch) { +struct pci_host_bridge *host_bridge =to_pci_host_bridge (dport->dport_dev); + +if (!host_bridge->native_aer) +return ; + +cxl_dport_map_rch_aer (dport); +cxl_disable_rch_root_ints (dport); + } +} +EXPORT_SYMBOL_NS_GPL (cxl_dport_init_ras_reporting,"CXL" ); + +static void cxl_handle_rdport_cor_ras (struct cxl_dev_state *cxlds, +struct cxl_dport *dport) +{ +return __cxl_handle_cor_ras (cxlds, dport->regs.ras); +} + +static bool cxl_handle_rdport_ras (struct cxl_dev_state *cxlds, +struct cxl_dport *dport) +{ +return __cxl_handle_ras (cxlds, dport->regs.ras); +} + +/* +* Copy the AER capability registers using 32 bit read accesses. +* This is necessary because RCRB AER capability is MMIO mapped. Clear the +* status after copying. +* +* @aer_base: base address of AER capability block in RCRB +* @aer_regs: destination for copying AER capability +*/ +static bool cxl_rch_get_aer_info (void __iomem *aer_base, +struct aer_capability_regs *aer_regs) +{ +int read_cnt =sizeof (struct aer_capability_regs ) /sizeof (u32); +u32 *aer_regs_buf = (u32 *)aer_regs; +int n; + +if (!aer_base) +return false; + +/* Use readl() to guarantee 32-bit accesses */ +for (n =0 ; n < read_cnt; n++) + aer_regs_buf[n] =readl (aer_base + n *sizeof (u32)); + +writel (aer_regs->uncor_status, aer_base +PCI_ERR_UNCOR_STATUS ); +writel (aer_regs->cor_status, aer_base +PCI_ERR_COR_STATUS ); + +return true; +} + +/* Get AER severity. Return false if there is no error. */ +static bool cxl_rch_get_aer_severity (struct aer_capability_regs *aer_regs, +int *severity) +{ +if (aer_regs->uncor_status & ~aer_regs->uncor_mask) { +if (aer_regs->uncor_status &PCI_ERR_ROOT_FATAL_RCV ) + *severity =AER_FATAL ; +else + *severity =AER_NONFATAL ; +return true; + } + +if (aer_regs->cor_status & ~aer_regs->cor_mask) { + *severity =AER_CORRECTABLE ; +return true; + } + +return false; +} + +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) +{ +struct pci_dev *pdev =to_pci_dev (cxlds->dev); +struct aer_capability_regs aer_regs; +struct cxl_dport *dport; +int severity; + + struct cxl_port *port __free(put_cxl_port) = +cxl_pci_find_port (pdev, &dport); +if (!port) +return ; + +if (!cxl_rch_get_aer_info (dport->regs.dport_aer, &aer_regs)) +return ; + +if (!cxl_rch_get_aer_severity (&aer_regs, &severity)) +return ; + +pci_print_aer (pdev, severity, &aer_regs); + +if (severity ==AER_CORRECTABLE ) +cxl_handle_rdport_cor_ras (cxlds, dport); +else +cxl_handle_rdport_ras (cxlds, dport); +}drivers/cxl/cxl.h +0
-8@@ -759,14 +759,6 @@ struct cxl_dport *devm_cxl_add_rch_dport(struct cxl_port *port, struct device *dport_dev, int port_id, resource_size_t rcrb); -#ifdef CONFIG_PCIEAER_CXL -void cxl_setup_parent_dport (struct device *host,struct cxl_dport *dport); -void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host); -#else -static inline void cxl_dport_init_ras_reporting (struct cxl_dport *dport, -struct device *host) { } -#endif -struct cxl_decoder *to_cxl_decoder (struct device *dev);struct cxl_root_decoder *to_cxl_root_decoder (struct device *dev);struct cxl_switch_decoder *to_cxl_switch_decoder (struct device *dev);drivers/cxl/cxlpci.h +8
-0@@ -135,4 +135,12 @@ void read_cdat_data(struct cxl_port *port);void cxl_cor_error_detected (struct pci_dev *pdev);pci_ers_result_t cxl_error_detected (struct pci_dev *pdev,pci_channel_state_t state); + +#ifdef CONFIG_PCIEAER_CXL +void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host); +#else +static inline void cxl_dport_init_ras_reporting (struct cxl_dport *dport, +struct device *host) { } +#endif +#endif /* __CXL_PCI_H__ */ tools/testing/cxl/Kbuild +1
-0@@ -68,6 +68,7 @@ cxl_core-$(CONFIG_CXL_REGION) += $(CXL_CORE_SRC)/region.o cxl_core-$(CONFIG_CXL_MCE) += $(CXL_CORE_SRC)/mce.o cxl_core-$(CONFIG_CXL_FEATURES) += $(CXL_CORE_SRC)/features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += $(CXL_CORE_SRC)/edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += $(CXL_CORE_SRC)/pci_aer.o cxl_core-y += config_check.o cxl_core-y += cxl_core_test.o cxl_core-y += cxl_core_exports.obase-commit:
347e9f5043c89695b01e66b3ed111755afcf1911
Dave Jiang
maintainerQuoted 263 lines · Terry
drivers/cxl/core/pci_aer.c
Quoted 250 lines · Dave Jiang
Remove from
core/pci_aer.cand move theCONFIG_PCIEAER_CXLifdef tocxlpci.hto deal with the exported symbols as needed. There is enough AER handling code (and more to come) to move the AER code to its own C file.Cc Robert Richter <[email protected]>Cc Terry Bowman <[email protected]>Signed-off-by Dave Jiang <[email protected]>---
v2:
- Address 0-day reprot ofcxl_handle_rdport_errorssymbol undefined--- drivers/cxl/core/Makefile | 1 + drivers/cxl/core/core.h | 9 ++ drivers/cxl/core/pci.c | 171 +------------------------------------ drivers/cxl/core/pci_aer.c | 168 ++++++++++++++++++++++++++++++++++++ drivers/cxl/cxl.h | 8 -- drivers/cxl/cxlpci.h | 8 ++ tools/testing/cxl/Kbuild | 1 + 7 files changed, 189 insertions(+), 177 deletions(-) create mode 100644 drivers/cxl/core/pci_aer.cdrivers/cxl/core/Makefile +1
-0@@ -21,3 +21,4 @@ cxl_core-$(CONFIG_CXL_REGION) += region.o cxl_core-$(CONFIG_CXL_MCE) += mce.o cxl_core-$(CONFIG_CXL_FEATURES) += features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += pci_aer.odrivers/cxl/core/core.h +9
-0@@ -136,4 +136,13 @@ int cxl_set_feature(struct cxl_mailbox *cxl_mbox, const uuid_t *feat_uuid, u16 *return_code);#endif +#ifdef CONFIG_PCIEAER_CXL +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds); +#else +static inline void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) {} +#endif + +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +#endif /* __CXL_CORE_H__ */ drivers/cxl/core/pci.c +2
-169@@ -664,8 +664,7 @@ void read_cdat_data(struct cxl_port *port) }EXPORT_SYMBOL_NS_GPL (read_cdat_data,"CXL" ); -static void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds, -void __iomem *ras_base) +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base) {void __iomem *addr;u32 status;@@ -707,8 +706,7 @@ static void header_log_copy(void __iomem *ras_base, u32 *log)* Log the state of the RAS status registers and prepare them to log the * next error status. Return 1 if reset needed. */ -static bool __cxl_handle_ras (struct cxl_dev_state *cxlds, -void __iomem *ras_base) +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base) {u32 hl[CXL_HEADERLOG_SIZE_U32 ];void __iomem *addr;@@ -746,171 +744,6 @@ static bool cxl_handle_endpoint_ras(struct cxl_dev_state *cxlds)return __cxl_handle_ras (cxlds, cxlds->regs.ras); } -#ifdef CONFIG_PCIEAER_CXL - -static void cxl_dport_map_rch_aer (struct cxl_dport *dport) -{ -resource_size_t aer_phys; -struct device *host; -u16 aer_cap; - - aer_cap =cxl_rcrb_to_aer (dport->dport_dev, dport->rcrb.base); -if (aer_cap) { - host = dport->reg_map.host; - aer_phys = aer_cap + dport->rcrb.base; - dport->regs.dport_aer =devm_cxl_iomap_block (host, aer_phys, -sizeof (struct aer_capability_regs )); - } -} - -static void cxl_dport_map_ras (struct cxl_dport *dport) -{ -struct cxl_register_map *map = &dport->reg_map; -struct device *dev = dport->dport_dev; - -if (!map->component_map.ras.valid) -dev_dbg (dev,"RAS registers not found\n" ); -else if (cxl_map_component_regs (map, &dport->regs.component, -BIT (CXL_CM_CAP_CAP_ID_RAS ))) -dev_dbg (dev,"Failed to map RAS capability.\n" ); -} - -static void cxl_disable_rch_root_ints (struct cxl_dport *dport) -{ -void __iomem *aer_base = dport->regs.dport_aer; -u32 aer_cmd_mask, aer_cmd; - -if (!aer_base) -return ; - -/* -* Disable RCH root port command interrupts. -* CXL 3.0 12.2.1.1 - RCH Downstream Port-detected Errors -* -* This sequence may not be necessary. CXL spec states disabling -* the root cmd register's interrupts is required. But, PCI spec -* shows these are disabled by default on reset. -*/ - aer_cmd_mask = (PCI_ERR_ROOT_CMD_COR_EN | -PCI_ERR_ROOT_CMD_NONFATAL_EN | -PCI_ERR_ROOT_CMD_FATAL_EN ); - aer_cmd =readl (aer_base +PCI_ERR_ROOT_COMMAND ); - aer_cmd &= ~aer_cmd_mask; -writel (aer_cmd, aer_base +PCI_ERR_ROOT_COMMAND ); -} - -/** -* cxl_dport_init_ras_reporting - Setup CXL RAS report on this dport -* @dport: the cxl_dport that needs to be initialized -* @host: host device for devm operations -*/ -void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host) -{ - dport->reg_map.host = host; -cxl_dport_map_ras (dport); - -if (dport->rch) { -struct pci_host_bridge *host_bridge =to_pci_host_bridge (dport->dport_dev); - -if (!host_bridge->native_aer) -return ; - -cxl_dport_map_rch_aer (dport); -cxl_disable_rch_root_ints (dport); - } -} -EXPORT_SYMBOL_NS_GPL (cxl_dport_init_ras_reporting,"CXL" ); - -static void cxl_handle_rdport_cor_ras (struct cxl_dev_state *cxlds, -struct cxl_dport *dport) -{ -return __cxl_handle_cor_ras (cxlds, dport->regs.ras); -} - -static bool cxl_handle_rdport_ras (struct cxl_dev_state *cxlds, -struct cxl_dport *dport) -{ -return __cxl_handle_ras (cxlds, dport->regs.ras); -} - -/* -* Copy the AER capability registers using 32 bit read accesses. -* This is necessary because RCRB AER capability is MMIO mapped. Clear the -* status after copying. -* -* @aer_base: base address of AER capability block in RCRB -* @aer_regs: destination for copying AER capability -*/ -static bool cxl_rch_get_aer_info (void __iomem *aer_base, -struct aer_capability_regs *aer_regs) -{ -int read_cnt =sizeof (struct aer_capability_regs ) /sizeof (u32); -u32 *aer_regs_buf = (u32 *)aer_regs; -int n; - -if (!aer_base) -return false; - -/* Use readl() to guarantee 32-bit accesses */ -for (n =0 ; n < read_cnt; n++) - aer_regs_buf[n] =readl (aer_base + n *sizeof (u32)); - -writel (aer_regs->uncor_status, aer_base +PCI_ERR_UNCOR_STATUS ); -writel (aer_regs->cor_status, aer_base +PCI_ERR_COR_STATUS ); - -return true; -} - -/* Get AER severity. Return false if there is no error. */ -static bool cxl_rch_get_aer_severity (struct aer_capability_regs *aer_regs, -int *severity) -{ -if (aer_regs->uncor_status & ~aer_regs->uncor_mask) { -if (aer_regs->uncor_status &PCI_ERR_ROOT_FATAL_RCV ) - *severity =AER_FATAL ; -else - *severity =AER_NONFATAL ; -return true; - } - -if (aer_regs->cor_status & ~aer_regs->cor_mask) { - *severity =AER_CORRECTABLE ; -return true; - } - -return false; -} - -static void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) -{ -struct pci_dev *pdev =to_pci_dev (cxlds->dev); -struct aer_capability_regs aer_regs; -struct cxl_dport *dport; -int severity; - - struct cxl_port *port __free(put_cxl_port) = -cxl_pci_find_port (pdev, &dport); -if (!port) -return ; - -if (!cxl_rch_get_aer_info (dport->regs.dport_aer, &aer_regs)) -return ; - -if (!cxl_rch_get_aer_severity (&aer_regs, &severity)) -return ; - -pci_print_aer (pdev, severity, &aer_regs); - -if (severity ==AER_CORRECTABLE ) -cxl_handle_rdport_cor_ras (cxlds, dport); -else -cxl_handle_rdport_ras (cxlds, dport); -} - -#else -static void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) { } -#endif -void cxl_cor_error_detected (struct pci_dev *pdev) {struct cxl_dev_state *cxlds =pci_get_drvdata (pdev);drivers/cxl/core/pci_aer.c new file +0
-0Hi Dave,
While adding this into my series and looking closer, I believe that
pci_aer.cfilename does not accurately reflect its content. The only AER-related code pertains to CX1.1, which is now a corner case. The AER driver, not CXL driver, logs AER errors for all other scenarios. There is no AER processing in the CXL driver except for CXL1.1 dports.Most of the code in this file is related to or will be directed towards CXL RAS. It's important to note that AER and RAS are distinct and separate registers, with CXL RAS only utilizing the AER interrupt notification and UIE/CIE.
Quoted 6 lines
I am concerned we are conflating AER and RAS, which may be misleading for others. We have other instances of this confusion, such as in the
cxl_aer_correctable_error()andcxl_aer_uncorrectable_error()functions, which log RAS instead of AER.My series aims to clarify the distinction between AER and CXL errors, categorizing CXL errors as a separate error class. At the least, can we add a documentation block at the top to indicate this source file provides functionality for CXL RAS?
Sure we can do that. Also, we have a core/ras.c already. Any suggestions on how do we want improve this? Move things to ras.c? Rename cxl_aer.c to something else? Rename current ras.c to something else?
Quoted 223 lines
-Terry
Quoted 219 lines
drivers/cxl/core/pci_aer.c new file +168
-0@@ -0,0 +1,168 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* Copyright(c) 2025 AMD Corporation. All rights reserved. */ +/* Copyright(c) 2025 Intel Corporation. All rights reserved. */ +#include <linux/pci.h> +#include <linux/aer.h> +#include <cxlpci.h> +#include <cxlmem.h> +#include <cxl.h> +#include "core.h" + +static void cxl_dport_map_rch_aer (struct cxl_dport *dport) +{ +resource_size_t aer_phys; +struct device *host; +u16 aer_cap; + + aer_cap =cxl_rcrb_to_aer (dport->dport_dev, dport->rcrb.base); +if (aer_cap) { + host = dport->reg_map.host; + aer_phys = aer_cap + dport->rcrb.base; + dport->regs.dport_aer =devm_cxl_iomap_block (host, aer_phys, +sizeof (struct aer_capability_regs )); + } +} + +static void cxl_dport_map_ras (struct cxl_dport *dport) +{ +struct cxl_register_map *map = &dport->reg_map; +struct device *dev = dport->dport_dev; + +if (!map->component_map.ras.valid) +dev_dbg (dev,"RAS registers not found\n" ); +else if (cxl_map_component_regs (map, &dport->regs.component, +BIT (CXL_CM_CAP_CAP_ID_RAS ))) +dev_dbg (dev,"Failed to map RAS capability.\n" ); +} + +static void cxl_disable_rch_root_ints (struct cxl_dport *dport) +{ +void __iomem *aer_base = dport->regs.dport_aer; +u32 aer_cmd_mask, aer_cmd; + +if (!aer_base) +return ; + +/* +* Disable RCH root port command interrupts. +* CXL 3.0 12.2.1.1 - RCH Downstream Port-detected Errors +* +* This sequence may not be necessary. CXL spec states disabling +* the root cmd register's interrupts is required. But, PCI spec +* shows these are disabled by default on reset. +*/ + aer_cmd_mask = (PCI_ERR_ROOT_CMD_COR_EN | +PCI_ERR_ROOT_CMD_NONFATAL_EN | +PCI_ERR_ROOT_CMD_FATAL_EN ); + aer_cmd =readl (aer_base +PCI_ERR_ROOT_COMMAND ); + aer_cmd &= ~aer_cmd_mask; +writel (aer_cmd, aer_base +PCI_ERR_ROOT_COMMAND ); +} + +/** +* cxl_dport_init_ras_reporting - Setup CXL RAS report on this dport +* @dport: the cxl_dport that needs to be initialized +* @host: host device for devm operations +*/ +void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host) +{ + dport->reg_map.host = host; +cxl_dport_map_ras (dport); + +if (dport->rch) { +struct pci_host_bridge *host_bridge =to_pci_host_bridge (dport->dport_dev); + +if (!host_bridge->native_aer) +return ; + +cxl_dport_map_rch_aer (dport); +cxl_disable_rch_root_ints (dport); + } +} +EXPORT_SYMBOL_NS_GPL (cxl_dport_init_ras_reporting,"CXL" ); + +static void cxl_handle_rdport_cor_ras (struct cxl_dev_state *cxlds, +struct cxl_dport *dport) +{ +return __cxl_handle_cor_ras (cxlds, dport->regs.ras); +} + +static bool cxl_handle_rdport_ras (struct cxl_dev_state *cxlds, +struct cxl_dport *dport) +{ +return __cxl_handle_ras (cxlds, dport->regs.ras); +} + +/* +* Copy the AER capability registers using 32 bit read accesses. +* This is necessary because RCRB AER capability is MMIO mapped. Clear the +* status after copying. +* +* @aer_base: base address of AER capability block in RCRB +* @aer_regs: destination for copying AER capability +*/ +static bool cxl_rch_get_aer_info (void __iomem *aer_base, +struct aer_capability_regs *aer_regs) +{ +int read_cnt =sizeof (struct aer_capability_regs ) /sizeof (u32); +u32 *aer_regs_buf = (u32 *)aer_regs; +int n; + +if (!aer_base) +return false; + +/* Use readl() to guarantee 32-bit accesses */ +for (n =0 ; n < read_cnt; n++) + aer_regs_buf[n] =readl (aer_base + n *sizeof (u32)); + +writel (aer_regs->uncor_status, aer_base +PCI_ERR_UNCOR_STATUS ); +writel (aer_regs->cor_status, aer_base +PCI_ERR_COR_STATUS ); + +return true; +} + +/* Get AER severity. Return false if there is no error. */ +static bool cxl_rch_get_aer_severity (struct aer_capability_regs *aer_regs, +int *severity) +{ +if (aer_regs->uncor_status & ~aer_regs->uncor_mask) { +if (aer_regs->uncor_status &PCI_ERR_ROOT_FATAL_RCV ) + *severity =AER_FATAL ; +else + *severity =AER_NONFATAL ; +return true; + } + +if (aer_regs->cor_status & ~aer_regs->cor_mask) { + *severity =AER_CORRECTABLE ; +return true; + } + +return false; +} + +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) +{ +struct pci_dev *pdev =to_pci_dev (cxlds->dev); +struct aer_capability_regs aer_regs; +struct cxl_dport *dport; +int severity; + + struct cxl_port *port __free(put_cxl_port) = +cxl_pci_find_port (pdev, &dport); +if (!port) +return ; + +if (!cxl_rch_get_aer_info (dport->regs.dport_aer, &aer_regs)) +return ; + +if (!cxl_rch_get_aer_severity (&aer_regs, &severity)) +return ; + +pci_print_aer (pdev, severity, &aer_regs); + +if (severity ==AER_CORRECTABLE ) +cxl_handle_rdport_cor_ras (cxlds, dport); +else +cxl_handle_rdport_ras (cxlds, dport); +}drivers/cxl/cxl.h +0
-8@@ -759,14 +759,6 @@ struct cxl_dport *devm_cxl_add_rch_dport(struct cxl_port *port, struct device *dport_dev, int port_id, resource_size_t rcrb); -#ifdef CONFIG_PCIEAER_CXL -void cxl_setup_parent_dport (struct device *host,struct cxl_dport *dport); -void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host); -#else -static inline void cxl_dport_init_ras_reporting (struct cxl_dport *dport, -struct device *host) { } -#endif -struct cxl_decoder *to_cxl_decoder (struct device *dev);struct cxl_root_decoder *to_cxl_root_decoder (struct device *dev);struct cxl_switch_decoder *to_cxl_switch_decoder (struct device *dev);drivers/cxl/cxlpci.h +8
-0@@ -135,4 +135,12 @@ void read_cdat_data(struct cxl_port *port);void cxl_cor_error_detected (struct pci_dev *pdev);pci_ers_result_t cxl_error_detected (struct pci_dev *pdev,pci_channel_state_t state); + +#ifdef CONFIG_PCIEAER_CXL +void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host); +#else +static inline void cxl_dport_init_ras_reporting (struct cxl_dport *dport, +struct device *host) { } +#endif +#endif /* __CXL_PCI_H__ */ tools/testing/cxl/Kbuild +1
-0@@ -68,6 +68,7 @@ cxl_core-$(CONFIG_CXL_REGION) += $(CXL_CORE_SRC)/region.o cxl_core-$(CONFIG_CXL_MCE) += $(CXL_CORE_SRC)/mce.o cxl_core-$(CONFIG_CXL_FEATURES) += $(CXL_CORE_SRC)/features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += $(CXL_CORE_SRC)/edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += $(CXL_CORE_SRC)/pci_aer.o cxl_core-y += config_check.o cxl_core-y += cxl_core_test.o cxl_core-y += cxl_core_exports.obase-commit:
347e9f5043c89695b01e66b3ed111755afcf1911
Quoted 266 lines · Dave Jiang
Quoted 263 lines · Terry
drivers/cxl/core/pci_aer.c
Quoted 250 lines · Dave Jiang
Remove from
core/pci_aer.cand move theCONFIG_PCIEAER_CXLifdef tocxlpci.hto deal with the exported symbols as needed. There is enough AER handling code (and more to come) to move the AER code to its own C file.Cc Robert Richter <[email protected]>Cc Terry Bowman <[email protected]>Signed-off-by Dave Jiang <[email protected]>---
v2:
- Address 0-day reprot ofcxl_handle_rdport_errorssymbol undefined--- drivers/cxl/core/Makefile | 1 + drivers/cxl/core/core.h | 9 ++ drivers/cxl/core/pci.c | 171 +------------------------------------ drivers/cxl/core/pci_aer.c | 168 ++++++++++++++++++++++++++++++++++++ drivers/cxl/cxl.h | 8 -- drivers/cxl/cxlpci.h | 8 ++ tools/testing/cxl/Kbuild | 1 + 7 files changed, 189 insertions(+), 177 deletions(-) create mode 100644 drivers/cxl/core/pci_aer.cdrivers/cxl/core/Makefile +1
-0@@ -21,3 +21,4 @@ cxl_core-$(CONFIG_CXL_REGION) += region.o cxl_core-$(CONFIG_CXL_MCE) += mce.o cxl_core-$(CONFIG_CXL_FEATURES) += features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += pci_aer.odrivers/cxl/core/core.h +9
-0@@ -136,4 +136,13 @@ int cxl_set_feature(struct cxl_mailbox *cxl_mbox, const uuid_t *feat_uuid, u16 *return_code);#endif +#ifdef CONFIG_PCIEAER_CXL +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds); +#else +static inline void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) {} +#endif + +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +#endif /* __CXL_CORE_H__ */ drivers/cxl/core/pci.c +2
-169@@ -664,8 +664,7 @@ void read_cdat_data(struct cxl_port *port) }EXPORT_SYMBOL_NS_GPL (read_cdat_data,"CXL" ); -static void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds, -void __iomem *ras_base) +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base) {void __iomem *addr;u32 status;@@ -707,8 +706,7 @@ static void header_log_copy(void __iomem *ras_base, u32 *log)* Log the state of the RAS status registers and prepare them to log the * next error status. Return 1 if reset needed. */ -static bool __cxl_handle_ras (struct cxl_dev_state *cxlds, -void __iomem *ras_base) +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base) {u32 hl[CXL_HEADERLOG_SIZE_U32 ];void __iomem *addr;@@ -746,171 +744,6 @@ static bool cxl_handle_endpoint_ras(struct cxl_dev_state *cxlds)return __cxl_handle_ras (cxlds, cxlds->regs.ras); } -#ifdef CONFIG_PCIEAER_CXL - -static void cxl_dport_map_rch_aer (struct cxl_dport *dport) -{ -resource_size_t aer_phys; -struct device *host; -u16 aer_cap; - - aer_cap =cxl_rcrb_to_aer (dport->dport_dev, dport->rcrb.base); -if (aer_cap) { - host = dport->reg_map.host; - aer_phys = aer_cap + dport->rcrb.base; - dport->regs.dport_aer =devm_cxl_iomap_block (host, aer_phys, -sizeof (struct aer_capability_regs )); - } -} - -static void cxl_dport_map_ras (struct cxl_dport *dport) -{ -struct cxl_register_map *map = &dport->reg_map; -struct device *dev = dport->dport_dev; - -if (!map->component_map.ras.valid) -dev_dbg (dev,"RAS registers not found\n" ); -else if (cxl_map_component_regs (map, &dport->regs.component, -BIT (CXL_CM_CAP_CAP_ID_RAS ))) -dev_dbg (dev,"Failed to map RAS capability.\n" ); -} - -static void cxl_disable_rch_root_ints (struct cxl_dport *dport) -{ -void __iomem *aer_base = dport->regs.dport_aer; -u32 aer_cmd_mask, aer_cmd; - -if (!aer_base) -return ; - -/* -* Disable RCH root port command interrupts. -* CXL 3.0 12.2.1.1 - RCH Downstream Port-detected Errors -* -* This sequence may not be necessary. CXL spec states disabling -* the root cmd register's interrupts is required. But, PCI spec -* shows these are disabled by default on reset. -*/ - aer_cmd_mask = (PCI_ERR_ROOT_CMD_COR_EN | -PCI_ERR_ROOT_CMD_NONFATAL_EN | -PCI_ERR_ROOT_CMD_FATAL_EN ); - aer_cmd =readl (aer_base +PCI_ERR_ROOT_COMMAND ); - aer_cmd &= ~aer_cmd_mask; -writel (aer_cmd, aer_base +PCI_ERR_ROOT_COMMAND ); -} - -/** -* cxl_dport_init_ras_reporting - Setup CXL RAS report on this dport -* @dport: the cxl_dport that needs to be initialized -* @host: host device for devm operations -*/ -void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host) -{ - dport->reg_map.host = host; -cxl_dport_map_ras (dport); - -if (dport->rch) { -struct pci_host_bridge *host_bridge =to_pci_host_bridge (dport->dport_dev); - -if (!host_bridge->native_aer) -return ; - -cxl_dport_map_rch_aer (dport); -cxl_disable_rch_root_ints (dport); - } -} -EXPORT_SYMBOL_NS_GPL (cxl_dport_init_ras_reporting,"CXL" ); - -static void cxl_handle_rdport_cor_ras (struct cxl_dev_state *cxlds, -struct cxl_dport *dport) -{ -return __cxl_handle_cor_ras (cxlds, dport->regs.ras); -} - -static bool cxl_handle_rdport_ras (struct cxl_dev_state *cxlds, -struct cxl_dport *dport) -{ -return __cxl_handle_ras (cxlds, dport->regs.ras); -} - -/* -* Copy the AER capability registers using 32 bit read accesses. -* This is necessary because RCRB AER capability is MMIO mapped. Clear the -* status after copying. -* -* @aer_base: base address of AER capability block in RCRB -* @aer_regs: destination for copying AER capability -*/ -static bool cxl_rch_get_aer_info (void __iomem *aer_base, -struct aer_capability_regs *aer_regs) -{ -int read_cnt =sizeof (struct aer_capability_regs ) /sizeof (u32); -u32 *aer_regs_buf = (u32 *)aer_regs; -int n; - -if (!aer_base) -return false; - -/* Use readl() to guarantee 32-bit accesses */ -for (n =0 ; n < read_cnt; n++) - aer_regs_buf[n] =readl (aer_base + n *sizeof (u32)); - -writel (aer_regs->uncor_status, aer_base +PCI_ERR_UNCOR_STATUS ); -writel (aer_regs->cor_status, aer_base +PCI_ERR_COR_STATUS ); - -return true; -} - -/* Get AER severity. Return false if there is no error. */ -static bool cxl_rch_get_aer_severity (struct aer_capability_regs *aer_regs, -int *severity) -{ -if (aer_regs->uncor_status & ~aer_regs->uncor_mask) { -if (aer_regs->uncor_status &PCI_ERR_ROOT_FATAL_RCV ) - *severity =AER_FATAL ; -else - *severity =AER_NONFATAL ; -return true; - } - -if (aer_regs->cor_status & ~aer_regs->cor_mask) { - *severity =AER_CORRECTABLE ; -return true; - } - -return false; -} - -static void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) -{ -struct pci_dev *pdev =to_pci_dev (cxlds->dev); -struct aer_capability_regs aer_regs; -struct cxl_dport *dport; -int severity; - - struct cxl_port *port __free(put_cxl_port) = -cxl_pci_find_port (pdev, &dport); -if (!port) -return ; - -if (!cxl_rch_get_aer_info (dport->regs.dport_aer, &aer_regs)) -return ; - -if (!cxl_rch_get_aer_severity (&aer_regs, &severity)) -return ; - -pci_print_aer (pdev, severity, &aer_regs); - -if (severity ==AER_CORRECTABLE ) -cxl_handle_rdport_cor_ras (cxlds, dport); -else -cxl_handle_rdport_ras (cxlds, dport); -} - -#else -static void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) { } -#endif -void cxl_cor_error_detected (struct pci_dev *pdev) {struct cxl_dev_state *cxlds =pci_get_drvdata (pdev);drivers/cxl/core/pci_aer.c new file +0
-0Hi Dave,
While adding this into my series and looking closer, I believe that
pci_aer.cfilename does not accurately reflect its content. The only AER-related code pertains to CX1.1, which is now a corner case. The AER driver, not CXL driver, logs AER errors for all other scenarios. There is no AER processing in the CXL driver except for CXL1.1 dports.Most of the code in this file is related to or will be directed towards CXL RAS. It's important to note that AER and RAS are distinct and separate registers, with CXL RAS only utilizing the AER interrupt notification and UIE/CIE.
Quoted 10 lines
Quoted 6 lines
I am concerned we are conflating AER and RAS, which may be misleading for others. We have other instances of this confusion, such as in the
cxl_aer_correctable_error()andcxl_aer_uncorrectable_error()functions, which log RAS instead of AER.My series aims to clarify the distinction between AER and CXL errors, categorizing CXL errors as a separate error class. At the least, can we add a documentation block at the top to indicate this source file provides functionality for CXL RAS?
Sure we can do that. Also, we have a
core/ras.calready. Any suggestions on how do we want improve this? Move things toras.c? Renamecxl_aer.cto something else? Rename currentras.cto something else?Quoted 1 line
I don't think we want to combine the os-first and fw-first code in a single file. The separation leveraging the Makefile conditional compile is good because fw-first doesn't use PCIEAER_CXL while os-first does. I think core/ras.c is good as-is.
cxl_aer.c is in the drivers/pci/pcie directory and is logically part of the AER service driver. This file includes the kfifo and logic for offloading CXL RAS errors to the CXL driver. I think the cxl_aer.c naming is fine but understand it could be argued to use cxl_ras.c instead. My concern was in the CXL driver AER-RAS naming but the same rules could be applied here as well.
To improve the original situation mentioned I recommend we use 'ras' in what is named pci_aer.c here? Maybe 'pci_ras.c' instead? It's not perfect but would be more precise. This feels somewhat nitpicky on my part and will not be an issue if we leave as-is. Thanks for considering.
-Terry
Quoted 223 lines
Quoted 222 lines
-Terry
Quoted 219 lines
drivers/cxl/core/pci_aer.c new file +168
-0@@ -0,0 +1,168 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* Copyright(c) 2025 AMD Corporation. All rights reserved. */ +/* Copyright(c) 2025 Intel Corporation. All rights reserved. */ +#include <linux/pci.h> +#include <linux/aer.h> +#include <cxlpci.h> +#include <cxlmem.h> +#include <cxl.h> +#include "core.h" + +static void cxl_dport_map_rch_aer (struct cxl_dport *dport) +{ +resource_size_t aer_phys; +struct device *host; +u16 aer_cap; + + aer_cap =cxl_rcrb_to_aer (dport->dport_dev, dport->rcrb.base); +if (aer_cap) { + host = dport->reg_map.host; + aer_phys = aer_cap + dport->rcrb.base; + dport->regs.dport_aer =devm_cxl_iomap_block (host, aer_phys, +sizeof (struct aer_capability_regs )); + } +} + +static void cxl_dport_map_ras (struct cxl_dport *dport) +{ +struct cxl_register_map *map = &dport->reg_map; +struct device *dev = dport->dport_dev; + +if (!map->component_map.ras.valid) +dev_dbg (dev,"RAS registers not found\n" ); +else if (cxl_map_component_regs (map, &dport->regs.component, +BIT (CXL_CM_CAP_CAP_ID_RAS ))) +dev_dbg (dev,"Failed to map RAS capability.\n" ); +} + +static void cxl_disable_rch_root_ints (struct cxl_dport *dport) +{ +void __iomem *aer_base = dport->regs.dport_aer; +u32 aer_cmd_mask, aer_cmd; + +if (!aer_base) +return ; + +/* +* Disable RCH root port command interrupts. +* CXL 3.0 12.2.1.1 - RCH Downstream Port-detected Errors +* +* This sequence may not be necessary. CXL spec states disabling +* the root cmd register's interrupts is required. But, PCI spec +* shows these are disabled by default on reset. +*/ + aer_cmd_mask = (PCI_ERR_ROOT_CMD_COR_EN | +PCI_ERR_ROOT_CMD_NONFATAL_EN | +PCI_ERR_ROOT_CMD_FATAL_EN ); + aer_cmd =readl (aer_base +PCI_ERR_ROOT_COMMAND ); + aer_cmd &= ~aer_cmd_mask; +writel (aer_cmd, aer_base +PCI_ERR_ROOT_COMMAND ); +} + +/** +* cxl_dport_init_ras_reporting - Setup CXL RAS report on this dport +* @dport: the cxl_dport that needs to be initialized +* @host: host device for devm operations +*/ +void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host) +{ + dport->reg_map.host = host; +cxl_dport_map_ras (dport); + +if (dport->rch) { +struct pci_host_bridge *host_bridge =to_pci_host_bridge (dport->dport_dev); + +if (!host_bridge->native_aer) +return ; + +cxl_dport_map_rch_aer (dport); +cxl_disable_rch_root_ints (dport); + } +} +EXPORT_SYMBOL_NS_GPL (cxl_dport_init_ras_reporting,"CXL" ); + +static void cxl_handle_rdport_cor_ras (struct cxl_dev_state *cxlds, +struct cxl_dport *dport) +{ +return __cxl_handle_cor_ras (cxlds, dport->regs.ras); +} + +static bool cxl_handle_rdport_ras (struct cxl_dev_state *cxlds, +struct cxl_dport *dport) +{ +return __cxl_handle_ras (cxlds, dport->regs.ras); +} + +/* +* Copy the AER capability registers using 32 bit read accesses. +* This is necessary because RCRB AER capability is MMIO mapped. Clear the +* status after copying. +* +* @aer_base: base address of AER capability block in RCRB +* @aer_regs: destination for copying AER capability +*/ +static bool cxl_rch_get_aer_info (void __iomem *aer_base, +struct aer_capability_regs *aer_regs) +{ +int read_cnt =sizeof (struct aer_capability_regs ) /sizeof (u32); +u32 *aer_regs_buf = (u32 *)aer_regs; +int n; + +if (!aer_base) +return false; + +/* Use readl() to guarantee 32-bit accesses */ +for (n =0 ; n < read_cnt; n++) + aer_regs_buf[n] =readl (aer_base + n *sizeof (u32)); + +writel (aer_regs->uncor_status, aer_base +PCI_ERR_UNCOR_STATUS ); +writel (aer_regs->cor_status, aer_base +PCI_ERR_COR_STATUS ); + +return true; +} + +/* Get AER severity. Return false if there is no error. */ +static bool cxl_rch_get_aer_severity (struct aer_capability_regs *aer_regs, +int *severity) +{ +if (aer_regs->uncor_status & ~aer_regs->uncor_mask) { +if (aer_regs->uncor_status &PCI_ERR_ROOT_FATAL_RCV ) + *severity =AER_FATAL ; +else + *severity =AER_NONFATAL ; +return true; + } + +if (aer_regs->cor_status & ~aer_regs->cor_mask) { + *severity =AER_CORRECTABLE ; +return true; + } + +return false; +} + +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) +{ +struct pci_dev *pdev =to_pci_dev (cxlds->dev); +struct aer_capability_regs aer_regs; +struct cxl_dport *dport; +int severity; + + struct cxl_port *port __free(put_cxl_port) = +cxl_pci_find_port (pdev, &dport); +if (!port) +return ; + +if (!cxl_rch_get_aer_info (dport->regs.dport_aer, &aer_regs)) +return ; + +if (!cxl_rch_get_aer_severity (&aer_regs, &severity)) +return ; + +pci_print_aer (pdev, severity, &aer_regs); + +if (severity ==AER_CORRECTABLE ) +cxl_handle_rdport_cor_ras (cxlds, dport); +else +cxl_handle_rdport_ras (cxlds, dport); +}drivers/cxl/cxl.h +0
-8@@ -759,14 +759,6 @@ struct cxl_dport *devm_cxl_add_rch_dport(struct cxl_port *port, struct device *dport_dev, int port_id, resource_size_t rcrb); -#ifdef CONFIG_PCIEAER_CXL -void cxl_setup_parent_dport (struct device *host,struct cxl_dport *dport); -void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host); -#else -static inline void cxl_dport_init_ras_reporting (struct cxl_dport *dport, -struct device *host) { } -#endif -struct cxl_decoder *to_cxl_decoder (struct device *dev);struct cxl_root_decoder *to_cxl_root_decoder (struct device *dev);struct cxl_switch_decoder *to_cxl_switch_decoder (struct device *dev);drivers/cxl/cxlpci.h +8
-0@@ -135,4 +135,12 @@ void read_cdat_data(struct cxl_port *port);void cxl_cor_error_detected (struct pci_dev *pdev);pci_ers_result_t cxl_error_detected (struct pci_dev *pdev,pci_channel_state_t state); + +#ifdef CONFIG_PCIEAER_CXL +void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host); +#else +static inline void cxl_dport_init_ras_reporting (struct cxl_dport *dport, +struct device *host) { } +#endif +#endif /* __CXL_PCI_H__ */ tools/testing/cxl/Kbuild +1
-0@@ -68,6 +68,7 @@ cxl_core-$(CONFIG_CXL_REGION) += $(CXL_CORE_SRC)/region.o cxl_core-$(CONFIG_CXL_MCE) += $(CXL_CORE_SRC)/mce.o cxl_core-$(CONFIG_CXL_FEATURES) += $(CXL_CORE_SRC)/features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += $(CXL_CORE_SRC)/edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += $(CXL_CORE_SRC)/pci_aer.o cxl_core-y += config_check.o cxl_core-y += cxl_core_test.o cxl_core-y += cxl_core_exports.obase-commit:
347e9f5043c89695b01e66b3ed111755afcf1911
Dave Jiang
maintainerQuoted 282 lines · Terry
Quoted 266 lines · Dave Jiang
Quoted 263 lines · Terry
drivers/cxl/core/pci_aer.c
Quoted 250 lines · Dave Jiang
Remove from
core/pci_aer.cand move theCONFIG_PCIEAER_CXLifdef tocxlpci.hto deal with the exported symbols as needed. There is enough AER handling code (and more to come) to move the AER code to its own C file.Cc Robert Richter <[email protected]>Cc Terry Bowman <[email protected]>Signed-off-by Dave Jiang <[email protected]>---
v2:
- Address 0-day reprot ofcxl_handle_rdport_errorssymbol undefined--- drivers/cxl/core/Makefile | 1 + drivers/cxl/core/core.h | 9 ++ drivers/cxl/core/pci.c | 171 +------------------------------------ drivers/cxl/core/pci_aer.c | 168 ++++++++++++++++++++++++++++++++++++ drivers/cxl/cxl.h | 8 -- drivers/cxl/cxlpci.h | 8 ++ tools/testing/cxl/Kbuild | 1 + 7 files changed, 189 insertions(+), 177 deletions(-) create mode 100644 drivers/cxl/core/pci_aer.cdrivers/cxl/core/Makefile +1
-0@@ -21,3 +21,4 @@ cxl_core-$(CONFIG_CXL_REGION) += region.o cxl_core-$(CONFIG_CXL_MCE) += mce.o cxl_core-$(CONFIG_CXL_FEATURES) += features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += pci_aer.odrivers/cxl/core/core.h +9
-0@@ -136,4 +136,13 @@ int cxl_set_feature(struct cxl_mailbox *cxl_mbox, const uuid_t *feat_uuid, u16 *return_code);#endif +#ifdef CONFIG_PCIEAER_CXL +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds); +#else +static inline void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) {} +#endif + +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +#endif /* __CXL_CORE_H__ */ drivers/cxl/core/pci.c +2
-169@@ -664,8 +664,7 @@ void read_cdat_data(struct cxl_port *port) }EXPORT_SYMBOL_NS_GPL (read_cdat_data,"CXL" ); -static void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds, -void __iomem *ras_base) +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base) {void __iomem *addr;u32 status;@@ -707,8 +706,7 @@ static void header_log_copy(void __iomem *ras_base, u32 *log)* Log the state of the RAS status registers and prepare them to log the * next error status. Return 1 if reset needed. */ -static bool __cxl_handle_ras (struct cxl_dev_state *cxlds, -void __iomem *ras_base) +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base) {u32 hl[CXL_HEADERLOG_SIZE_U32 ];void __iomem *addr;@@ -746,171 +744,6 @@ static bool cxl_handle_endpoint_ras(struct cxl_dev_state *cxlds)return __cxl_handle_ras (cxlds, cxlds->regs.ras); } -#ifdef CONFIG_PCIEAER_CXL - -static void cxl_dport_map_rch_aer (struct cxl_dport *dport) -{ -resource_size_t aer_phys; -struct device *host; -u16 aer_cap; - - aer_cap =cxl_rcrb_to_aer (dport->dport_dev, dport->rcrb.base); -if (aer_cap) { - host = dport->reg_map.host; - aer_phys = aer_cap + dport->rcrb.base; - dport->regs.dport_aer =devm_cxl_iomap_block (host, aer_phys, -sizeof (struct aer_capability_regs )); - } -} - -static void cxl_dport_map_ras (struct cxl_dport *dport) -{ -struct cxl_register_map *map = &dport->reg_map; -struct device *dev = dport->dport_dev; - -if (!map->component_map.ras.valid) -dev_dbg (dev,"RAS registers not found\n" ); -else if (cxl_map_component_regs (map, &dport->regs.component, -BIT (CXL_CM_CAP_CAP_ID_RAS ))) -dev_dbg (dev,"Failed to map RAS capability.\n" ); -} - -static void cxl_disable_rch_root_ints (struct cxl_dport *dport) -{ -void __iomem *aer_base = dport->regs.dport_aer; -u32 aer_cmd_mask, aer_cmd; - -if (!aer_base) -return ; - -/* -* Disable RCH root port command interrupts. -* CXL 3.0 12.2.1.1 - RCH Downstream Port-detected Errors -* -* This sequence may not be necessary. CXL spec states disabling -* the root cmd register's interrupts is required. But, PCI spec -* shows these are disabled by default on reset. -*/ - aer_cmd_mask = (PCI_ERR_ROOT_CMD_COR_EN | -PCI_ERR_ROOT_CMD_NONFATAL_EN | -PCI_ERR_ROOT_CMD_FATAL_EN ); - aer_cmd =readl (aer_base +PCI_ERR_ROOT_COMMAND ); - aer_cmd &= ~aer_cmd_mask; -writel (aer_cmd, aer_base +PCI_ERR_ROOT_COMMAND ); -} - -/** -* cxl_dport_init_ras_reporting - Setup CXL RAS report on this dport -* @dport: the cxl_dport that needs to be initialized -* @host: host device for devm operations -*/ -void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host) -{ - dport->reg_map.host = host; -cxl_dport_map_ras (dport); - -if (dport->rch) { -struct pci_host_bridge *host_bridge =to_pci_host_bridge (dport->dport_dev); - -if (!host_bridge->native_aer) -return ; - -cxl_dport_map_rch_aer (dport); -cxl_disable_rch_root_ints (dport); - } -} -EXPORT_SYMBOL_NS_GPL (cxl_dport_init_ras_reporting,"CXL" ); - -static void cxl_handle_rdport_cor_ras (struct cxl_dev_state *cxlds, -struct cxl_dport *dport) -{ -return __cxl_handle_cor_ras (cxlds, dport->regs.ras); -} - -static bool cxl_handle_rdport_ras (struct cxl_dev_state *cxlds, -struct cxl_dport *dport) -{ -return __cxl_handle_ras (cxlds, dport->regs.ras); -} - -/* -* Copy the AER capability registers using 32 bit read accesses. -* This is necessary because RCRB AER capability is MMIO mapped. Clear the -* status after copying. -* -* @aer_base: base address of AER capability block in RCRB -* @aer_regs: destination for copying AER capability -*/ -static bool cxl_rch_get_aer_info (void __iomem *aer_base, -struct aer_capability_regs *aer_regs) -{ -int read_cnt =sizeof (struct aer_capability_regs ) /sizeof (u32); -u32 *aer_regs_buf = (u32 *)aer_regs; -int n; - -if (!aer_base) -return false; - -/* Use readl() to guarantee 32-bit accesses */ -for (n =0 ; n < read_cnt; n++) - aer_regs_buf[n] =readl (aer_base + n *sizeof (u32)); - -writel (aer_regs->uncor_status, aer_base +PCI_ERR_UNCOR_STATUS ); -writel (aer_regs->cor_status, aer_base +PCI_ERR_COR_STATUS ); - -return true; -} - -/* Get AER severity. Return false if there is no error. */ -static bool cxl_rch_get_aer_severity (struct aer_capability_regs *aer_regs, -int *severity) -{ -if (aer_regs->uncor_status & ~aer_regs->uncor_mask) { -if (aer_regs->uncor_status &PCI_ERR_ROOT_FATAL_RCV ) - *severity =AER_FATAL ; -else - *severity =AER_NONFATAL ; -return true; - } - -if (aer_regs->cor_status & ~aer_regs->cor_mask) { - *severity =AER_CORRECTABLE ; -return true; - } - -return false; -} - -static void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) -{ -struct pci_dev *pdev =to_pci_dev (cxlds->dev); -struct aer_capability_regs aer_regs; -struct cxl_dport *dport; -int severity; - - struct cxl_port *port __free(put_cxl_port) = -cxl_pci_find_port (pdev, &dport); -if (!port) -return ; - -if (!cxl_rch_get_aer_info (dport->regs.dport_aer, &aer_regs)) -return ; - -if (!cxl_rch_get_aer_severity (&aer_regs, &severity)) -return ; - -pci_print_aer (pdev, severity, &aer_regs); - -if (severity ==AER_CORRECTABLE ) -cxl_handle_rdport_cor_ras (cxlds, dport); -else -cxl_handle_rdport_ras (cxlds, dport); -} - -#else -static void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) { } -#endif -void cxl_cor_error_detected (struct pci_dev *pdev) {struct cxl_dev_state *cxlds =pci_get_drvdata (pdev);drivers/cxl/core/pci_aer.c new file +0
-0Hi Dave,
While adding this into my series and looking closer, I believe that
pci_aer.cfilename does not accurately reflect its content. The only AER-related code pertains to CX1.1, which is now a corner case. The AER driver, not CXL driver, logs AER errors for all other scenarios. There is no AER processing in the CXL driver except for CXL1.1 dports.Most of the code in this file is related to or will be directed towards CXL RAS. It's important to note that AER and RAS are distinct and separate registers, with CXL RAS only utilizing the AER interrupt notification and UIE/CIE.
Quoted 10 lines
Quoted 6 lines
I am concerned we are conflating AER and RAS, which may be misleading for others. We have other instances of this confusion, such as in the
cxl_aer_correctable_error()andcxl_aer_uncorrectable_error()functions, which log RAS instead of AER.My series aims to clarify the distinction between AER and CXL errors, categorizing CXL errors as a separate error class. At the least, can we add a documentation block at the top to indicate this source file provides functionality for CXL RAS?
Sure we can do that. Also, we have a
core/ras.calready. Any suggestions on how do we want improve this? Move things toras.c? Renamecxl_aer.cto something else? Rename currentras.cto something else?Quoted 1 line
I don't think we want to combine the os-first and fw-first code in a single file. The separation leveraging the
Makefileconditional compile is good because fw-first doesn't usePCIEAER_CXLwhile os-first does. I thinkcore/ras.cis good as-is.
Quoted 8 lines
cxl_aer.cis in thedrivers/pci/pciedirectory and is logically part of the AER service driver. This file includes the kfifo and logic for offloading CXL RAS errors to the CXL driver. I think thecxl_aer.cnaming is fine but understand it could be argued to usecxl_ras.cinstead. My concern was in the CXL driver AER-RAS naming but the same rules could be applied here as well.To improve the original situation mentioned I recommend we use 'ras' in what is named
pci_aer.chere? Maybe 'pci_ras.c' instead? It's not perfect but would be more precise. This feels somewhat nitpicky on my part and will not be an issue if we leave as-is. Thanks for considering.
Thanks for the detailed explanation. I'll rename it to pci_ras.c.
Quoted 227 lines
-Terry
Quoted 223 lines
Quoted 222 lines
-Terry
Quoted 219 lines
drivers/cxl/core/pci_aer.c new file +168
-0@@ -0,0 +1,168 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* Copyright(c) 2025 AMD Corporation. All rights reserved. */ +/* Copyright(c) 2025 Intel Corporation. All rights reserved. */ +#include <linux/pci.h> +#include <linux/aer.h> +#include <cxlpci.h> +#include <cxlmem.h> +#include <cxl.h> +#include "core.h" + +static void cxl_dport_map_rch_aer (struct cxl_dport *dport) +{ +resource_size_t aer_phys; +struct device *host; +u16 aer_cap; + + aer_cap =cxl_rcrb_to_aer (dport->dport_dev, dport->rcrb.base); +if (aer_cap) { + host = dport->reg_map.host; + aer_phys = aer_cap + dport->rcrb.base; + dport->regs.dport_aer =devm_cxl_iomap_block (host, aer_phys, +sizeof (struct aer_capability_regs )); + } +} + +static void cxl_dport_map_ras (struct cxl_dport *dport) +{ +struct cxl_register_map *map = &dport->reg_map; +struct device *dev = dport->dport_dev; + +if (!map->component_map.ras.valid) +dev_dbg (dev,"RAS registers not found\n" ); +else if (cxl_map_component_regs (map, &dport->regs.component, +BIT (CXL_CM_CAP_CAP_ID_RAS ))) +dev_dbg (dev,"Failed to map RAS capability.\n" ); +} + +static void cxl_disable_rch_root_ints (struct cxl_dport *dport) +{ +void __iomem *aer_base = dport->regs.dport_aer; +u32 aer_cmd_mask, aer_cmd; + +if (!aer_base) +return ; + +/* +* Disable RCH root port command interrupts. +* CXL 3.0 12.2.1.1 - RCH Downstream Port-detected Errors +* +* This sequence may not be necessary. CXL spec states disabling +* the root cmd register's interrupts is required. But, PCI spec +* shows these are disabled by default on reset. +*/ + aer_cmd_mask = (PCI_ERR_ROOT_CMD_COR_EN | +PCI_ERR_ROOT_CMD_NONFATAL_EN | +PCI_ERR_ROOT_CMD_FATAL_EN ); + aer_cmd =readl (aer_base +PCI_ERR_ROOT_COMMAND ); + aer_cmd &= ~aer_cmd_mask; +writel (aer_cmd, aer_base +PCI_ERR_ROOT_COMMAND ); +} + +/** +* cxl_dport_init_ras_reporting - Setup CXL RAS report on this dport +* @dport: the cxl_dport that needs to be initialized +* @host: host device for devm operations +*/ +void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host) +{ + dport->reg_map.host = host; +cxl_dport_map_ras (dport); + +if (dport->rch) { +struct pci_host_bridge *host_bridge =to_pci_host_bridge (dport->dport_dev); + +if (!host_bridge->native_aer) +return ; + +cxl_dport_map_rch_aer (dport); +cxl_disable_rch_root_ints (dport); + } +} +EXPORT_SYMBOL_NS_GPL (cxl_dport_init_ras_reporting,"CXL" ); + +static void cxl_handle_rdport_cor_ras (struct cxl_dev_state *cxlds, +struct cxl_dport *dport) +{ +return __cxl_handle_cor_ras (cxlds, dport->regs.ras); +} + +static bool cxl_handle_rdport_ras (struct cxl_dev_state *cxlds, +struct cxl_dport *dport) +{ +return __cxl_handle_ras (cxlds, dport->regs.ras); +} + +/* +* Copy the AER capability registers using 32 bit read accesses. +* This is necessary because RCRB AER capability is MMIO mapped. Clear the +* status after copying. +* +* @aer_base: base address of AER capability block in RCRB +* @aer_regs: destination for copying AER capability +*/ +static bool cxl_rch_get_aer_info (void __iomem *aer_base, +struct aer_capability_regs *aer_regs) +{ +int read_cnt =sizeof (struct aer_capability_regs ) /sizeof (u32); +u32 *aer_regs_buf = (u32 *)aer_regs; +int n; + +if (!aer_base) +return false; + +/* Use readl() to guarantee 32-bit accesses */ +for (n =0 ; n < read_cnt; n++) + aer_regs_buf[n] =readl (aer_base + n *sizeof (u32)); + +writel (aer_regs->uncor_status, aer_base +PCI_ERR_UNCOR_STATUS ); +writel (aer_regs->cor_status, aer_base +PCI_ERR_COR_STATUS ); + +return true; +} + +/* Get AER severity. Return false if there is no error. */ +static bool cxl_rch_get_aer_severity (struct aer_capability_regs *aer_regs, +int *severity) +{ +if (aer_regs->uncor_status & ~aer_regs->uncor_mask) { +if (aer_regs->uncor_status &PCI_ERR_ROOT_FATAL_RCV ) + *severity =AER_FATAL ; +else + *severity =AER_NONFATAL ; +return true; + } + +if (aer_regs->cor_status & ~aer_regs->cor_mask) { + *severity =AER_CORRECTABLE ; +return true; + } + +return false; +} + +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) +{ +struct pci_dev *pdev =to_pci_dev (cxlds->dev); +struct aer_capability_regs aer_regs; +struct cxl_dport *dport; +int severity; + + struct cxl_port *port __free(put_cxl_port) = +cxl_pci_find_port (pdev, &dport); +if (!port) +return ; + +if (!cxl_rch_get_aer_info (dport->regs.dport_aer, &aer_regs)) +return ; + +if (!cxl_rch_get_aer_severity (&aer_regs, &severity)) +return ; + +pci_print_aer (pdev, severity, &aer_regs); + +if (severity ==AER_CORRECTABLE ) +cxl_handle_rdport_cor_ras (cxlds, dport); +else +cxl_handle_rdport_ras (cxlds, dport); +}drivers/cxl/cxl.h +0
-8@@ -759,14 +759,6 @@ struct cxl_dport *devm_cxl_add_rch_dport(struct cxl_port *port, struct device *dport_dev, int port_id, resource_size_t rcrb); -#ifdef CONFIG_PCIEAER_CXL -void cxl_setup_parent_dport (struct device *host,struct cxl_dport *dport); -void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host); -#else -static inline void cxl_dport_init_ras_reporting (struct cxl_dport *dport, -struct device *host) { } -#endif -struct cxl_decoder *to_cxl_decoder (struct device *dev);struct cxl_root_decoder *to_cxl_root_decoder (struct device *dev);struct cxl_switch_decoder *to_cxl_switch_decoder (struct device *dev);drivers/cxl/cxlpci.h +8
-0@@ -135,4 +135,12 @@ void read_cdat_data(struct cxl_port *port);void cxl_cor_error_detected (struct pci_dev *pdev);pci_ers_result_t cxl_error_detected (struct pci_dev *pdev,pci_channel_state_t state); + +#ifdef CONFIG_PCIEAER_CXL +void cxl_dport_init_ras_reporting (struct cxl_dport *dport,struct device *host); +#else +static inline void cxl_dport_init_ras_reporting (struct cxl_dport *dport, +struct device *host) { } +#endif +#endif /* __CXL_PCI_H__ */ tools/testing/cxl/Kbuild +1
-0@@ -68,6 +68,7 @@ cxl_core-$(CONFIG_CXL_REGION) += $(CXL_CORE_SRC)/region.o cxl_core-$(CONFIG_CXL_MCE) += $(CXL_CORE_SRC)/mce.o cxl_core-$(CONFIG_CXL_FEATURES) += $(CXL_CORE_SRC)/features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += $(CXL_CORE_SRC)/edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += $(CXL_CORE_SRC)/pci_aer.o cxl_core-y += config_check.o cxl_core-y += cxl_core_test.o cxl_core-y += cxl_core_exports.obase-commit:
347e9f5043c89695b01e66b3ed111755afcf1911
Dave Jiang wrote:
[..]
Quoted 12 lines
Quoted 12 lines
I don't think we want to combine the os-first and fw-first code in a single file. The separation leveraging the
Makefileconditional compile is good because fw-first doesn't usePCIEAER_CXLwhile os-first does. I thinkcore/ras.cis good as-is.
cxl_aer.cis in thedrivers/pci/pciedirectory and is logically part of the AER service driver. This file includes the kfifo and logic for offloading CXL RAS errors to the CXL driver. I think thecxl_aer.cnaming is fine but understand it could be argued to usecxl_ras.cinstead. My concern was in the CXL driver AER-RAS naming but the same rules could be applied here as well.
Quoted 7 lines
Quoted 5 lines
To improve the original situation mentioned I recommend we use 'ras' in what is named
pci_aer.chere? Maybe 'pci_ras.c' instead? It's not perfect but would be more precise. This feels somewhat nitpicky on my part and will not be an issue if we leave as-is. Thanks for considering.Thanks for the detailed explanation. I'll rename it to
pci_ras.c.
Hey guys, please trim your responses.
Between RAS, AER, and EDAC there is significant overlap. You could say that pci_ras.c is also misnamed because the error details being reported are CXL not PCI. I am ok with "RAS" for anything that handles decorating an AER notification with extra information, and "EDAC" for all the other error notification / reporting / and management flows that handled separately from AER.
Dave Jiang
maintainerQuoted 22 lines · On 7/23/25 11:53 AM, [email protected] wrote:
Dave Jiang wrote:
[..]Quoted 12 lines
Quoted 12 lines
I don't think we want to combine the os-first and fw-first code in a single file. The separation leveraging the
Makefileconditional compile is good because fw-first doesn't usePCIEAER_CXLwhile os-first does. I thinkcore/ras.cis good as-is.
cxl_aer.cis in thedrivers/pci/pciedirectory and is logically part of the AER service driver. This file includes the kfifo and logic for offloading CXL RAS errors to the CXL driver. I think thecxl_aer.cnaming is fine but understand it could be argued to usecxl_ras.cinstead. My concern was in the CXL driver AER-RAS naming but the same rules could be applied here as well.Quoted 7 lines
Quoted 5 lines
To improve the original situation mentioned I recommend we use 'ras' in what is named
pci_aer.chere? Maybe 'pci_ras.c' instead? It's not perfect but would be more precise. This feels somewhat nitpicky on my part and will not be an issue if we leave as-is. Thanks for considering.Thanks for the detailed explanation. I'll rename it to
pci_ras.c.
Quoted 8 lines
Hey guys, please trim your responses.
Between RAS, AER, and EDAC there is significant overlap. You could say that
pci_ras.cis also misnamed because the error details being reported are CXL not PCI. I am ok with "RAS" for anything that handles decorating an AER notification with extra information, and "EDAC" for all the other error notification / reporting / and management flows that handled separately from AER.
I wonder if we should rename ras.c to cper.c and move the new code to ras.c.... or cxl_ras.c.
Dave Jiang wrote:
[..]
Quoted 8 lines
Quoted 6 lines
Between RAS, AER, and EDAC there is significant overlap. You could say that
pci_ras.cis also misnamed because the error details being reported are CXL not PCI. I am ok with "RAS" for anything that handles decorating an AER notification with extra information, and "EDAC" for all the other error notification / reporting / and management flows that handled separately from AER.I wonder if we should rename
ras.ctocper.cand move the new code toras.c.... orcxl_ras.c.
In the protocol error thread my ask was to just colocate all this error handling code in ras.c [1]. No real value I can see to do fine grained separation of these concerns. CPER notifications will eventually want to trigger a cxl_port walk, so keep it all together unless someone can point to a good reason to split them up.
Quoted 39 lines · Dave Jiang
Remove from
core/pci_aer.cand move theCONFIG_PCIEAER_CXLifdef tocxlpci.hto deal with the exported symbols as needed. There is enough AER handling code (and more to come) to move the AER code to its own C file.Cc Robert Richter <[email protected]>Cc Terry Bowman <[email protected]>Signed-off-by Dave Jiang <[email protected]>---
v2:
- Address 0-day reprot ofcxl_handle_rdport_errorssymbol undefined--- drivers/cxl/core/Makefile | 1 + drivers/cxl/core/core.h | 9 ++ drivers/cxl/core/pci.c | 171 +------------------------------------ drivers/cxl/core/pci_aer.c | 168 ++++++++++++++++++++++++++++++++++++ drivers/cxl/cxl.h | 8 -- drivers/cxl/cxlpci.h | 8 ++ tools/testing/cxl/Kbuild | 1 + 7 files changed, 189 insertions(+), 177 deletions(-) create mode 100644 drivers/cxl/core/pci_aer.cdrivers/cxl/core/Makefile +1
-0@@ -21,3 +21,4 @@ cxl_core-$(CONFIG_CXL_REGION) += region.o cxl_core-$(CONFIG_CXL_MCE) += mce.o cxl_core-$(CONFIG_CXL_FEATURES) += features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += pci_aer.odrivers/cxl/core/core.h +0
-0@@ -136,4 +136,13 @@ int cxl_set_feature(struct cxl_mailbox *cxl_mbox, const uuid_t *feat_uuid, u16 *return_code);#endif
drivers/cxl/core/core.h int cxl_set_feature(struct cxl_mailbox *cxl_mbox, const uuid_t *feat_uuid, · line 148
Quoted 10 lines
drivers/cxl/core/core.h +9
-0+#ifdef CONFIG_PCIEAER_CXL +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds); +#else +static inline void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) {} +#endif + +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +#endif /* __CXL_CORE_H__ */
Hi Dave,
Any reason not to pull in cxl_handle_cor_ras()/cxl_handle_ras() into pci_ras.c? These 2 functions read the RAS from the register block and then call the trace routine. These are only called by the handlers in pci.c. I think it might round-out your changes nicely if they are moved into pci_ras.c s.t. when CONFIG_PCIEAER_CXL is disabled then they are do nothing. Or another way to look at it is they should only be enabled if CONFIG_PCIEAER_CXL is enabled.
-Terry
[snip]
Dave Jiang
maintainerQuoted 53 lines · Terry
Quoted 39 lines · Dave Jiang
Remove from
core/pci_aer.cand move theCONFIG_PCIEAER_CXLifdef tocxlpci.hto deal with the exported symbols as needed. There is enough AER handling code (and more to come) to move the AER code to its own C file.Cc Robert Richter <[email protected]>Cc Terry Bowman <[email protected]>Signed-off-by Dave Jiang <[email protected]>---
v2:
- Address 0-day reprot ofcxl_handle_rdport_errorssymbol undefined--- drivers/cxl/core/Makefile | 1 + drivers/cxl/core/core.h | 9 ++ drivers/cxl/core/pci.c | 171 +------------------------------------ drivers/cxl/core/pci_aer.c | 168 ++++++++++++++++++++++++++++++++++++ drivers/cxl/cxl.h | 8 -- drivers/cxl/cxlpci.h | 8 ++ tools/testing/cxl/Kbuild | 1 + 7 files changed, 189 insertions(+), 177 deletions(-) create mode 100644 drivers/cxl/core/pci_aer.cdrivers/cxl/core/Makefile +1
-0@@ -21,3 +21,4 @@ cxl_core-$(CONFIG_CXL_REGION) += region.o cxl_core-$(CONFIG_CXL_MCE) += mce.o cxl_core-$(CONFIG_CXL_FEATURES) += features.o cxl_core-$(CONFIG_CXL_EDAC_MEM_FEATURES) += edac.o +cxl_core-$(CONFIG_PCIEAER_CXL) += pci_aer.odrivers/cxl/core/core.h +0
-0@@ -136,4 +136,13 @@ int cxl_set_feature(struct cxl_mailbox *cxl_mbox, const uuid_t *feat_uuid, u16 *return_code);#endif drivers/cxl/core/core.h int cxl_set_feature(struct cxl_mailbox *cxl_mbox, const uuid_t *feat_uuid, · line 148
Quoted 10 lines
drivers/cxl/core/core.h +9
-0+#ifdef CONFIG_PCIEAER_CXL +void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds); +#else +static inline void cxl_handle_rdport_errors (struct cxl_dev_state *cxlds) {} +#endif + +void __cxl_handle_cor_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +bool __cxl_handle_ras (struct cxl_dev_state *cxlds,void __iomem *ras_base); +#endif /* __CXL_CORE_H__ */ Hi Dave,
Quoted 7 lines
Any reason not to pull in cxl_handle_cor_ras()/cxl_handle_ras() into
pci_ras.c? These 2 functions read the RAS from the register block and then call the trace routine. These are only called by the handlers inpci.c. I think it might round-out your changes nicely if they are moved intopci_ras.cs.t. whenCONFIG_PCIEAER_CXLis disabled then they are do nothing. Or another way to look at it is they should only be enabled ifCONFIG_PCIEAER_CXLis enabled.
Missed them. Yes I can do that.
Quoted 5 lines
-Terry
[snip]