| Message ID | 20240226222704.1079449-3-Benjamin.Cheatham@amd.com |
|---|---|
| State | New |
| Headers | show |
| Series | cxl, EINJ: Update EINJ for CXL error types | expand |
> Remove CXL protocol error types from the EINJ module and move them to > a new einj_cxl module. The einj_cxl module implements the necessary > +config ACPI_APEI_EINJ_CXL > + bool "CXL Error INJection Support" It's not really a module anymore. Need to update commit messages. -Tony
This patch had an outdated commit message, so here's the patch with an updated description. I also realized that I was wrong about letting CXL 2.0+ error types (discussed a revision or two ago) and I wasn't actually letting them through. I've went ahead and added the ability to inject CXL 2.0+ error through the legacy interface. This pretty much amounts to returning an error for CXL 1.0/1.1 injection types in einj_error_inject() and instead routing them through a new einj_cxl_rch_error_inject() function called in einj-cxl.c If this change is too big I can send in another revision, I just wanted to avoid spamming the list(s).
Ben Cheatham wrote: > This patch had an outdated commit message, so here's the patch with an updated description. > > I also realized that I was wrong about letting CXL 2.0+ error types (discussed a revision > or two ago) and I wasn't actually letting them through. I've went ahead and added > the ability to inject CXL 2.0+ error through the legacy interface. This pretty > much amounts to returning an error for CXL 1.0/1.1 injection types in einj_error_inject() > and instead routing them through a new einj_cxl_rch_error_inject() function called > in einj-cxl.c > > If this change is too big I can send in another revision, I just wanted to avoid > spamming the list(s). > > From eea1cf991dc2a551f6db2e3bb9510ed43c86762d Mon Sep 17 00:00:00 2001 > From: Ben Cheatham <Benjamin.Cheatham@amd.com> > Date: Fri, 16 Feb 2024 11:12:51 -0600 > Subject: [PATCH v14 2/4] EINJ: Add CXL error type support > > Move CXL protocol error types from einj.c (now einj-core.c) to einj-cxl.c. > einj-cxl.c implements the necessary handling for CXL protocol error > injection and exposes an API for the CXL core to use said functionality, > while also allowing the EINJ module to be built without CXL support. > Because CXL error types targeting CXL 1.0/1.1 ports require special > handling, only allow them to be injected through the new cxl debugfs > interface (next commit) and return an error when attempting to inject > through the legacy interface. > > Signed-off-by: Ben Cheatham <Benjamin.Cheatham@amd.com> > --- > MAINTAINERS | 1 + > drivers/acpi/apei/Kconfig | 12 +++ > drivers/acpi/apei/Makefile | 2 + > drivers/acpi/apei/apei-internal.h | 18 ++++ > drivers/acpi/apei/{einj.c => einj-core.c} | 85 +++++++++++---- > drivers/acpi/apei/einj-cxl.c | 121 ++++++++++++++++++++++ > include/linux/einj-cxl.h | 40 +++++++ > 7 files changed, 257 insertions(+), 22 deletions(-) > rename drivers/acpi/apei/{einj.c => einj-core.c} (94%) > create mode 100644 drivers/acpi/apei/einj-cxl.c > create mode 100644 include/linux/einj-cxl.h > > diff --git a/MAINTAINERS b/MAINTAINERS > index 2ecaaec6a6bf..90cf8403dd17 100644 > --- a/MAINTAINERS > +++ b/MAINTAINERS > @@ -5289,6 +5289,7 @@ M: Dan Williams <dan.j.williams@intel.com> > L: linux-cxl@vger.kernel.org > S: Maintained > F: drivers/cxl/ > +F: include/linux/cxl-einj.h > F: include/linux/cxl-event.h > F: include/uapi/linux/cxl_mem.h > F: tools/testing/cxl/ > diff --git a/drivers/acpi/apei/Kconfig b/drivers/acpi/apei/Kconfig > index 6b18f8bc7be3..f01afa2805be 100644 > --- a/drivers/acpi/apei/Kconfig > +++ b/drivers/acpi/apei/Kconfig > @@ -60,6 +60,18 @@ config ACPI_APEI_EINJ > mainly used for debugging and testing the other parts of > APEI and some other RAS features. > > +config ACPI_APEI_EINJ_CXL > + bool "CXL Error INJection Support" > + default ACPI_APEI_EINJ > + depends on ACPI_APEI_EINJ && CXL_BUS <= ACPI_APEI_EINJ > + help > + Support for CXL protocol Error INJection through debugfs/cxl. > + Availability and which errors are supported is dependent on > + the host platform. Look to ACPI v6.5 section 18.6.4 and kernel > + EINJ documentation for more information. > + > + If unsure say 'n' > + > config ACPI_APEI_ERST_DEBUG > tristate "APEI Error Record Serialization Table (ERST) Debug Support" > depends on ACPI_APEI > diff --git a/drivers/acpi/apei/Makefile b/drivers/acpi/apei/Makefile > index 4dfac2128737..2c474e6477e1 100644 > --- a/drivers/acpi/apei/Makefile > +++ b/drivers/acpi/apei/Makefile > @@ -2,6 +2,8 @@ > obj-$(CONFIG_ACPI_APEI) += apei.o > obj-$(CONFIG_ACPI_APEI_GHES) += ghes.o > obj-$(CONFIG_ACPI_APEI_EINJ) += einj.o > +einj-y := einj-core.o > +einj-$(CONFIG_ACPI_APEI_EINJ_CXL) += einj-cxl.o > obj-$(CONFIG_ACPI_APEI_ERST_DEBUG) += erst-dbg.o > > apei-y := apei-base.o hest.o erst.o bert.o > diff --git a/drivers/acpi/apei/apei-internal.h b/drivers/acpi/apei/apei-internal.h > index 67c2c3b959e1..cd2766c69d78 100644 > --- a/drivers/acpi/apei/apei-internal.h > +++ b/drivers/acpi/apei/apei-internal.h > @@ -130,4 +130,22 @@ static inline u32 cper_estatus_len(struct acpi_hest_generic_status *estatus) > } > > int apei_osc_setup(void); > + > +int einj_get_available_error_type(u32 *type); > +int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, u64 param3, > + u64 param4); > +int einj_cxl_rch_error_inject(u32 type, u32 flags, u64 param1, u64 param2, > + u64 param3, u64 param4); > +bool einj_is_cxl_error_type(u64 type); > +int einj_validate_error_type(u64 type); > + > +#ifndef ACPI_EINJ_CXL_CACHE_CORRECTABLE > +#define ACPI_EINJ_CXL_CACHE_CORRECTABLE BIT(12) > +#define ACPI_EINJ_CXL_CACHE_UNCORRECTABLE BIT(13) > +#define ACPI_EINJ_CXL_CACHE_FATAL BIT(14) > +#define ACPI_EINJ_CXL_MEM_CORRECTABLE BIT(15) > +#define ACPI_EINJ_CXL_MEM_UNCORRECTABLE BIT(16) > +#define ACPI_EINJ_CXL_MEM_FATAL BIT(17) > +#endif > + > #endif > diff --git a/drivers/acpi/apei/einj.c b/drivers/acpi/apei/einj-core.c > similarity index 94% > rename from drivers/acpi/apei/einj.c > rename to drivers/acpi/apei/einj-core.c > index 937c69844dac..437c13949be7 100644 > --- a/drivers/acpi/apei/einj.c > +++ b/drivers/acpi/apei/einj-core.c > @@ -37,6 +37,12 @@ > #define MEM_ERROR_MASK (ACPI_EINJ_MEMORY_CORRECTABLE | \ > ACPI_EINJ_MEMORY_UNCORRECTABLE | \ > ACPI_EINJ_MEMORY_FATAL) > +#define CXL_ERROR_MASK (ACPI_EINJ_CXL_CACHE_CORRECTABLE | \ > + ACPI_EINJ_CXL_CACHE_UNCORRECTABLE | \ > + ACPI_EINJ_CXL_CACHE_FATAL | \ > + ACPI_EINJ_CXL_MEM_CORRECTABLE | \ > + ACPI_EINJ_CXL_MEM_UNCORRECTABLE | \ > + ACPI_EINJ_CXL_MEM_FATAL) > > /* > * ACPI version 5 provides a SET_ERROR_TYPE_WITH_ADDRESS action. > @@ -141,7 +147,7 @@ static DEFINE_MUTEX(einj_mutex); > /* > * Exported APIs use this flag to exit early if einj_probe() failed. > */ > -static bool einj_initialized __ro_after_init; > +bool einj_initialized __ro_after_init; > > static void *einj_param; > > @@ -166,7 +172,7 @@ static int __einj_get_available_error_type(u32 *type) > } > > /* Get error injection capabilities of the platform */ > -static int einj_get_available_error_type(u32 *type) > +int einj_get_available_error_type(u32 *type) > { > int rc; > > @@ -536,8 +542,8 @@ static int __einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, > } > > /* Inject the specified hardware error */ > -static int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, > - u64 param3, u64 param4) > +int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, u64 param3, > + u64 param4) > { > int rc; > u64 base_addr, size; > @@ -560,8 +566,18 @@ static int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, > if (type & ACPI5_VENDOR_BIT) { > if (vendor_flags != SETWA_FLAGS_MEM) > goto inject; > - } else if (!(type & MEM_ERROR_MASK) && !(flags & SETWA_FLAGS_MEM)) > + } else if (!(type & MEM_ERROR_MASK) && !(flags & SETWA_FLAGS_MEM)) { > goto inject; > + } > + > + /* > + * Injections targeting a CXL 1.0/1.1 port have to be injected > + * from the CXL debugfs interface so that we can guarantee a > + * correct MMIO address. > + */ Given that the CXL debugfs is not present in this file I would update this comment to give a hints using local references. Something like: /* * Injections targeting a CXL 1.0/1.1 port have to be injected * via the einj_cxl_rch_error_inject() path as that does proper * input validation that passed address is an RCRB base address. */ ...that said, how does this work for the CXL 2.0 path? Does not einj_cxl_inject_error() need to use __einj_error_inject() rather than einj_error_inject() to get by this check?
Ben Cheatham wrote: > This patch had an outdated commit message, so here's the patch with an updated description. > > I also realized that I was wrong about letting CXL 2.0+ error types (discussed a revision > or two ago) and I wasn't actually letting them through. I've went ahead and added > the ability to inject CXL 2.0+ error through the legacy interface. This pretty > much amounts to returning an error for CXL 1.0/1.1 injection types in einj_error_inject() > and instead routing them through a new einj_cxl_rch_error_inject() function called > in einj-cxl.c > > If this change is too big I can send in another revision, I just wanted to avoid > spamming the list(s). The thing to avoid is sending too many versions too quickly, and making sure that upstream can reassemble the set with minimal effort. In this case now that I have a few review comments and the fact that b4 likely has no chance at assembling this correctly I would go ahead and fixup those comments for a v15.
On 2/28/24 12:00 AM, Dan Williams wrote: > Ben Cheatham wrote: >> This patch had an outdated commit message, so here's the patch with an updated description. >> >> I also realized that I was wrong about letting CXL 2.0+ error types (discussed a revision >> or two ago) and I wasn't actually letting them through. I've went ahead and added >> the ability to inject CXL 2.0+ error through the legacy interface. This pretty >> much amounts to returning an error for CXL 1.0/1.1 injection types in einj_error_inject() >> and instead routing them through a new einj_cxl_rch_error_inject() function called >> in einj-cxl.c >> >> If this change is too big I can send in another revision, I just wanted to avoid >> spamming the list(s). >> >> From eea1cf991dc2a551f6db2e3bb9510ed43c86762d Mon Sep 17 00:00:00 2001 >> From: Ben Cheatham <Benjamin.Cheatham@amd.com> >> Date: Fri, 16 Feb 2024 11:12:51 -0600 >> Subject: [PATCH v14 2/4] EINJ: Add CXL error type support >> >> Move CXL protocol error types from einj.c (now einj-core.c) to einj-cxl.c. >> einj-cxl.c implements the necessary handling for CXL protocol error >> injection and exposes an API for the CXL core to use said functionality, >> while also allowing the EINJ module to be built without CXL support. >> Because CXL error types targeting CXL 1.0/1.1 ports require special >> handling, only allow them to be injected through the new cxl debugfs >> interface (next commit) and return an error when attempting to inject >> through the legacy interface. >> >> Signed-off-by: Ben Cheatham <Benjamin.Cheatham@amd.com> >> --- >> MAINTAINERS | 1 + >> drivers/acpi/apei/Kconfig | 12 +++ >> drivers/acpi/apei/Makefile | 2 + >> drivers/acpi/apei/apei-internal.h | 18 ++++ >> drivers/acpi/apei/{einj.c => einj-core.c} | 85 +++++++++++---- >> drivers/acpi/apei/einj-cxl.c | 121 ++++++++++++++++++++++ >> include/linux/einj-cxl.h | 40 +++++++ >> 7 files changed, 257 insertions(+), 22 deletions(-) >> rename drivers/acpi/apei/{einj.c => einj-core.c} (94%) >> create mode 100644 drivers/acpi/apei/einj-cxl.c >> create mode 100644 include/linux/einj-cxl.h >> >> diff --git a/MAINTAINERS b/MAINTAINERS >> index 2ecaaec6a6bf..90cf8403dd17 100644 >> --- a/MAINTAINERS >> +++ b/MAINTAINERS >> @@ -5289,6 +5289,7 @@ M: Dan Williams <dan.j.williams@intel.com> >> L: linux-cxl@vger.kernel.org >> S: Maintained >> F: drivers/cxl/ >> +F: include/linux/cxl-einj.h >> F: include/linux/cxl-event.h >> F: include/uapi/linux/cxl_mem.h >> F: tools/testing/cxl/ >> diff --git a/drivers/acpi/apei/Kconfig b/drivers/acpi/apei/Kconfig >> index 6b18f8bc7be3..f01afa2805be 100644 >> --- a/drivers/acpi/apei/Kconfig >> +++ b/drivers/acpi/apei/Kconfig >> @@ -60,6 +60,18 @@ config ACPI_APEI_EINJ >> mainly used for debugging and testing the other parts of >> APEI and some other RAS features. >> >> +config ACPI_APEI_EINJ_CXL >> + bool "CXL Error INJection Support" >> + default ACPI_APEI_EINJ >> + depends on ACPI_APEI_EINJ && CXL_BUS <= ACPI_APEI_EINJ >> + help >> + Support for CXL protocol Error INJection through debugfs/cxl. >> + Availability and which errors are supported is dependent on >> + the host platform. Look to ACPI v6.5 section 18.6.4 and kernel >> + EINJ documentation for more information. >> + >> + If unsure say 'n' >> + >> config ACPI_APEI_ERST_DEBUG >> tristate "APEI Error Record Serialization Table (ERST) Debug Support" >> depends on ACPI_APEI >> diff --git a/drivers/acpi/apei/Makefile b/drivers/acpi/apei/Makefile >> index 4dfac2128737..2c474e6477e1 100644 >> --- a/drivers/acpi/apei/Makefile >> +++ b/drivers/acpi/apei/Makefile >> @@ -2,6 +2,8 @@ >> obj-$(CONFIG_ACPI_APEI) += apei.o >> obj-$(CONFIG_ACPI_APEI_GHES) += ghes.o >> obj-$(CONFIG_ACPI_APEI_EINJ) += einj.o >> +einj-y := einj-core.o >> +einj-$(CONFIG_ACPI_APEI_EINJ_CXL) += einj-cxl.o >> obj-$(CONFIG_ACPI_APEI_ERST_DEBUG) += erst-dbg.o >> >> apei-y := apei-base.o hest.o erst.o bert.o >> diff --git a/drivers/acpi/apei/apei-internal.h b/drivers/acpi/apei/apei-internal.h >> index 67c2c3b959e1..cd2766c69d78 100644 >> --- a/drivers/acpi/apei/apei-internal.h >> +++ b/drivers/acpi/apei/apei-internal.h >> @@ -130,4 +130,22 @@ static inline u32 cper_estatus_len(struct acpi_hest_generic_status *estatus) >> } >> >> int apei_osc_setup(void); >> + >> +int einj_get_available_error_type(u32 *type); >> +int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, u64 param3, >> + u64 param4); >> +int einj_cxl_rch_error_inject(u32 type, u32 flags, u64 param1, u64 param2, >> + u64 param3, u64 param4); >> +bool einj_is_cxl_error_type(u64 type); >> +int einj_validate_error_type(u64 type); >> + >> +#ifndef ACPI_EINJ_CXL_CACHE_CORRECTABLE >> +#define ACPI_EINJ_CXL_CACHE_CORRECTABLE BIT(12) >> +#define ACPI_EINJ_CXL_CACHE_UNCORRECTABLE BIT(13) >> +#define ACPI_EINJ_CXL_CACHE_FATAL BIT(14) >> +#define ACPI_EINJ_CXL_MEM_CORRECTABLE BIT(15) >> +#define ACPI_EINJ_CXL_MEM_UNCORRECTABLE BIT(16) >> +#define ACPI_EINJ_CXL_MEM_FATAL BIT(17) >> +#endif >> + >> #endif >> diff --git a/drivers/acpi/apei/einj.c b/drivers/acpi/apei/einj-core.c >> similarity index 94% >> rename from drivers/acpi/apei/einj.c >> rename to drivers/acpi/apei/einj-core.c >> index 937c69844dac..437c13949be7 100644 >> --- a/drivers/acpi/apei/einj.c >> +++ b/drivers/acpi/apei/einj-core.c >> @@ -37,6 +37,12 @@ >> #define MEM_ERROR_MASK (ACPI_EINJ_MEMORY_CORRECTABLE | \ >> ACPI_EINJ_MEMORY_UNCORRECTABLE | \ >> ACPI_EINJ_MEMORY_FATAL) >> +#define CXL_ERROR_MASK (ACPI_EINJ_CXL_CACHE_CORRECTABLE | \ >> + ACPI_EINJ_CXL_CACHE_UNCORRECTABLE | \ >> + ACPI_EINJ_CXL_CACHE_FATAL | \ >> + ACPI_EINJ_CXL_MEM_CORRECTABLE | \ >> + ACPI_EINJ_CXL_MEM_UNCORRECTABLE | \ >> + ACPI_EINJ_CXL_MEM_FATAL) >> >> /* >> * ACPI version 5 provides a SET_ERROR_TYPE_WITH_ADDRESS action. >> @@ -141,7 +147,7 @@ static DEFINE_MUTEX(einj_mutex); >> /* >> * Exported APIs use this flag to exit early if einj_probe() failed. >> */ >> -static bool einj_initialized __ro_after_init; >> +bool einj_initialized __ro_after_init; >> >> static void *einj_param; >> >> @@ -166,7 +172,7 @@ static int __einj_get_available_error_type(u32 *type) >> } >> >> /* Get error injection capabilities of the platform */ >> -static int einj_get_available_error_type(u32 *type) >> +int einj_get_available_error_type(u32 *type) >> { >> int rc; >> >> @@ -536,8 +542,8 @@ static int __einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, >> } >> >> /* Inject the specified hardware error */ >> -static int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, >> - u64 param3, u64 param4) >> +int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, u64 param3, >> + u64 param4) >> { >> int rc; >> u64 base_addr, size; >> @@ -560,8 +566,18 @@ static int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, >> if (type & ACPI5_VENDOR_BIT) { >> if (vendor_flags != SETWA_FLAGS_MEM) >> goto inject; >> - } else if (!(type & MEM_ERROR_MASK) && !(flags & SETWA_FLAGS_MEM)) >> + } else if (!(type & MEM_ERROR_MASK) && !(flags & SETWA_FLAGS_MEM)) { >> goto inject; >> + } >> + >> + /* >> + * Injections targeting a CXL 1.0/1.1 port have to be injected >> + * from the CXL debugfs interface so that we can guarantee a >> + * correct MMIO address. >> + */ > > Given that the CXL debugfs is not present in this file I would update > this comment to give a hints using local references. Something like: > > /* > * Injections targeting a CXL 1.0/1.1 port have to be injected > * via the einj_cxl_rch_error_inject() path as that does proper > * input validation that passed address is an RCRB base address. > */ > I agree, I'll change it to something along those lines. > ...that said, how does this work for the CXL 2.0 path? Does not > einj_cxl_inject_error() need to use __einj_error_inject() rather than > einj_error_inject() to get by this check? > The way that the target port type (1.0 vs. 2.0+) is differentiated is whether we are targeting a memory address, which is what SETWA_FLAGS_MEM indicates. So for the CXL 2.0+ path, the check will pass. Although the error type is a CXL error type, the flags variable will be set to SETWA_FLAGS_SBDF (I'm 90% sure that's what it's called, but it's something along those lines). Let me know if that doesn't answer your question!
On 2/28/24 12:04 AM, Dan Williams wrote: > Ben Cheatham wrote: >> This patch had an outdated commit message, so here's the patch with an updated description. >> >> I also realized that I was wrong about letting CXL 2.0+ error types (discussed a revision >> or two ago) and I wasn't actually letting them through. I've went ahead and added >> the ability to inject CXL 2.0+ error through the legacy interface. This pretty >> much amounts to returning an error for CXL 1.0/1.1 injection types in einj_error_inject() >> and instead routing them through a new einj_cxl_rch_error_inject() function called >> in einj-cxl.c >> >> If this change is too big I can send in another revision, I just wanted to avoid >> spamming the list(s). > > The thing to avoid is sending too many versions too quickly, and making > sure that upstream can reassemble the set with minimal effort. In this > case now that I have a few review comments and the fact that b4 likely > has no chance at assembling this correctly I would go ahead and fixup > those comments for a v15. Alright that makes sense. I'll put together a v15 and send it out later this week/early next week. Thanks, Ben
On Mon, 26 Feb 2024 16:27:02 -0600 Ben Cheatham <Benjamin.Cheatham@amd.com> wrote: > Remove CXL protocol error types from the EINJ module and move them to > a new einj_cxl module. The einj_cxl module implements the necessary > handling for CXL protocol error injection and exposes an API for the > CXL core to use said functionality. Because the CXL error types > require special handling, only allow them to be injected through the > einj_cxl module and return an error when attempting to inject through > "regular" EINJ. > > Signed-off-by: Ben Cheatham <Benjamin.Cheatham@amd.com> Hi Ben, Some minor comments inline given you are doing a v15 (yikes!) Jonathan > diff --git a/drivers/acpi/apei/einj.c b/drivers/acpi/apei/einj-core.c > similarity index 94% > rename from drivers/acpi/apei/einj.c > rename to drivers/acpi/apei/einj-core.c > index 937c69844dac..1a5f53d81d09 100644 > --- a/drivers/acpi/apei/einj.c > +++ b/drivers/acpi/apei/einj-core.c ... > @@ -640,29 +648,21 @@ static int available_error_type_show(struct seq_file *m, void *v) > > DEFINE_SHOW_ATTRIBUTE(available_error_type); > > -static int error_type_get(void *data, u64 *val) > -{ > - *val = error_type; > - > - return 0; > -} > - > -static int error_type_set(void *data, u64 val) > +int einj_validate_error_type(u64 type) > { > + u32 tval, vendor, available_error_type = 0; > int rc; > - u32 available_error_type = 0; > - u32 tval, vendor; > > /* Only low 32 bits for error type are valid */ > - if (val & GENMASK_ULL(63, 32)) > + if (type & GENMASK_ULL(63, 32)) > return -EINVAL; > > /* > * Vendor defined types have 0x80000000 bit set, and > * are not enumerated by ACPI_EINJ_GET_ERROR_TYPE > */ > - vendor = val & ACPI5_VENDOR_BIT; > - tval = val & 0x7fffffff; > + vendor = type & ACPI5_VENDOR_BIT; > + tval = type & GENMASK(30, 0); > > /* Only one error type can be specified */ > if (tval & (tval - 1)) > @@ -671,9 +671,37 @@ static int error_type_set(void *data, u64 val) > rc = einj_get_available_error_type(&available_error_type); > if (rc) > return rc; > - if (!(val & available_error_type)) > + if (!(type & available_error_type)) > return -EINVAL; > } > + > + return 0; > +} > + > +bool einj_is_cxl_error_type(u64 type) > +{ > + return (type & CXL_ERROR_MASK) && (!(type & ACPI5_VENDOR_BIT)); > +} > + > +static int error_type_get(void *data, u64 *val) This is reordered, but fairly sure no need to do so and it will make patch cleaner to leave it above the validation code. > +{ > + *val = error_type; > + > + return 0; > +} > + > +static int error_type_set(void *data, u64 val) > +{ > + int rc; > + > + /* CXL error types have to be injected from cxl debugfs */ > + if (einj_is_cxl_error_type(val)) > + return -EINVAL; > + > + rc = einj_validate_error_type(val); Trivial but I'd have preferred this factoring out in a precursor patch just to make reviewing this a tiny bit easier. > + if (rc) > + return rc; > + > error_type = val; > > return 0; > diff --git a/drivers/acpi/apei/einj-cxl.c b/drivers/acpi/apei/einj-cxl.c > new file mode 100644 > index 000000000000..34badc6a801e > --- /dev/null > +++ b/drivers/acpi/apei/einj-cxl.c > @@ -0,0 +1,120 @@ > +// SPDX-License-Identifier: GPL-2.0-only > +/* > + * CXL Error INJection support. Used by CXL core to inject > + * protocol errors into CXL ports. > + * > + * Copyright (C) 2023 Advanced Micro Devices, Inc. > + * > + * Author: Ben Cheatham <benjamin.cheatham@amd.com> > + */ > +#include <linux/einj-cxl.h> > +#include <linux/debugfs.h> Follow include what you use principles a little closer (there are always exceptions in kernel world...). The following seems resonable to me. #include <linux/array_size.h> #include <linux/seq_file.h> > + > +#include "apei-internal.h" > + > +/* Defined in einj-core.c */ > +extern bool einj_initialized; > + > +static struct { u32 mask; const char *str; } const einj_cxl_error_type_string[] = { > + { BIT(12), "CXL.cache Protocol Correctable" }, Use the defines for the bits? Not sure why the original code didn't do so other than maybe long line lengths? > + { BIT(13), "CXL.cache Protocol Uncorrectable non-fatal" }, > + { BIT(14), "CXL.cache Protocol Uncorrectable fatal" }, > + { BIT(15), "CXL.mem Protocol Correctable" }, > + { BIT(16), "CXL.mem Protocol Uncorrectable non-fatal" }, > + { BIT(17), "CXL.mem Protocol Uncorrectable fatal" }, > +}; > + > +int einj_cxl_available_error_type_show(struct seq_file *m, void *v) > +{ > + int cxl_err, rc; > + u32 available_error_type = 0; > + > + if (!einj_initialized) > + return -ENXIO; > + > + rc = einj_get_available_error_type(&available_error_type); > + if (rc) > + return rc; > + > + for (int pos = 0; pos < ARRAY_SIZE(einj_cxl_error_type_string); pos++) { > + cxl_err = ACPI_EINJ_CXL_CACHE_CORRECTABLE << pos; Hmm. This is a little ugly. Could do something like the following bit it's of similar level of ugly so up to you. int bit_pos = ACPI_EINJ_CXL_CACHE_CORRECTABLE; for_each_bit_set_bit_from(bit_pos, &available_error_type, ARRAY_SIZE(einj_cxl_error_type_string)) { int pos = bit_pos - ACPI_EINJ_CXL_CACHE_CORRECTABLE; > + > + if (available_error_type & cxl_err) > + seq_printf(m, "0x%08x\t%s\n", > + einj_cxl_error_type_string[pos].mask, > + einj_cxl_error_type_string[pos].str); > + } > + > + return 0; > +} > +EXPORT_SYMBOL_NS_GPL(einj_cxl_available_error_type_show, CXL); > diff --git a/include/linux/einj-cxl.h b/include/linux/einj-cxl.h > new file mode 100644 > index 000000000000..4a1f4600539a > --- /dev/null > +++ b/include/linux/einj-cxl.h > @@ -0,0 +1,40 @@ > +/* SPDX-License-Identifier: GPL-2.0-or-later */ > +/* > + * CXL protocol Error INJection support. > + * > + * Copyright (c) 2023 Advanced Micro Devices, Inc. > + * All Rights Reserved. > + * > + * Author: Ben Cheatham <benjamin.cheatham@amd.com> > + */ > +#ifndef EINJ_CXL_H > +#define EINJ_CXL_H > + > +#include <linux/pci.h> Use a forwards def struct pci_dev; and drop the include. Also need struct seq_file; > + > +#if IS_ENABLED(CONFIG_ACPI_APEI_EINJ_CXL) > +int einj_cxl_available_error_type_show(struct seq_file *m, void *v); > +int einj_cxl_inject_error(struct pci_dev *dport_dev, u64 type); > +int einj_cxl_inject_rch_error(u64 rcrb, u64 type); > +bool einj_cxl_is_initialized(void); > +#else /* !IS_ENABLED(CONFIG_ACPI_APEI_EINJ_CXL) */ > +static inline int einj_cxl_available_error_type_show(struct seq_file *m, > + void *v) > +{ > + return -ENXIO; > +} > + > +static inline int einj_cxl_inject_error(struct pci_dev *dport_dev, u64 type) > +{ > + return -ENXIO; > +} > + > +static inline int einj_cxl_inject_rch_error(u64 rcrb, u64 type) > +{ > + return -ENXIO; > +} > + > +static inline bool einj_cxl_is_initialized(void) { return false; } > +#endif /* CONFIG_ACPI_APEI_EINJ_CXL */ > + > +#endif /* EINJ_CXL_H */
Hey Jonathan, thanks for taking a look! On 3/7/24 6:09 AM, Jonathan Cameron wrote: > On Mon, 26 Feb 2024 16:27:02 -0600 > Ben Cheatham <Benjamin.Cheatham@amd.com> wrote: > >> Remove CXL protocol error types from the EINJ module and move them to >> a new einj_cxl module. The einj_cxl module implements the necessary >> handling for CXL protocol error injection and exposes an API for the >> CXL core to use said functionality. Because the CXL error types >> require special handling, only allow them to be injected through the >> einj_cxl module and return an error when attempting to inject through >> "regular" EINJ. >> >> Signed-off-by: Ben Cheatham <Benjamin.Cheatham@amd.com> > Hi Ben, > > Some minor comments inline given you are doing a v15 (yikes!) > Yeah I know :(. > Jonathan > >> diff --git a/drivers/acpi/apei/einj.c b/drivers/acpi/apei/einj-core.c >> similarity index 94% >> rename from drivers/acpi/apei/einj.c >> rename to drivers/acpi/apei/einj-core.c >> index 937c69844dac..1a5f53d81d09 100644 >> --- a/drivers/acpi/apei/einj.c >> +++ b/drivers/acpi/apei/einj-core.c > > ... > >> @@ -640,29 +648,21 @@ static int available_error_type_show(struct seq_file *m, void *v) >> >> DEFINE_SHOW_ATTRIBUTE(available_error_type); >> >> -static int error_type_get(void *data, u64 *val) >> -{ >> - *val = error_type; >> - >> - return 0; >> -} >> - >> -static int error_type_set(void *data, u64 val) >> +int einj_validate_error_type(u64 type) >> { >> + u32 tval, vendor, available_error_type = 0; >> int rc; >> - u32 available_error_type = 0; >> - u32 tval, vendor; >> >> /* Only low 32 bits for error type are valid */ >> - if (val & GENMASK_ULL(63, 32)) >> + if (type & GENMASK_ULL(63, 32)) >> return -EINVAL; >> >> /* >> * Vendor defined types have 0x80000000 bit set, and >> * are not enumerated by ACPI_EINJ_GET_ERROR_TYPE >> */ >> - vendor = val & ACPI5_VENDOR_BIT; >> - tval = val & 0x7fffffff; >> + vendor = type & ACPI5_VENDOR_BIT; >> + tval = type & GENMASK(30, 0); >> >> /* Only one error type can be specified */ >> if (tval & (tval - 1)) >> @@ -671,9 +671,37 @@ static int error_type_set(void *data, u64 val) >> rc = einj_get_available_error_type(&available_error_type); >> if (rc) >> return rc; >> - if (!(val & available_error_type)) >> + if (!(type & available_error_type)) >> return -EINVAL; >> } >> + >> + return 0; >> +} >> + >> +bool einj_is_cxl_error_type(u64 type) >> +{ >> + return (type & CXL_ERROR_MASK) && (!(type & ACPI5_VENDOR_BIT)); >> +} >> + >> +static int error_type_get(void *data, u64 *val) > This is reordered, but fairly sure no need to do so and it will > make patch cleaner to leave it above the validation code. > Sorry it's been a bit, but I think I moved it to go with the other error_type functions. I don't mind leaving it where it was though. >> +{ >> + *val = error_type; >> + >> + return 0; >> +} >> + >> +static int error_type_set(void *data, u64 val) >> +{ >> + int rc; >> + >> + /* CXL error types have to be injected from cxl debugfs */ >> + if (einj_is_cxl_error_type(val)) >> + return -EINVAL; >> + >> + rc = einj_validate_error_type(val); > > Trivial but I'd have preferred this factoring out in a precursor patch > just to make reviewing this a tiny bit easier. > It would probably make sense, but at this point I just want to get this across the finish line :). > >> + if (rc) >> + return rc; >> + >> error_type = val; >> >> return 0; >> diff --git a/drivers/acpi/apei/einj-cxl.c b/drivers/acpi/apei/einj-cxl.c >> new file mode 100644 >> index 000000000000..34badc6a801e >> --- /dev/null >> +++ b/drivers/acpi/apei/einj-cxl.c >> @@ -0,0 +1,120 @@ >> +// SPDX-License-Identifier: GPL-2.0-only >> +/* >> + * CXL Error INJection support. Used by CXL core to inject >> + * protocol errors into CXL ports. >> + * >> + * Copyright (C) 2023 Advanced Micro Devices, Inc. >> + * >> + * Author: Ben Cheatham <benjamin.cheatham@amd.com> >> + */ >> +#include <linux/einj-cxl.h> >> +#include <linux/debugfs.h> > Follow include what you use principles a little closer > (there are always exceptions in kernel world...). > The following seems resonable to me. > > #include <linux/array_size.h> > #include <linux/seq_file.h> > Will do! > >> + >> +#include "apei-internal.h" >> + >> +/* Defined in einj-core.c */ >> +extern bool einj_initialized; >> + >> +static struct { u32 mask; const char *str; } const einj_cxl_error_type_string[] = { >> + { BIT(12), "CXL.cache Protocol Correctable" }, > > Use the defines for the bits? Not sure why the original code didn't do so other > than maybe long line lengths? > I agree, this was merged from some ACPI tree changes in rc4 and I don't think they have the defines in that tree yet. > > >> + { BIT(13), "CXL.cache Protocol Uncorrectable non-fatal" }, >> + { BIT(14), "CXL.cache Protocol Uncorrectable fatal" }, >> + { BIT(15), "CXL.mem Protocol Correctable" }, >> + { BIT(16), "CXL.mem Protocol Uncorrectable non-fatal" }, >> + { BIT(17), "CXL.mem Protocol Uncorrectable fatal" }, >> +}; >> + >> +int einj_cxl_available_error_type_show(struct seq_file *m, void *v) >> +{ >> + int cxl_err, rc; >> + u32 available_error_type = 0; >> + >> + if (!einj_initialized) >> + return -ENXIO; >> + >> + rc = einj_get_available_error_type(&available_error_type); >> + if (rc) >> + return rc; >> + >> + for (int pos = 0; pos < ARRAY_SIZE(einj_cxl_error_type_string); pos++) { >> + cxl_err = ACPI_EINJ_CXL_CACHE_CORRECTABLE << pos; > Hmm. This is a little ugly. > Could do something like the following bit it's of similar level of ugly > so up to you. > > int bit_pos = ACPI_EINJ_CXL_CACHE_CORRECTABLE; > for_each_bit_set_bit_from(bit_pos, &available_error_type, > ARRAY_SIZE(einj_cxl_error_type_string)) { > int pos = bit_pos - ACPI_EINJ_CXL_CACHE_CORRECTABLE; > I agree it's ugly. I think this version has the added benfit of parity with einj_available_error_type_show() in einj-core.c, so I think it's better to keep it this way if it's the same to you. > > >> + >> + if (available_error_type & cxl_err) >> + seq_printf(m, "0x%08x\t%s\n", >> + einj_cxl_error_type_string[pos].mask, >> + einj_cxl_error_type_string[pos].str); >> + } >> + >> + return 0; >> +} >> +EXPORT_SYMBOL_NS_GPL(einj_cxl_available_error_type_show, CXL); > > >> diff --git a/include/linux/einj-cxl.h b/include/linux/einj-cxl.h >> new file mode 100644 >> index 000000000000..4a1f4600539a >> --- /dev/null >> +++ b/include/linux/einj-cxl.h >> @@ -0,0 +1,40 @@ >> +/* SPDX-License-Identifier: GPL-2.0-or-later */ >> +/* >> + * CXL protocol Error INJection support. >> + * >> + * Copyright (c) 2023 Advanced Micro Devices, Inc. >> + * All Rights Reserved. >> + * >> + * Author: Ben Cheatham <benjamin.cheatham@amd.com> >> + */ >> +#ifndef EINJ_CXL_H >> +#define EINJ_CXL_H >> + >> +#include <linux/pci.h> > Use a forwards def > > struct pci_dev; > > and drop the include. Also need > > struct seq_file; > Will do. > >> + >> +#if IS_ENABLED(CONFIG_ACPI_APEI_EINJ_CXL) >> +int einj_cxl_available_error_type_show(struct seq_file *m, void *v); >> +int einj_cxl_inject_error(struct pci_dev *dport_dev, u64 type); >> +int einj_cxl_inject_rch_error(u64 rcrb, u64 type); >> +bool einj_cxl_is_initialized(void); >> +#else /* !IS_ENABLED(CONFIG_ACPI_APEI_EINJ_CXL) */ >> +static inline int einj_cxl_available_error_type_show(struct seq_file *m, >> + void *v) >> +{ >> + return -ENXIO; >> +} >> + >> +static inline int einj_cxl_inject_error(struct pci_dev *dport_dev, u64 type) >> +{ >> + return -ENXIO; >> +} >> + >> +static inline int einj_cxl_inject_rch_error(u64 rcrb, u64 type) >> +{ >> + return -ENXIO; >> +} >> + >> +static inline bool einj_cxl_is_initialized(void) { return false; } >> +#endif /* CONFIG_ACPI_APEI_EINJ_CXL */ >> + >> +#endif /* EINJ_CXL_H */ >
On Thu, 7 Mar 2024 08:46:49 -0600 Ben Cheatham <benjamin.cheatham@amd.com> wrote: > Hey Jonathan, thanks for taking a look! > > On 3/7/24 6:09 AM, Jonathan Cameron wrote: > > On Mon, 26 Feb 2024 16:27:02 -0600 > > Ben Cheatham <Benjamin.Cheatham@amd.com> wrote: > > > >> Remove CXL protocol error types from the EINJ module and move them to > >> a new einj_cxl module. The einj_cxl module implements the necessary > >> handling for CXL protocol error injection and exposes an API for the > >> CXL core to use said functionality. Because the CXL error types > >> require special handling, only allow them to be injected through the > >> einj_cxl module and return an error when attempting to inject through > >> "regular" EINJ. > >> > >> Signed-off-by: Ben Cheatham <Benjamin.Cheatham@amd.com> > > Hi Ben, > > > > Some minor comments inline given you are doing a v15 (yikes!) > > > > Yeah I know :(. > With headers tidied up. Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> > > > > > >> + { BIT(13), "CXL.cache Protocol Uncorrectable non-fatal" }, > >> + { BIT(14), "CXL.cache Protocol Uncorrectable fatal" }, > >> + { BIT(15), "CXL.mem Protocol Correctable" }, > >> + { BIT(16), "CXL.mem Protocol Uncorrectable non-fatal" }, > >> + { BIT(17), "CXL.mem Protocol Uncorrectable fatal" }, > >> +}; > >> + > >> +int einj_cxl_available_error_type_show(struct seq_file *m, void *v) > >> +{ > >> + int cxl_err, rc; > >> + u32 available_error_type = 0; > >> + > >> + if (!einj_initialized) > >> + return -ENXIO; > >> + > >> + rc = einj_get_available_error_type(&available_error_type); > >> + if (rc) > >> + return rc; > >> + > >> + for (int pos = 0; pos < ARRAY_SIZE(einj_cxl_error_type_string); pos++) { > >> + cxl_err = ACPI_EINJ_CXL_CACHE_CORRECTABLE << pos; > > Hmm. This is a little ugly. > > Could do something like the following bit it's of similar level of ugly > > so up to you. > > > > int bit_pos = ACPI_EINJ_CXL_CACHE_CORRECTABLE; > > for_each_bit_set_bit_from(bit_pos, &available_error_type, > > ARRAY_SIZE(einj_cxl_error_type_string)) { > > int pos = bit_pos - ACPI_EINJ_CXL_CACHE_CORRECTABLE; > > > > I agree it's ugly. I think this version has the added benfit of parity > with einj_available_error_type_show() in einj-core.c, so I think it's > better to keep it this way if it's the same to you. > Sure. We can always (maybe) tidy them both up later :) J
diff --git a/MAINTAINERS b/MAINTAINERS index 2ecaaec6a6bf..90cf8403dd17 100644 --- a/MAINTAINERS +++ b/MAINTAINERS @@ -5289,6 +5289,7 @@ M: Dan Williams <dan.j.williams@intel.com> L: linux-cxl@vger.kernel.org S: Maintained F: drivers/cxl/ +F: include/linux/cxl-einj.h F: include/linux/cxl-event.h F: include/uapi/linux/cxl_mem.h F: tools/testing/cxl/ diff --git a/drivers/acpi/apei/Kconfig b/drivers/acpi/apei/Kconfig index 6b18f8bc7be3..f01afa2805be 100644 --- a/drivers/acpi/apei/Kconfig +++ b/drivers/acpi/apei/Kconfig @@ -60,6 +60,18 @@ config ACPI_APEI_EINJ mainly used for debugging and testing the other parts of APEI and some other RAS features. +config ACPI_APEI_EINJ_CXL + bool "CXL Error INJection Support" + default ACPI_APEI_EINJ + depends on ACPI_APEI_EINJ && CXL_BUS <= ACPI_APEI_EINJ + help + Support for CXL protocol Error INJection through debugfs/cxl. + Availability and which errors are supported is dependent on + the host platform. Look to ACPI v6.5 section 18.6.4 and kernel + EINJ documentation for more information. + + If unsure say 'n' + config ACPI_APEI_ERST_DEBUG tristate "APEI Error Record Serialization Table (ERST) Debug Support" depends on ACPI_APEI diff --git a/drivers/acpi/apei/Makefile b/drivers/acpi/apei/Makefile index 4dfac2128737..2c474e6477e1 100644 --- a/drivers/acpi/apei/Makefile +++ b/drivers/acpi/apei/Makefile @@ -2,6 +2,8 @@ obj-$(CONFIG_ACPI_APEI) += apei.o obj-$(CONFIG_ACPI_APEI_GHES) += ghes.o obj-$(CONFIG_ACPI_APEI_EINJ) += einj.o +einj-y := einj-core.o +einj-$(CONFIG_ACPI_APEI_EINJ_CXL) += einj-cxl.o obj-$(CONFIG_ACPI_APEI_ERST_DEBUG) += erst-dbg.o apei-y := apei-base.o hest.o erst.o bert.o diff --git a/drivers/acpi/apei/apei-internal.h b/drivers/acpi/apei/apei-internal.h index 67c2c3b959e1..ab039d24cd22 100644 --- a/drivers/acpi/apei/apei-internal.h +++ b/drivers/acpi/apei/apei-internal.h @@ -130,4 +130,20 @@ static inline u32 cper_estatus_len(struct acpi_hest_generic_status *estatus) } int apei_osc_setup(void); + +int einj_get_available_error_type(u32 *type); +int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, u64 param3, + u64 param4); +bool einj_is_cxl_error_type(u64 type); +int einj_validate_error_type(u64 type); + +#ifndef ACPI_EINJ_CXL_CACHE_CORRECTABLE +#define ACPI_EINJ_CXL_CACHE_CORRECTABLE BIT(12) +#define ACPI_EINJ_CXL_CACHE_UNCORRECTABLE BIT(13) +#define ACPI_EINJ_CXL_CACHE_FATAL BIT(14) +#define ACPI_EINJ_CXL_MEM_CORRECTABLE BIT(15) +#define ACPI_EINJ_CXL_MEM_UNCORRECTABLE BIT(16) +#define ACPI_EINJ_CXL_MEM_FATAL BIT(17) +#endif + #endif diff --git a/drivers/acpi/apei/einj.c b/drivers/acpi/apei/einj-core.c similarity index 94% rename from drivers/acpi/apei/einj.c rename to drivers/acpi/apei/einj-core.c index 937c69844dac..1a5f53d81d09 100644 --- a/drivers/acpi/apei/einj.c +++ b/drivers/acpi/apei/einj-core.c @@ -37,6 +37,12 @@ #define MEM_ERROR_MASK (ACPI_EINJ_MEMORY_CORRECTABLE | \ ACPI_EINJ_MEMORY_UNCORRECTABLE | \ ACPI_EINJ_MEMORY_FATAL) +#define CXL_ERROR_MASK (ACPI_EINJ_CXL_CACHE_CORRECTABLE | \ + ACPI_EINJ_CXL_CACHE_UNCORRECTABLE | \ + ACPI_EINJ_CXL_CACHE_FATAL | \ + ACPI_EINJ_CXL_MEM_CORRECTABLE | \ + ACPI_EINJ_CXL_MEM_UNCORRECTABLE | \ + ACPI_EINJ_CXL_MEM_FATAL) /* * ACPI version 5 provides a SET_ERROR_TYPE_WITH_ADDRESS action. @@ -141,7 +147,7 @@ static DEFINE_MUTEX(einj_mutex); /* * Exported APIs use this flag to exit early if einj_probe() failed. */ -static bool einj_initialized __ro_after_init; +bool einj_initialized __ro_after_init; static void *einj_param; @@ -166,7 +172,7 @@ static int __einj_get_available_error_type(u32 *type) } /* Get error injection capabilities of the platform */ -static int einj_get_available_error_type(u32 *type) +int einj_get_available_error_type(u32 *type) { int rc; @@ -536,8 +542,8 @@ static int __einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, } /* Inject the specified hardware error */ -static int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, - u64 param3, u64 param4) +int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, u64 param3, + u64 param4) { int rc; u64 base_addr, size; @@ -560,8 +566,16 @@ static int einj_error_inject(u32 type, u32 flags, u64 param1, u64 param2, if (type & ACPI5_VENDOR_BIT) { if (vendor_flags != SETWA_FLAGS_MEM) goto inject; - } else if (!(type & MEM_ERROR_MASK) && !(flags & SETWA_FLAGS_MEM)) + } else if (!(type & MEM_ERROR_MASK) && !(flags & SETWA_FLAGS_MEM)) { goto inject; + /* + * If the error type is for a CXL 1.0/1.1 port then skip address + * checking here. The einj-cxl module should have already verified the + * supplied MMIO address is correct. + */ + } else if ((type & CXL_ERROR_MASK) && (flags & SETWA_FLAGS_MEM)) { + goto inject; + } /* * Disallow crazy address masks that give BIOS leeway to pick @@ -613,12 +627,6 @@ static struct { u32 mask; const char *str; } const einj_error_type_string[] = { { BIT(9), "Platform Correctable" }, { BIT(10), "Platform Uncorrectable non-fatal" }, { BIT(11), "Platform Uncorrectable fatal"}, - { BIT(12), "CXL.cache Protocol Correctable" }, - { BIT(13), "CXL.cache Protocol Uncorrectable non-fatal" }, - { BIT(14), "CXL.cache Protocol Uncorrectable fatal" }, - { BIT(15), "CXL.mem Protocol Correctable" }, - { BIT(16), "CXL.mem Protocol Uncorrectable non-fatal" }, - { BIT(17), "CXL.mem Protocol Uncorrectable fatal" }, { BIT(31), "Vendor Defined Error Types" }, }; @@ -640,29 +648,21 @@ static int available_error_type_show(struct seq_file *m, void *v) DEFINE_SHOW_ATTRIBUTE(available_error_type); -static int error_type_get(void *data, u64 *val) -{ - *val = error_type; - - return 0; -} - -static int error_type_set(void *data, u64 val) +int einj_validate_error_type(u64 type) { + u32 tval, vendor, available_error_type = 0; int rc; - u32 available_error_type = 0; - u32 tval, vendor; /* Only low 32 bits for error type are valid */ - if (val & GENMASK_ULL(63, 32)) + if (type & GENMASK_ULL(63, 32)) return -EINVAL; /* * Vendor defined types have 0x80000000 bit set, and * are not enumerated by ACPI_EINJ_GET_ERROR_TYPE */ - vendor = val & ACPI5_VENDOR_BIT; - tval = val & 0x7fffffff; + vendor = type & ACPI5_VENDOR_BIT; + tval = type & GENMASK(30, 0); /* Only one error type can be specified */ if (tval & (tval - 1)) @@ -671,9 +671,37 @@ static int error_type_set(void *data, u64 val) rc = einj_get_available_error_type(&available_error_type); if (rc) return rc; - if (!(val & available_error_type)) + if (!(type & available_error_type)) return -EINVAL; } + + return 0; +} + +bool einj_is_cxl_error_type(u64 type) +{ + return (type & CXL_ERROR_MASK) && (!(type & ACPI5_VENDOR_BIT)); +} + +static int error_type_get(void *data, u64 *val) +{ + *val = error_type; + + return 0; +} + +static int error_type_set(void *data, u64 val) +{ + int rc; + + /* CXL error types have to be injected from cxl debugfs */ + if (einj_is_cxl_error_type(val)) + return -EINVAL; + + rc = einj_validate_error_type(val); + if (rc) + return rc; + error_type = val; return 0; diff --git a/drivers/acpi/apei/einj-cxl.c b/drivers/acpi/apei/einj-cxl.c new file mode 100644 index 000000000000..34badc6a801e --- /dev/null +++ b/drivers/acpi/apei/einj-cxl.c @@ -0,0 +1,120 @@ +// SPDX-License-Identifier: GPL-2.0-only +/* + * CXL Error INJection support. Used by CXL core to inject + * protocol errors into CXL ports. + * + * Copyright (C) 2023 Advanced Micro Devices, Inc. + * + * Author: Ben Cheatham <benjamin.cheatham@amd.com> + */ +#include <linux/einj-cxl.h> +#include <linux/debugfs.h> + +#include "apei-internal.h" + +/* Defined in einj-core.c */ +extern bool einj_initialized; + +static struct { u32 mask; const char *str; } const einj_cxl_error_type_string[] = { + { BIT(12), "CXL.cache Protocol Correctable" }, + { BIT(13), "CXL.cache Protocol Uncorrectable non-fatal" }, + { BIT(14), "CXL.cache Protocol Uncorrectable fatal" }, + { BIT(15), "CXL.mem Protocol Correctable" }, + { BIT(16), "CXL.mem Protocol Uncorrectable non-fatal" }, + { BIT(17), "CXL.mem Protocol Uncorrectable fatal" }, +}; + +int einj_cxl_available_error_type_show(struct seq_file *m, void *v) +{ + int cxl_err, rc; + u32 available_error_type = 0; + + if (!einj_initialized) + return -ENXIO; + + rc = einj_get_available_error_type(&available_error_type); + if (rc) + return rc; + + for (int pos = 0; pos < ARRAY_SIZE(einj_cxl_error_type_string); pos++) { + cxl_err = ACPI_EINJ_CXL_CACHE_CORRECTABLE << pos; + + if (available_error_type & cxl_err) + seq_printf(m, "0x%08x\t%s\n", + einj_cxl_error_type_string[pos].mask, + einj_cxl_error_type_string[pos].str); + } + + return 0; +} +EXPORT_SYMBOL_NS_GPL(einj_cxl_available_error_type_show, CXL); + +static int cxl_dport_get_sbdf(struct pci_dev *dport_dev, u64 *sbdf) +{ + struct pci_bus *pbus; + struct pci_host_bridge *bridge; + u64 seg = 0, bus; + + pbus = dport_dev->bus; + bridge = pci_find_host_bridge(pbus); + + if (!bridge) + return -ENODEV; + + if (bridge->domain_nr != PCI_DOMAIN_NR_NOT_SET) + seg = bridge->domain_nr; + + bus = pbus->number; + *sbdf = (seg << 24) | (bus << 16) | dport_dev->devfn; + + return 0; +} + +int einj_cxl_inject_rch_error(u64 rcrb, u64 type) +{ + int rc; + + if (!einj_initialized) + return -ENXIO; + + /* Only CXL error types can be specified */ + if (!einj_is_cxl_error_type(type)) + return -EINVAL; + + rc = einj_validate_error_type(type); + if (rc) + return rc; + + return einj_error_inject(type, 0x2, rcrb, GENMASK_ULL(63, 12), 0, 0); +} +EXPORT_SYMBOL_NS_GPL(einj_cxl_inject_rch_error, CXL); + +int einj_cxl_inject_error(struct pci_dev *dport, u64 type) +{ + u64 param4 = 0; + int rc; + + if (!einj_initialized) + return -ENXIO; + + /* Only CXL error types can be specified */ + if (!einj_is_cxl_error_type(type)) + return -EINVAL; + + rc = einj_validate_error_type(type); + if (rc) + return rc; + + rc = cxl_dport_get_sbdf(dport, ¶m4); + if (rc) + return rc; + + return einj_error_inject(type, 0x4, 0, 0, 0, param4); +} +EXPORT_SYMBOL_NS_GPL(einj_cxl_inject_error, CXL); + +bool einj_cxl_is_initialized(void) +{ + return einj_initialized; +} +EXPORT_SYMBOL_NS_GPL(einj_cxl_is_initialized, CXL); diff --git a/include/linux/einj-cxl.h b/include/linux/einj-cxl.h new file mode 100644 index 000000000000..4a1f4600539a --- /dev/null +++ b/include/linux/einj-cxl.h @@ -0,0 +1,40 @@ +/* SPDX-License-Identifier: GPL-2.0-or-later */ +/* + * CXL protocol Error INJection support. + * + * Copyright (c) 2023 Advanced Micro Devices, Inc. + * All Rights Reserved. + * + * Author: Ben Cheatham <benjamin.cheatham@amd.com> + */ +#ifndef EINJ_CXL_H +#define EINJ_CXL_H + +#include <linux/pci.h> + +#if IS_ENABLED(CONFIG_ACPI_APEI_EINJ_CXL) +int einj_cxl_available_error_type_show(struct seq_file *m, void *v); +int einj_cxl_inject_error(struct pci_dev *dport_dev, u64 type); +int einj_cxl_inject_rch_error(u64 rcrb, u64 type); +bool einj_cxl_is_initialized(void); +#else /* !IS_ENABLED(CONFIG_ACPI_APEI_EINJ_CXL) */ +static inline int einj_cxl_available_error_type_show(struct seq_file *m, + void *v) +{ + return -ENXIO; +} + +static inline int einj_cxl_inject_error(struct pci_dev *dport_dev, u64 type) +{ + return -ENXIO; +} + +static inline int einj_cxl_inject_rch_error(u64 rcrb, u64 type) +{ + return -ENXIO; +} + +static inline bool einj_cxl_is_initialized(void) { return false; } +#endif /* CONFIG_ACPI_APEI_EINJ_CXL */ + +#endif /* EINJ_CXL_H */
Remove CXL protocol error types from the EINJ module and move them to a new einj_cxl module. The einj_cxl module implements the necessary handling for CXL protocol error injection and exposes an API for the CXL core to use said functionality. Because the CXL error types require special handling, only allow them to be injected through the einj_cxl module and return an error when attempting to inject through "regular" EINJ. Signed-off-by: Ben Cheatham <Benjamin.Cheatham@amd.com> --- MAINTAINERS | 1 + drivers/acpi/apei/Kconfig | 12 +++ drivers/acpi/apei/Makefile | 2 + drivers/acpi/apei/apei-internal.h | 16 +++ drivers/acpi/apei/{einj.c => einj-core.c} | 78 +++++++++----- drivers/acpi/apei/einj-cxl.c | 120 ++++++++++++++++++++++ include/linux/einj-cxl.h | 40 ++++++++ 7 files changed, 244 insertions(+), 25 deletions(-) rename drivers/acpi/apei/{einj.c => einj-core.c} (94%) create mode 100644 drivers/acpi/apei/einj-cxl.c create mode 100644 include/linux/einj-cxl.h