Message ID | 20220407105120.1280-1-lorenzo.pieralisi@arm.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | ACPI: osl: Fix BERT error region memory mapping | expand |
On Thu, Apr 7, 2022 at 12:51 PM Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> wrote: > > Currently the sysfs interface maps the BERT error region as "memory" > (through acpi_os_map_memory()) in order to copy the error records into > memory buffers through memory operations (eg memory_read_from_buffer()). > > The OS system cannot detect whether the BERT error region is part of > system RAM or it is "device memory" (eg BMC memory) and therefore it > cannot detect which memory attributes the bus to memory support (and > corresponding kernel mapping, unless firmware provides the required > information). > > The acpi_os_map_memory() arch backend implementation determines the > mapping attributes. On arm64, if the BERT error region is not present in > the EFI memory map, the error region is mapped as device-nGnRnE; this > triggers alignment faults since memcpy unaligned accesses are not > allowed in device-nGnRnE regions. > > The ACPI sysfs code cannot therefore map by default the BERT error > region with memory semantics but should use a safer default. > > Change the sysfs code to map the BERT error region as MMIO (through > acpi_os_map_iomem()) and use the memcpy_fromio() interface to read the > error region into the kernel buffer. > Hi, I tested this patch on top of the arm tree for-kernelci branch (a2c0b0fbe014). I wasn't able to trigger the original problem on the same HW, and the patch didn't introduce any new issues on these runs, nor on other randomly chosen aarch64 machines. Tested-by: Veronika Kabatova <vkabatov@redhat.com> > Link: https://lore.kernel.org/linux-arm-kernel/31ffe8fc-f5ee-2858-26c5-0fd8bdd68702@arm.com > Link: https://lore.kernel.org/linux-acpi/CAJZ5v0g+OVbhuUUDrLUCfX_mVqY_e8ubgLTU98=jfjTeb4t+Pw@mail.gmail.com > Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> > Cc: Ard Biesheuvel <ardb@kernel.org> > Cc: Will Deacon <will@kernel.org> > Cc: Hanjun Guo <guohanjun@huawei.com> > Cc: Sudeep Holla <sudeep.holla@arm.com> > Cc: Catalin Marinas <catalin.marinas@arm.com> > Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> > --- > drivers/acpi/sysfs.c | 25 ++++++++++++++++++------- > 1 file changed, 18 insertions(+), 7 deletions(-) > > diff --git a/drivers/acpi/sysfs.c b/drivers/acpi/sysfs.c > index a4b638bea6f1..cc2fe0618178 100644 > --- a/drivers/acpi/sysfs.c > +++ b/drivers/acpi/sysfs.c > @@ -415,19 +415,30 @@ static ssize_t acpi_data_show(struct file *filp, struct kobject *kobj, > loff_t offset, size_t count) > { > struct acpi_data_attr *data_attr; > - void *base; > - ssize_t rc; > + void __iomem *base; > + ssize_t size; > > data_attr = container_of(bin_attr, struct acpi_data_attr, attr); > + size = data_attr->attr.size; > + > + if (offset < 0) > + return -EINVAL; > + > + if (offset >= size) > + return 0; > > - base = acpi_os_map_memory(data_attr->addr, data_attr->attr.size); > + if (count > size - offset) > + count = size - offset; > + > + base = acpi_os_map_iomem(data_attr->addr, size); > if (!base) > return -ENOMEM; > - rc = memory_read_from_buffer(buf, count, &offset, base, > - data_attr->attr.size); > - acpi_os_unmap_memory(base, data_attr->attr.size); > > - return rc; > + memcpy_fromio(buf, base + offset, count); > + > + acpi_os_unmap_iomem(base, size); > + > + return count; > } > > static int acpi_bert_data_init(void *th, struct acpi_data_attr *data_attr) > -- > 2.31.0 >
On Thu, Apr 07, 2022 at 11:51:20AM +0100, Lorenzo Pieralisi wrote: > Currently the sysfs interface maps the BERT error region as "memory" > (through acpi_os_map_memory()) in order to copy the error records into > memory buffers through memory operations (eg memory_read_from_buffer()). > > The OS system cannot detect whether the BERT error region is part of > system RAM or it is "device memory" (eg BMC memory) and therefore it > cannot detect which memory attributes the bus to memory support (and > corresponding kernel mapping, unless firmware provides the required > information). > > The acpi_os_map_memory() arch backend implementation determines the > mapping attributes. On arm64, if the BERT error region is not present in > the EFI memory map, the error region is mapped as device-nGnRnE; this > triggers alignment faults since memcpy unaligned accesses are not > allowed in device-nGnRnE regions. > > The ACPI sysfs code cannot therefore map by default the BERT error > region with memory semantics but should use a safer default. > > Change the sysfs code to map the BERT error region as MMIO (through > acpi_os_map_iomem()) and use the memcpy_fromio() interface to read the > error region into the kernel buffer. > > Link: https://lore.kernel.org/linux-arm-kernel/31ffe8fc-f5ee-2858-26c5-0fd8bdd68702@arm.com > Link: https://lore.kernel.org/linux-acpi/CAJZ5v0g+OVbhuUUDrLUCfX_mVqY_e8ubgLTU98=jfjTeb4t+Pw@mail.gmail.com > Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> > Cc: Ard Biesheuvel <ardb@kernel.org> > Cc: Will Deacon <will@kernel.org> > Cc: Hanjun Guo <guohanjun@huawei.com> > Cc: Sudeep Holla <sudeep.holla@arm.com> > Cc: Catalin Marinas <catalin.marinas@arm.com> > Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> > --- > drivers/acpi/sysfs.c | 25 ++++++++++++++++++------- > 1 file changed, 18 insertions(+), 7 deletions(-) > > diff --git a/drivers/acpi/sysfs.c b/drivers/acpi/sysfs.c > index a4b638bea6f1..cc2fe0618178 100644 > --- a/drivers/acpi/sysfs.c > +++ b/drivers/acpi/sysfs.c > @@ -415,19 +415,30 @@ static ssize_t acpi_data_show(struct file *filp, struct kobject *kobj, > loff_t offset, size_t count) > { > struct acpi_data_attr *data_attr; > - void *base; > - ssize_t rc; > + void __iomem *base; > + ssize_t size; > > data_attr = container_of(bin_attr, struct acpi_data_attr, attr); > + size = data_attr->attr.size; > + > + if (offset < 0) > + return -EINVAL; > + > + if (offset >= size) > + return 0; > > - base = acpi_os_map_memory(data_attr->addr, data_attr->attr.size); > + if (count > size - offset) > + count = size - offset; > + > + base = acpi_os_map_iomem(data_attr->addr, size); > if (!base) > return -ENOMEM; > - rc = memory_read_from_buffer(buf, count, &offset, base, > - data_attr->attr.size); > - acpi_os_unmap_memory(base, data_attr->attr.size); > > - return rc; > + memcpy_fromio(buf, base + offset, count); > + > + acpi_os_unmap_iomem(base, size); > + > + return count; > } > > static int acpi_bert_data_init(void *th, struct acpi_data_attr *data_attr) I've tested this patch with the reproducer and I was unable to reproduce the issue. Tested-by: Aristeu Rozanski <aris@redhat.com>
On Thu, 7 Apr 2022 at 12:51, Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> wrote: > > Currently the sysfs interface maps the BERT error region as "memory" > (through acpi_os_map_memory()) in order to copy the error records into > memory buffers through memory operations (eg memory_read_from_buffer()). > > The OS system cannot detect whether the BERT error region is part of > system RAM or it is "device memory" (eg BMC memory) and therefore it > cannot detect which memory attributes the bus to memory support (and > corresponding kernel mapping, unless firmware provides the required > information). > > The acpi_os_map_memory() arch backend implementation determines the > mapping attributes. On arm64, if the BERT error region is not present in > the EFI memory map, the error region is mapped as device-nGnRnE; this > triggers alignment faults since memcpy unaligned accesses are not > allowed in device-nGnRnE regions. > > The ACPI sysfs code cannot therefore map by default the BERT error > region with memory semantics but should use a safer default. > > Change the sysfs code to map the BERT error region as MMIO (through > acpi_os_map_iomem()) and use the memcpy_fromio() interface to read the > error region into the kernel buffer. > > Link: https://lore.kernel.org/linux-arm-kernel/31ffe8fc-f5ee-2858-26c5-0fd8bdd68702@arm.com > Link: https://lore.kernel.org/linux-acpi/CAJZ5v0g+OVbhuUUDrLUCfX_mVqY_e8ubgLTU98=jfjTeb4t+Pw@mail.gmail.com > Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> > Cc: Ard Biesheuvel <ardb@kernel.org> > Cc: Will Deacon <will@kernel.org> > Cc: Hanjun Guo <guohanjun@huawei.com> > Cc: Sudeep Holla <sudeep.holla@arm.com> > Cc: Catalin Marinas <catalin.marinas@arm.com> > Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> Acked-by: Ard Biesheuvel <ardb@kernel.org> > --- > drivers/acpi/sysfs.c | 25 ++++++++++++++++++------- > 1 file changed, 18 insertions(+), 7 deletions(-) > > diff --git a/drivers/acpi/sysfs.c b/drivers/acpi/sysfs.c > index a4b638bea6f1..cc2fe0618178 100644 > --- a/drivers/acpi/sysfs.c > +++ b/drivers/acpi/sysfs.c > @@ -415,19 +415,30 @@ static ssize_t acpi_data_show(struct file *filp, struct kobject *kobj, > loff_t offset, size_t count) > { > struct acpi_data_attr *data_attr; > - void *base; > - ssize_t rc; > + void __iomem *base; > + ssize_t size; > > data_attr = container_of(bin_attr, struct acpi_data_attr, attr); > + size = data_attr->attr.size; > + > + if (offset < 0) > + return -EINVAL; > + > + if (offset >= size) > + return 0; > > - base = acpi_os_map_memory(data_attr->addr, data_attr->attr.size); > + if (count > size - offset) > + count = size - offset; > + > + base = acpi_os_map_iomem(data_attr->addr, size); > if (!base) > return -ENOMEM; > - rc = memory_read_from_buffer(buf, count, &offset, base, > - data_attr->attr.size); > - acpi_os_unmap_memory(base, data_attr->attr.size); > > - return rc; > + memcpy_fromio(buf, base + offset, count); > + > + acpi_os_unmap_iomem(base, size); > + > + return count; > } > > static int acpi_bert_data_init(void *th, struct acpi_data_attr *data_attr) > -- > 2.31.0 >
On Wed, Apr 13, 2022 at 7:41 PM Ard Biesheuvel <ardb@kernel.org> wrote: > > On Thu, 7 Apr 2022 at 12:51, Lorenzo Pieralisi > <lorenzo.pieralisi@arm.com> wrote: > > > > Currently the sysfs interface maps the BERT error region as "memory" > > (through acpi_os_map_memory()) in order to copy the error records into > > memory buffers through memory operations (eg memory_read_from_buffer()). > > > > The OS system cannot detect whether the BERT error region is part of > > system RAM or it is "device memory" (eg BMC memory) and therefore it > > cannot detect which memory attributes the bus to memory support (and > > corresponding kernel mapping, unless firmware provides the required > > information). > > > > The acpi_os_map_memory() arch backend implementation determines the > > mapping attributes. On arm64, if the BERT error region is not present in > > the EFI memory map, the error region is mapped as device-nGnRnE; this > > triggers alignment faults since memcpy unaligned accesses are not > > allowed in device-nGnRnE regions. > > > > The ACPI sysfs code cannot therefore map by default the BERT error > > region with memory semantics but should use a safer default. > > > > Change the sysfs code to map the BERT error region as MMIO (through > > acpi_os_map_iomem()) and use the memcpy_fromio() interface to read the > > error region into the kernel buffer. > > > > Link: https://lore.kernel.org/linux-arm-kernel/31ffe8fc-f5ee-2858-26c5-0fd8bdd68702@arm.com > > Link: https://lore.kernel.org/linux-acpi/CAJZ5v0g+OVbhuUUDrLUCfX_mVqY_e8ubgLTU98=jfjTeb4t+Pw@mail.gmail.com > > Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> > > Cc: Ard Biesheuvel <ardb@kernel.org> > > Cc: Will Deacon <will@kernel.org> > > Cc: Hanjun Guo <guohanjun@huawei.com> > > Cc: Sudeep Holla <sudeep.holla@arm.com> > > Cc: Catalin Marinas <catalin.marinas@arm.com> > > Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> > > Acked-by: Ard Biesheuvel <ardb@kernel.org> Applied as 5.19 material, thanks! > > --- > > drivers/acpi/sysfs.c | 25 ++++++++++++++++++------- > > 1 file changed, 18 insertions(+), 7 deletions(-) > > > > diff --git a/drivers/acpi/sysfs.c b/drivers/acpi/sysfs.c > > index a4b638bea6f1..cc2fe0618178 100644 > > --- a/drivers/acpi/sysfs.c > > +++ b/drivers/acpi/sysfs.c > > @@ -415,19 +415,30 @@ static ssize_t acpi_data_show(struct file *filp, struct kobject *kobj, > > loff_t offset, size_t count) > > { > > struct acpi_data_attr *data_attr; > > - void *base; > > - ssize_t rc; > > + void __iomem *base; > > + ssize_t size; > > > > data_attr = container_of(bin_attr, struct acpi_data_attr, attr); > > + size = data_attr->attr.size; > > + > > + if (offset < 0) > > + return -EINVAL; > > + > > + if (offset >= size) > > + return 0; > > > > - base = acpi_os_map_memory(data_attr->addr, data_attr->attr.size); > > + if (count > size - offset) > > + count = size - offset; > > + > > + base = acpi_os_map_iomem(data_attr->addr, size); > > if (!base) > > return -ENOMEM; > > - rc = memory_read_from_buffer(buf, count, &offset, base, > > - data_attr->attr.size); > > - acpi_os_unmap_memory(base, data_attr->attr.size); > > > > - return rc; > > + memcpy_fromio(buf, base + offset, count); > > + > > + acpi_os_unmap_iomem(base, size); > > + > > + return count; > > } > > > > static int acpi_bert_data_init(void *th, struct acpi_data_attr *data_attr) > > -- > > 2.31.0 > >
diff --git a/drivers/acpi/sysfs.c b/drivers/acpi/sysfs.c index a4b638bea6f1..cc2fe0618178 100644 --- a/drivers/acpi/sysfs.c +++ b/drivers/acpi/sysfs.c @@ -415,19 +415,30 @@ static ssize_t acpi_data_show(struct file *filp, struct kobject *kobj, loff_t offset, size_t count) { struct acpi_data_attr *data_attr; - void *base; - ssize_t rc; + void __iomem *base; + ssize_t size; data_attr = container_of(bin_attr, struct acpi_data_attr, attr); + size = data_attr->attr.size; + + if (offset < 0) + return -EINVAL; + + if (offset >= size) + return 0; - base = acpi_os_map_memory(data_attr->addr, data_attr->attr.size); + if (count > size - offset) + count = size - offset; + + base = acpi_os_map_iomem(data_attr->addr, size); if (!base) return -ENOMEM; - rc = memory_read_from_buffer(buf, count, &offset, base, - data_attr->attr.size); - acpi_os_unmap_memory(base, data_attr->attr.size); - return rc; + memcpy_fromio(buf, base + offset, count); + + acpi_os_unmap_iomem(base, size); + + return count; } static int acpi_bert_data_init(void *th, struct acpi_data_attr *data_attr)
Currently the sysfs interface maps the BERT error region as "memory" (through acpi_os_map_memory()) in order to copy the error records into memory buffers through memory operations (eg memory_read_from_buffer()). The OS system cannot detect whether the BERT error region is part of system RAM or it is "device memory" (eg BMC memory) and therefore it cannot detect which memory attributes the bus to memory support (and corresponding kernel mapping, unless firmware provides the required information). The acpi_os_map_memory() arch backend implementation determines the mapping attributes. On arm64, if the BERT error region is not present in the EFI memory map, the error region is mapped as device-nGnRnE; this triggers alignment faults since memcpy unaligned accesses are not allowed in device-nGnRnE regions. The ACPI sysfs code cannot therefore map by default the BERT error region with memory semantics but should use a safer default. Change the sysfs code to map the BERT error region as MMIO (through acpi_os_map_iomem()) and use the memcpy_fromio() interface to read the error region into the kernel buffer. Link: https://lore.kernel.org/linux-arm-kernel/31ffe8fc-f5ee-2858-26c5-0fd8bdd68702@arm.com Link: https://lore.kernel.org/linux-acpi/CAJZ5v0g+OVbhuUUDrLUCfX_mVqY_e8ubgLTU98=jfjTeb4t+Pw@mail.gmail.com Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com> Cc: Ard Biesheuvel <ardb@kernel.org> Cc: Will Deacon <will@kernel.org> Cc: Hanjun Guo <guohanjun@huawei.com> Cc: Sudeep Holla <sudeep.holla@arm.com> Cc: Catalin Marinas <catalin.marinas@arm.com> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net> --- drivers/acpi/sysfs.c | 25 ++++++++++++++++++------- 1 file changed, 18 insertions(+), 7 deletions(-)