[PATCH] ACPI / APEI: Boot Error Record Table processing was needlessly complicated

[PATCH] ACPI / APEI: Boot Error Record Table processing was needlessly complicated

[Luck, Tony]

From: Tony Luck <tony.luck@intel.com>

Quoting version 6.1 of the ACPI specification. Section 18.3.1 "Boot
Error Source" says:

  The Boot Error Region is a range of addressable memory OSPM can access
  during initialization to determine if an unhandled error condition
  occurred. System firmware must report this memory range as firmware
  reserved. The format of the Boot Error Region follow that of an Error
  Status Block, this is defined in Section 18.3.2.7. The format of the
  error status block is described by Table 18-342.

This clarifies some points that were obfuscated in earlier versions.
E.g. there is no longer a separate table to describe the format of the
"Boot Error Region" (which was identical to the "Error Status Block").
Also saying "follow that of *an* Error Status Block" makes it clear that
there is just one block (which can still contain multiple "Generic Error
Data Entry structures").

The loop inside bert_print_all() is unnecessary (but probably harmless
as the "while (remain > sizeof(struct acpi_bert_region))" loop should
terminate after we skipped over the first entry.

We can drop the "bert_print_all()" function and just move the four
relevant lines inline in "bert_init()". Since there are no remaining
users of "struct acpi_bert_region" we delete it from <acpi/actbl1.h>

Cc: Len Brown <lenb@kernel.org>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Tomasz Nowicki <tomasz.nowicki@linaro.org>
Cc: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
Cc: Tyler Baicar <tbaicar@codeaurora.org>
Cc: linux-acpi@vger.kernel.org
Signed-off-by: Tony Luck <tony.luck@intel.com>
---
 drivers/acpi/apei/bert.c | 54 +++++++-----------------------------------------
 include/acpi/actbl1.h    | 11 +---------
 2 files changed, 8 insertions(+), 57 deletions(-)

diff --git a/drivers/acpi/apei/bert.c b/drivers/acpi/apei/bert.c
index 12771fcf0417..b28e1573d4cf 100644
--- a/drivers/acpi/apei/bert.c
+++ b/drivers/acpi/apei/bert.c
@@ -34,50 +34,6 @@
 
 static int bert_disable;
 
-static void __init bert_print_all(struct acpi_bert_region *region,
-				  unsigned int region_len)
-{
-	struct acpi_hest_generic_status *estatus =
-		(struct acpi_hest_generic_status *)region;
-	int remain = region_len;
-	u32 estatus_len;
-
-	if (!estatus->block_status)
-		return;
-
-	while (remain > sizeof(struct acpi_bert_region)) {
-		if (cper_estatus_check(estatus)) {
-			pr_err(FW_BUG "Invalid error record.\n");
-			return;
-		}
-
-		estatus_len = cper_estatus_len(estatus);
-		if (remain < estatus_len) {
-			pr_err(FW_BUG "Truncated status block (length: %u).\n",
-			       estatus_len);
-			return;
-		}
-
-		pr_info_once("Error records from previous boot:\n");
-
-		cper_estatus_print(KERN_INFO HW_ERR, estatus);
-
-		/*
-		 * Because the boot error source is "one-time polled" type,
-		 * clear Block Status of current Generic Error Status Block,
-		 * once it's printed.
-		 */
-		estatus->block_status = 0;
-
-		estatus = (void *)estatus + estatus_len;
-		/* No more error records. */
-		if (!estatus->block_status)
-			return;
-
-		remain -= estatus_len;
-	}
-}
-
 static int __init setup_bert_disable(char *str)
 {
 	bert_disable = 1;
@@ -89,7 +45,7 @@ __setup("bert_disable", setup_bert_disable);
 static int __init bert_check_table(struct acpi_table_bert *bert_tab)
 {
 	if (bert_tab->header.length < sizeof(struct acpi_table_bert) ||
-	    bert_tab->region_length < sizeof(struct acpi_bert_region))
+	    bert_tab->region_length < sizeof(struct acpi_hest_generic_status))
 		return -EINVAL;
 
 	return 0;
@@ -98,7 +54,7 @@ static int __init bert_check_table(struct acpi_table_bert *bert_tab)
 static int __init bert_init(void)
 {
 	struct apei_resources bert_resources;
-	struct acpi_bert_region *boot_error_region;
+	struct acpi_hest_generic_status *boot_error_region;
 	struct acpi_table_bert *bert_tab;
 	unsigned int region_len;
 	acpi_status status;
@@ -138,7 +94,11 @@ static int __init bert_init(void)
 		goto out_fini;
 	boot_error_region = ioremap_cache(bert_tab->address, region_len);
 	if (boot_error_region) {
-		bert_print_all(boot_error_region, region_len);
+		if (boot_error_region->block_status) {
+			pr_info("Error records from previous boot:\n");
+			cper_estatus_print(KERN_INFO HW_ERR, boot_error_region);
+			boot_error_region->block_status = 0;
+		}
 		iounmap(boot_error_region);
 	} else {
 		rc = -ENOMEM;
diff --git a/include/acpi/actbl1.h b/include/acpi/actbl1.h
index b4ce55c008b0..cf0bd26774aa 100644
--- a/include/acpi/actbl1.h
+++ b/include/acpi/actbl1.h
@@ -127,16 +127,7 @@ struct acpi_table_bert {
 	struct acpi_table_header header;	/* Common ACPI table header */
 	u32 region_length;	/* Length of the boot error region */
 	u64 address;		/* Physical address of the error region */
-};
-
-/* Boot Error Region (not a subtable, pointed to by Address field above) */
-
-struct acpi_bert_region {
-	u32 block_status;	/* Type of error information */
-	u32 raw_data_offset;	/* Offset to raw error data */
-	u32 raw_data_length;	/* Length of raw error data */
-	u32 data_length;	/* Length of generic error data */
-	u32 error_severity;	/* Severity code */
+				/* which is a struct acpi_hest_generic_status */
 };
 
 /* Values for block_status flags above */
-- 
2.11.0

RE: [PATCH] ACPI / APEI: Boot Error Record Table processing was needlessly complicated

[Luck, Tony]

N.B. this really needs some testing by someone who has a system that
generates populated BERT records.  All I know at the moment is that it
works with en empty BERT log

-Tony

Re: [PATCH] ACPI / APEI: Boot Error Record Table processing was needlessly complicated

[Rafael J. Wysocki]

On Friday, May 05, 2017 11:10:37 AM Luck, Tony wrote:
> From: Tony Luck <tony.luck@intel.com>
> 
> Quoting version 6.1 of the ACPI specification. Section 18.3.1 "Boot
> Error Source" says:
> 
>   The Boot Error Region is a range of addressable memory OSPM can access
>   during initialization to determine if an unhandled error condition
>   occurred. System firmware must report this memory range as firmware
>   reserved. The format of the Boot Error Region follow that of an Error
>   Status Block, this is defined in Section 18.3.2.7. The format of the
>   error status block is described by Table 18-342.
> 
> This clarifies some points that were obfuscated in earlier versions.
> E.g. there is no longer a separate table to describe the format of the
> "Boot Error Region" (which was identical to the "Error Status Block").
> Also saying "follow that of *an* Error Status Block" makes it clear that
> there is just one block (which can still contain multiple "Generic Error
> Data Entry structures").
> 
> The loop inside bert_print_all() is unnecessary (but probably harmless
> as the "while (remain > sizeof(struct acpi_bert_region))" loop should
> terminate after we skipped over the first entry.
> 
> We can drop the "bert_print_all()" function and just move the four
> relevant lines inline in "bert_init()". Since there are no remaining
> users of "struct acpi_bert_region" we delete it from <acpi/actbl1.h>
> 
> Cc: Len Brown <lenb@kernel.org>
> Cc: Huang Ying <ying.huang@intel.com>
> Cc: Borislav Petkov <bp@suse.de>
> Cc: Tomasz Nowicki <tomasz.nowicki@linaro.org>
> Cc: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
> Cc: Tyler Baicar <tbaicar@codeaurora.org>
> Cc: linux-acpi@vger.kernel.org
> Signed-off-by: Tony Luck <tony.luck@intel.com>
> ---
>  drivers/acpi/apei/bert.c | 54 +++++++-----------------------------------------
>  include/acpi/actbl1.h    | 11 +---------
>  2 files changed, 8 insertions(+), 57 deletions(-)
> 
> diff --git a/drivers/acpi/apei/bert.c b/drivers/acpi/apei/bert.c
> index 12771fcf0417..b28e1573d4cf 100644
> --- a/drivers/acpi/apei/bert.c
> +++ b/drivers/acpi/apei/bert.c
> @@ -34,50 +34,6 @@
>  
>  static int bert_disable;
>  
> -static void __init bert_print_all(struct acpi_bert_region *region,
> -				  unsigned int region_len)
> -{
> -	struct acpi_hest_generic_status *estatus =
> -		(struct acpi_hest_generic_status *)region;
> -	int remain = region_len;
> -	u32 estatus_len;
> -
> -	if (!estatus->block_status)
> -		return;
> -
> -	while (remain > sizeof(struct acpi_bert_region)) {
> -		if (cper_estatus_check(estatus)) {
> -			pr_err(FW_BUG "Invalid error record.\n");
> -			return;
> -		}
> -
> -		estatus_len = cper_estatus_len(estatus);
> -		if (remain < estatus_len) {
> -			pr_err(FW_BUG "Truncated status block (length: %u).\n",
> -			       estatus_len);
> -			return;
> -		}
> -
> -		pr_info_once("Error records from previous boot:\n");
> -
> -		cper_estatus_print(KERN_INFO HW_ERR, estatus);
> -
> -		/*
> -		 * Because the boot error source is "one-time polled" type,
> -		 * clear Block Status of current Generic Error Status Block,
> -		 * once it's printed.
> -		 */
> -		estatus->block_status = 0;
> -
> -		estatus = (void *)estatus + estatus_len;
> -		/* No more error records. */
> -		if (!estatus->block_status)
> -			return;
> -
> -		remain -= estatus_len;
> -	}
> -}
> -
>  static int __init setup_bert_disable(char *str)
>  {
>  	bert_disable = 1;
> @@ -89,7 +45,7 @@ __setup("bert_disable", setup_bert_disable);
>  static int __init bert_check_table(struct acpi_table_bert *bert_tab)
>  {
>  	if (bert_tab->header.length < sizeof(struct acpi_table_bert) ||
> -	    bert_tab->region_length < sizeof(struct acpi_bert_region))
> +	    bert_tab->region_length < sizeof(struct acpi_hest_generic_status))
>  		return -EINVAL;
>  
>  	return 0;
> @@ -98,7 +54,7 @@ static int __init bert_check_table(struct acpi_table_bert *bert_tab)
>  static int __init bert_init(void)
>  {
>  	struct apei_resources bert_resources;
> -	struct acpi_bert_region *boot_error_region;
> +	struct acpi_hest_generic_status *boot_error_region;
>  	struct acpi_table_bert *bert_tab;
>  	unsigned int region_len;
>  	acpi_status status;
> @@ -138,7 +94,11 @@ static int __init bert_init(void)
>  		goto out_fini;
>  	boot_error_region = ioremap_cache(bert_tab->address, region_len);
>  	if (boot_error_region) {
> -		bert_print_all(boot_error_region, region_len);
> +		if (boot_error_region->block_status) {
> +			pr_info("Error records from previous boot:\n");
> +			cper_estatus_print(KERN_INFO HW_ERR, boot_error_region);
> +			boot_error_region->block_status = 0;
> +		}
>  		iounmap(boot_error_region);
>  	} else {
>  		rc = -ENOMEM;
> diff --git a/include/acpi/actbl1.h b/include/acpi/actbl1.h
> index b4ce55c008b0..cf0bd26774aa 100644
> --- a/include/acpi/actbl1.h
> +++ b/include/acpi/actbl1.h
> @@ -127,16 +127,7 @@ struct acpi_table_bert {
>  	struct acpi_table_header header;	/* Common ACPI table header */
>  	u32 region_length;	/* Length of the boot error region */
>  	u64 address;		/* Physical address of the error region */
> -};
> -
> -/* Boot Error Region (not a subtable, pointed to by Address field above) */
> -
> -struct acpi_bert_region {
> -	u32 block_status;	/* Type of error information */
> -	u32 raw_data_offset;	/* Offset to raw error data */
> -	u32 raw_data_length;	/* Length of raw error data */
> -	u32 data_length;	/* Length of generic error data */
> -	u32 error_severity;	/* Severity code */
> +				/* which is a struct acpi_hest_generic_status */
>  };
>  
>  /* Values for block_status flags above */
> 

The actbl1.h change should be routed through the upstream ACPICA I think
after we've dropped all things depending on it from the kernel.

Thanks,
Rafael

Re: [PATCH] ACPI / APEI: Boot Error Record Table processing was needlessly complicated

[Borislav Petkov]

On Fri, May 05, 2017 at 11:10:37AM -0700, Luck, Tony wrote:
> From: Tony Luck <tony.luck@intel.com>
> 
> Quoting version 6.1 of the ACPI specification. Section 18.3.1 "Boot
> Error Source" says:
> 
>   The Boot Error Region is a range of addressable memory OSPM can access
>   during initialization to determine if an unhandled error condition
>   occurred. System firmware must report this memory range as firmware
>   reserved. The format of the Boot Error Region follow that of an Error
>   Status Block, this is defined in Section 18.3.2.7. The format of the
>   error status block is described by Table 18-342.
> 
> This clarifies some points that were obfuscated in earlier versions.
> E.g. there is no longer a separate table to describe the format of the
> "Boot Error Region" (which was identical to the "Error Status Block").
> Also saying "follow that of *an* Error Status Block" makes it clear that
> there is just one block (which can still contain multiple "Generic Error
> Data Entry structures").
> 
> The loop inside bert_print_all() is unnecessary (but probably harmless
> as the "while (remain > sizeof(struct acpi_bert_region))" loop should
> terminate after we skipped over the first entry.
> 
> We can drop the "bert_print_all()" function and just move the four
> relevant lines inline in "bert_init()". Since there are no remaining
> users of "struct acpi_bert_region" we delete it from <acpi/actbl1.h>
> 
> Cc: Len Brown <lenb@kernel.org>
> Cc: Huang Ying <ying.huang@intel.com>
> Cc: Borislav Petkov <bp@suse.de>
> Cc: Tomasz Nowicki <tomasz.nowicki@linaro.org>
> Cc: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
> Cc: Tyler Baicar <tbaicar@codeaurora.org>
> Cc: linux-acpi@vger.kernel.org
> Signed-off-by: Tony Luck <tony.luck@intel.com>
> ---
>  drivers/acpi/apei/bert.c | 54 +++++++-----------------------------------------
>  include/acpi/actbl1.h    | 11 +---------
>  2 files changed, 8 insertions(+), 57 deletions(-)

...

> -struct acpi_bert_region {
> -	u32 block_status;	/* Type of error information */
> -	u32 raw_data_offset;	/* Offset to raw error data */
> -	u32 raw_data_length;	/* Length of raw error data */
> -	u32 data_length;	/* Length of generic error data */
> -	u32 error_severity;	/* Severity code */
> +				/* which is a struct acpi_hest_generic_status */

Just this nitpick: please merge this comment with the one above it into
a single:

        u64 address;            /*
				 * Physical address of the error region which is a Generic
				 * Error Status Block (struct acpi_hest_generic_status).
				 */

I would like to advocate to use the actual names from the spec too
because those are so many and so dumbly named, so that one's head
starts spinning pretty quickly, trying to keep 'em apart.

Other than that,

Reviewed-by: Borislav Petkov <bp@suse.de>

-- 
Regards/Gruss,
    Boris.

Good mailing practices for 400: avoid top-posting and trim the reply.

[PATCH v2 0/2] Clean up ACPI Boot Error Record Table code

[Luck, Tony]

From: Tony Luck <tony.luck@intel.com>

ACPI 6.1 spec clarified the structure of the Boot Error Record Table.
Linux code was overly complex, and included a redundant structure
definition of "acpi_bert_region".

v2 changes:
Rafael: Split the change to <acpi/actbl1.h> into a separate patch so
	it can be fed through ACPICA process.
Boris:  Better wording for the comment about the "address" field in 
	the acpi_table_bert structure.

Tony Luck (2):
  ACPI / APEI: Boot Error Record Table processing was needlessly
    complicated
  ACPI / APEI: No remaining users of struct acpi_bert_region

 drivers/acpi/apei/bert.c | 54 +++++++-----------------------------------------
 include/acpi/actbl1.h    | 16 +++++---------
 2 files changed, 12 insertions(+), 58 deletions(-)

Cc: Len Brown <lenb@kernel.org>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Tomasz Nowicki <tomasz.nowicki@linaro.org>
Cc: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
Cc: Tyler Baicar <tbaicar@codeaurora.org>
Cc: linux-acpi@vger.kernel.org
Reviewed-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Tony Luck <tony.luck@intel.com>

-- 
2.11.0

[PATCH v2 1/2] ACPI / APEI: Boot Error Record Table processing was needlessly complicated

[Luck, Tony]

From: Tony Luck <tony.luck@intel.com>

Quoting version 6.1 of the ACPI specification. Section 18.3.1 "Boot
Error Source" says:

  The Boot Error Region is a range of addressable memory OSPM can access
  during initialization to determine if an unhandled error condition
  occurred. System firmware must report this memory range as firmware
  reserved. The format of the Boot Error Region follow that of an Error
  Status Block, this is defined in Section 18.3.2.7. The format of the
  error status block is described by Table 18-342.

This clarifies some points that were obfuscated in earlier versions.
E.g. there is no longer a separate table to describe the format of the
"Boot Error Region" (which was identical to the "Error Status Block").
Also saying "follow that of *an* Error Status Block" makes it clear that
there is just one block (which can still contain multiple "Generic Error
Data Entry structures").

The loop inside bert_print_all() is unnecessary (but probably harmless
as the "while (remain > sizeof(struct acpi_bert_region))" loop should
terminate after we skipped over the first entry.

We can drop the "bert_print_all()" function and just move the four
relevant lines inline in "bert_init()".

Cc: Len Brown <lenb@kernel.org>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Tomasz Nowicki <tomasz.nowicki@linaro.org>
Cc: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
Cc: Tyler Baicar <tbaicar@codeaurora.org>
Cc: linux-acpi@vger.kernel.org
Reviewed-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Tony Luck <tony.luck@intel.com>
---
 drivers/acpi/apei/bert.c | 54 +++++++-----------------------------------------
 1 file changed, 7 insertions(+), 47 deletions(-)

diff --git a/drivers/acpi/apei/bert.c b/drivers/acpi/apei/bert.c
index 12771fcf0417..b28e1573d4cf 100644
--- a/drivers/acpi/apei/bert.c
+++ b/drivers/acpi/apei/bert.c
@@ -34,50 +34,6 @@
 
 static int bert_disable;
 
-static void __init bert_print_all(struct acpi_bert_region *region,
-				  unsigned int region_len)
-{
-	struct acpi_hest_generic_status *estatus =
-		(struct acpi_hest_generic_status *)region;
-	int remain = region_len;
-	u32 estatus_len;
-
-	if (!estatus->block_status)
-		return;
-
-	while (remain > sizeof(struct acpi_bert_region)) {
-		if (cper_estatus_check(estatus)) {
-			pr_err(FW_BUG "Invalid error record.\n");
-			return;
-		}
-
-		estatus_len = cper_estatus_len(estatus);
-		if (remain < estatus_len) {
-			pr_err(FW_BUG "Truncated status block (length: %u).\n",
-			       estatus_len);
-			return;
-		}
-
-		pr_info_once("Error records from previous boot:\n");
-
-		cper_estatus_print(KERN_INFO HW_ERR, estatus);
-
-		/*
-		 * Because the boot error source is "one-time polled" type,
-		 * clear Block Status of current Generic Error Status Block,
-		 * once it's printed.
-		 */
-		estatus->block_status = 0;
-
-		estatus = (void *)estatus + estatus_len;
-		/* No more error records. */
-		if (!estatus->block_status)
-			return;
-
-		remain -= estatus_len;
-	}
-}
-
 static int __init setup_bert_disable(char *str)
 {
 	bert_disable = 1;
@@ -89,7 +45,7 @@ __setup("bert_disable", setup_bert_disable);
 static int __init bert_check_table(struct acpi_table_bert *bert_tab)
 {
 	if (bert_tab->header.length < sizeof(struct acpi_table_bert) ||
-	    bert_tab->region_length < sizeof(struct acpi_bert_region))
+	    bert_tab->region_length < sizeof(struct acpi_hest_generic_status))
 		return -EINVAL;
 
 	return 0;
@@ -98,7 +54,7 @@ static int __init bert_check_table(struct acpi_table_bert *bert_tab)
 static int __init bert_init(void)
 {
 	struct apei_resources bert_resources;
-	struct acpi_bert_region *boot_error_region;
+	struct acpi_hest_generic_status *boot_error_region;
 	struct acpi_table_bert *bert_tab;
 	unsigned int region_len;
 	acpi_status status;
@@ -138,7 +94,11 @@ static int __init bert_init(void)
 		goto out_fini;
 	boot_error_region = ioremap_cache(bert_tab->address, region_len);
 	if (boot_error_region) {
-		bert_print_all(boot_error_region, region_len);
+		if (boot_error_region->block_status) {
+			pr_info("Error records from previous boot:\n");
+			cper_estatus_print(KERN_INFO HW_ERR, boot_error_region);
+			boot_error_region->block_status = 0;
+		}
 		iounmap(boot_error_region);
 	} else {
 		rc = -ENOMEM;
-- 
2.11.0

[PATCH v2 2/2] ACPI / APEI: No remaining users of struct acpi_bert_region

[Luck, Tony]

From: Tony Luck <tony.luck@intel.com>

Version 6.1 of the ACPI specification. Section 18.3.1 "Boot Error
Source" clarified that the format of the Boot Error Region is a "Generic
Error Status Block". So we don't need separate (but identical) "struct
acpi_bert_region" and "struct acpi_hest_generic_status".

Code using acpi_bert_region has been changed to use acpi_hest_generic_status
instead. So we can delete the definition now.

Cc: Len Brown <lenb@kernel.org>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Tomasz Nowicki <tomasz.nowicki@linaro.org>
Cc: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
Cc: Tyler Baicar <tbaicar@codeaurora.org>
Cc: linux-acpi@vger.kernel.org
Reviewed-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Tony Luck <tony.luck@intel.com>
---
 include/acpi/actbl1.h | 16 +++++-----------
 1 file changed, 5 insertions(+), 11 deletions(-)

diff --git a/include/acpi/actbl1.h b/include/acpi/actbl1.h
index b4ce55c008b0..b44663a87bd1 100644
--- a/include/acpi/actbl1.h
+++ b/include/acpi/actbl1.h
@@ -126,17 +126,11 @@ struct acpi_whea_header {
 struct acpi_table_bert {
 	struct acpi_table_header header;	/* Common ACPI table header */
 	u32 region_length;	/* Length of the boot error region */
-	u64 address;		/* Physical address of the error region */
-};
-
-/* Boot Error Region (not a subtable, pointed to by Address field above) */
-
-struct acpi_bert_region {
-	u32 block_status;	/* Type of error information */
-	u32 raw_data_offset;	/* Offset to raw error data */
-	u32 raw_data_length;	/* Length of raw error data */
-	u32 data_length;	/* Length of generic error data */
-	u32 error_severity;	/* Severity code */
+	u64 address;		/*
+				 * Physical address of the error region
+				 * which is a Generic Error Status Block
+				 * (struct acpi_hest_generic_status)
+				 */
 };
 
 /* Values for block_status flags above */
-- 
2.11.0

Re: [PATCH v2 2/2] ACPI / APEI: No remaining users of struct acpi_bert_region

[Rafael J. Wysocki]

On Friday, May 12, 2017 02:42:12 PM Luck, Tony wrote:
> From: Tony Luck <tony.luck@intel.com>
> 
> Version 6.1 of the ACPI specification. Section 18.3.1 "Boot Error
> Source" clarified that the format of the Boot Error Region is a "Generic
> Error Status Block". So we don't need separate (but identical) "struct
> acpi_bert_region" and "struct acpi_hest_generic_status".
> 
> Code using acpi_bert_region has been changed to use acpi_hest_generic_status
> instead. So we can delete the definition now.
> 
> Cc: Len Brown <lenb@kernel.org>
> Cc: Huang Ying <ying.huang@intel.com>
> Cc: Borislav Petkov <bp@suse.de>
> Cc: Tomasz Nowicki <tomasz.nowicki@linaro.org>
> Cc: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
> Cc: Tyler Baicar <tbaicar@codeaurora.org>
> Cc: linux-acpi@vger.kernel.org
> Reviewed-by: Borislav Petkov <bp@suse.de>
> Signed-off-by: Tony Luck <tony.luck@intel.com>
> ---
>  include/acpi/actbl1.h | 16 +++++-----------
>  1 file changed, 5 insertions(+), 11 deletions(-)
> 
> diff --git a/include/acpi/actbl1.h b/include/acpi/actbl1.h
> index b4ce55c008b0..b44663a87bd1 100644
> --- a/include/acpi/actbl1.h
> +++ b/include/acpi/actbl1.h
> @@ -126,17 +126,11 @@ struct acpi_whea_header {
>  struct acpi_table_bert {
>  	struct acpi_table_header header;	/* Common ACPI table header */
>  	u32 region_length;	/* Length of the boot error region */
> -	u64 address;		/* Physical address of the error region */
> -};
> -
> -/* Boot Error Region (not a subtable, pointed to by Address field above) */
> -
> -struct acpi_bert_region {
> -	u32 block_status;	/* Type of error information */
> -	u32 raw_data_offset;	/* Offset to raw error data */
> -	u32 raw_data_length;	/* Length of raw error data */
> -	u32 data_length;	/* Length of generic error data */
> -	u32 error_severity;	/* Severity code */
> +	u64 address;		/*
> +				 * Physical address of the error region
> +				 * which is a Generic Error Status Block
> +				 * (struct acpi_hest_generic_status)
> +				 */
>  };
>  
>  /* Values for block_status flags above */
> 

Well, again, this is an ACPICA header and by changing it in the kernel alone
we make it out of sync with the upstream, which may be a problem for the
maintainers thereof when they port changes from the upstream ACPICA to the
kernel.

So, (a) do we have to change it and (b) can that be routed through the
upstream?

Thanks,
Rafael

Re: [PATCH v2 2/2] ACPI / APEI: No remaining users of struct acpi_bert_region

[Luck, Tony]

On Fri, May 12, 2017 at 11:46:31PM +0200, Rafael J. Wysocki wrote:
> Well, again, this is an ACPICA header and by changing it in the kernel alone
> we make it out of sync with the upstream, which may be a problem for the
> maintainers thereof when they port changes from the upstream ACPICA to the
> kernel.
> 
> So, (a) do we have to change it and (b) can that be routed through the
> upstream?

a) No ... we don't have to

b) Yes. What is the process / mailing list to make an ACPICA change?

-Tony

Re: [PATCH v2 2/2] ACPI / APEI: No remaining users of struct acpi_bert_region

[Rafael J. Wysocki]

On Friday, May 12, 2017 03:32:51 PM Luck, Tony wrote:
> On Fri, May 12, 2017 at 11:46:31PM +0200, Rafael J. Wysocki wrote:
> > Well, again, this is an ACPICA header and by changing it in the kernel alone
> > we make it out of sync with the upstream, which may be a problem for the
> > maintainers thereof when they port changes from the upstream ACPICA to the
> > kernel.
> > 
> > So, (a) do we have to change it and (b) can that be routed through the
> > upstream?
> 
> a) No ... we don't have to

OK

> b) Yes. What is the process / mailing list to make an ACPICA change?

The relevant information is there in MAINTAINERS.

The process is to post changes to the devel mailing list, as per MAINTAINERS,
but it should be sufficient to CC the ACPICA maintainers, which I did, and talk
to them. :-)

Thanks,
Rafael

Re: [PATCH v2 0/2] Clean up ACPI Boot Error Record Table code

[Baicar, Tyler]

On 5/12/2017 3:42 PM, Luck, Tony wrote:
> From: Tony Luck <tony.luck@intel.com>
>
> ACPI 6.1 spec clarified the structure of the Boot Error Record Table.
> Linux code was overly complex, and included a redundant structure
> definition of "acpi_bert_region".
>
> v2 changes:
> Rafael: Split the change to <acpi/actbl1.h> into a separate patch so
> 	it can be fed through ACPICA process.
> Boris:  Better wording for the comment about the "address" field in
> 	the acpi_table_bert structure.
>
> Tony Luck (2):
>    ACPI / APEI: Boot Error Record Table processing was needlessly
>      complicated
>    ACPI / APEI: No remaining users of struct acpi_bert_region
>
>   drivers/acpi/apei/bert.c | 54 +++++++-----------------------------------------
>   include/acpi/actbl1.h    | 16 +++++---------
>   2 files changed, 12 insertions(+), 58 deletions(-)
>
> Cc: Len Brown <lenb@kernel.org>
> Cc: Huang Ying <ying.huang@intel.com>
> Cc: Borislav Petkov <bp@suse.de>
> Cc: Tomasz Nowicki <tomasz.nowicki@linaro.org>
> Cc: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
> Cc: Tyler Baicar <tbaicar@codeaurora.org>
> Cc: linux-acpi@vger.kernel.org
> Reviewed-by: Borislav Petkov <bp@suse.de>
> Signed-off-by: Tony Luck <tony.luck@intel.com>
Hello Tony,

I tested this out and can see everything works properly as long as the 
BERT record is valid.

Previously, if I made the CPER section size too small then the kernel 
would just print the Firmware Bug: Invalid error record. Doing the same 
thing with this patch series I can see an unlimited stream of unknown 
error sections being printed out (I tested this with my ACPI 6.1/UEFI 
2.6 patches which include the non-standard error record support).

[   11.482831] [Hardware Error]:  Error 1, type: recoverable
[   11.482832] [Hardware Error]:   section type: unknown, 
00000000-0000-0000-0000-000000000000
[   11.482833] [Hardware Error]:   section length: 0x0
[   11.482834] [Hardware Error]:  Error 2, type: recoverable
[   11.482835] [Hardware Error]:   section type: unknown, 
00000000-0000-0000-0000-000000000000
[   11.482836] [Hardware Error]:   section length: 0x0
[   11.482837] [Hardware Error]:  Error 3, type: recoverable
[   11.482839] [Hardware Error]:   section type: unknown, 
00000000-0000-0000-0000-000000000000
[   11.482839] [Hardware Error]:   section length: 0x0
...

For that reason, I'd suggest leaving this check in:

if (cper_estatus_check(estatus)) {
     pr_err(FW_BUG "Invalid error record.\n");
     return;
}

Thanks,
Tyler

-- 
Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc.
Qualcomm Technologies, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project.

[PATCH v3] ACPI / APEI: Boot Error Record Table processing was needlessly complicated

[Luck, Tony]

From: Tony Luck <tony.luck@intel.com>

Quoting version 6.1 of the ACPI specification. Section 18.3.1 "Boot
Error Source" says:

  The Boot Error Region is a range of addressable memory OSPM can access
  during initialization to determine if an unhandled error condition
  occurred. System firmware must report this memory range as firmware
  reserved. The format of the Boot Error Region follow that of an Error
  Status Block, this is defined in Section 18.3.2.7. The format of the
  error status block is described by Table 18-342.

This clarifies some points that were obfuscated in earlier versions.
E.g. there is no longer a separate table to describe the format of the
"Boot Error Region" (which was identical to the "Error Status Block").
Also saying "follow that of *an* Error Status Block" makes it clear that
there is just one block (which can still contain multiple "Generic Error
Data Entry structures").

The loop inside bert_print_all() is unnecessary (but probably harmless
as the "while (remain > sizeof(struct acpi_bert_region))" loop should
terminate after we skipped over the first entry.

We can drop the "bert_print_all()" function and just move the four
relevant lines inline in "bert_init()".

Cc: Len Brown <lenb@kernel.org>
Cc: Huang Ying <ying.huang@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Tomasz Nowicki <tomasz.nowicki@linaro.org>
Cc: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
Cc: Tyler Baicar <tbaicar@codeaurora.org>
Cc: linux-acpi@vger.kernel.org
Reviewed-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Tony Luck <tony.luck@intel.com>
---

v3:
1)  Back to just one patch (second part to delete the unused structure
    definition can fly later via ACPICA tree).
2)  Tyler Baicur ran some tests that included non-standard error support
    and found that with an invalid BERT record the kernel loops forever
    printing "section type: unknown" messages. He recommended putting
    back the cper_estatus_check() test that I'd dropped. Here's the net
    diff since v2:

                if (boot_error_region->block_status) {
+                       rc = cper_estatus_check(boot_error_region);
+                       if (rc) {
+                               pr_err(FW_BUG "Invalid error record.\n");
+                               iounmap(boot_error_region);
+                               return rc;
+                       }
                        pr_info("Error records from previous boot:\n");

 drivers/acpi/apei/bert.c | 60 +++++++++++-------------------------------------
 1 file changed, 13 insertions(+), 47 deletions(-)

diff --git a/drivers/acpi/apei/bert.c b/drivers/acpi/apei/bert.c
index 12771fcf0417..2cf4c6441821 100644
--- a/drivers/acpi/apei/bert.c
+++ b/drivers/acpi/apei/bert.c
@@ -34,50 +34,6 @@
 
 static int bert_disable;
 
-static void __init bert_print_all(struct acpi_bert_region *region,
-				  unsigned int region_len)
-{
-	struct acpi_hest_generic_status *estatus =
-		(struct acpi_hest_generic_status *)region;
-	int remain = region_len;
-	u32 estatus_len;
-
-	if (!estatus->block_status)
-		return;
-
-	while (remain > sizeof(struct acpi_bert_region)) {
-		if (cper_estatus_check(estatus)) {
-			pr_err(FW_BUG "Invalid error record.\n");
-			return;
-		}
-
-		estatus_len = cper_estatus_len(estatus);
-		if (remain < estatus_len) {
-			pr_err(FW_BUG "Truncated status block (length: %u).\n",
-			       estatus_len);
-			return;
-		}
-
-		pr_info_once("Error records from previous boot:\n");
-
-		cper_estatus_print(KERN_INFO HW_ERR, estatus);
-
-		/*
-		 * Because the boot error source is "one-time polled" type,
-		 * clear Block Status of current Generic Error Status Block,
-		 * once it's printed.
-		 */
-		estatus->block_status = 0;
-
-		estatus = (void *)estatus + estatus_len;
-		/* No more error records. */
-		if (!estatus->block_status)
-			return;
-
-		remain -= estatus_len;
-	}
-}
-
 static int __init setup_bert_disable(char *str)
 {
 	bert_disable = 1;
@@ -89,7 +45,7 @@ __setup("bert_disable", setup_bert_disable);
 static int __init bert_check_table(struct acpi_table_bert *bert_tab)
 {
 	if (bert_tab->header.length < sizeof(struct acpi_table_bert) ||
-	    bert_tab->region_length < sizeof(struct acpi_bert_region))
+	    bert_tab->region_length < sizeof(struct acpi_hest_generic_status))
 		return -EINVAL;
 
 	return 0;
@@ -98,7 +54,7 @@ static int __init bert_check_table(struct acpi_table_bert *bert_tab)
 static int __init bert_init(void)
 {
 	struct apei_resources bert_resources;
-	struct acpi_bert_region *boot_error_region;
+	struct acpi_hest_generic_status *boot_error_region;
 	struct acpi_table_bert *bert_tab;
 	unsigned int region_len;
 	acpi_status status;
@@ -138,7 +94,17 @@ static int __init bert_init(void)
 		goto out_fini;
 	boot_error_region = ioremap_cache(bert_tab->address, region_len);
 	if (boot_error_region) {
-		bert_print_all(boot_error_region, region_len);
+		if (boot_error_region->block_status) {
+			rc = cper_estatus_check(boot_error_region);
+			if (rc) {
+				pr_err(FW_BUG "Invalid error record.\n");
+				iounmap(boot_error_region);
+				return rc;
+			}
+			pr_info("Error records from previous boot:\n");
+			cper_estatus_print(KERN_INFO HW_ERR, boot_error_region);
+			boot_error_region->block_status = 0;
+		}
 		iounmap(boot_error_region);
 	} else {
 		rc = -ENOMEM;
-- 
2.11.0

Re: [PATCH v3] ACPI / APEI: Boot Error Record Table processing was needlessly complicated

[Baicar, Tyler]

On 5/16/2017 4:58 PM, Luck, Tony wrote:
> From: Tony Luck <tony.luck@intel.com>
>
> Quoting version 6.1 of the ACPI specification. Section 18.3.1 "Boot
> Error Source" says:
>
>    The Boot Error Region is a range of addressable memory OSPM can access
>    during initialization to determine if an unhandled error condition
>    occurred. System firmware must report this memory range as firmware
>    reserved. The format of the Boot Error Region follow that of an Error
>    Status Block, this is defined in Section 18.3.2.7. The format of the
>    error status block is described by Table 18-342.
>
> This clarifies some points that were obfuscated in earlier versions.
> E.g. there is no longer a separate table to describe the format of the
> "Boot Error Region" (which was identical to the "Error Status Block").
> Also saying "follow that of *an* Error Status Block" makes it clear that
> there is just one block (which can still contain multiple "Generic Error
> Data Entry structures").
>
> The loop inside bert_print_all() is unnecessary (but probably harmless
> as the "while (remain > sizeof(struct acpi_bert_region))" loop should
> terminate after we skipped over the first entry.
>
> We can drop the "bert_print_all()" function and just move the four
> relevant lines inline in "bert_init()".
>
> Cc: Len Brown <lenb@kernel.org>
> Cc: Huang Ying <ying.huang@intel.com>
> Cc: Borislav Petkov <bp@suse.de>
> Cc: Tomasz Nowicki <tomasz.nowicki@linaro.org>
> Cc: Jonathan (Zhixiong) Zhang <zjzhang@codeaurora.org>
> Cc: Tyler Baicar <tbaicar@codeaurora.org>
> Cc: linux-acpi@vger.kernel.org
> Reviewed-by: Borislav Petkov <bp@suse.de>
> Signed-off-by: Tony Luck <tony.luck@intel.com>
Tested-by: Tyler Baicar <tbaicar@codeaurora.org>

-- 
Qualcomm Datacenter Technologies, Inc. as an affiliate of Qualcomm Technologies, Inc.
Qualcomm Technologies, Inc. is a member of the Code Aurora Forum,
a Linux Foundation Collaborative Project.