Re: [PATCH v3 03/14] acpi/ghes: Use HEST table offsets when preparing GHES records

[Igor Mammedov]

On Fri, 31 Jan 2025 18:42:44 +0100
Mauro Carvalho Chehab <mchehab+huawei@xxxxxxxxxx> wrote:

> There are two pointers that are needed during error injection:
> 
> 1. The start address of the CPER block to be stored;
> 2. The address of the ack.
> 
> It is preferable to calculate them from the HEST table.  This allows
> checking the source ID, the size of the table and the type of the
> HEST error block structures.
> 
> Yet, keep the old code, as this is needed for migration purposes.
> 
> Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@xxxxxxxxxx>
> ---
>  hw/acpi/ghes.c         | 132 ++++++++++++++++++++++++++++++++++++-----
>  include/hw/acpi/ghes.h |   1 +
>  2 files changed, 119 insertions(+), 14 deletions(-)
> 
> diff --git a/hw/acpi/ghes.c b/hw/acpi/ghes.c
> index 27478f2d5674..8f284fd191a6 100644
> --- a/hw/acpi/ghes.c
> +++ b/hw/acpi/ghes.c
> @@ -41,6 +41,12 @@
>  /* Address offset in Generic Address Structure(GAS) */
>  #define GAS_ADDR_OFFSET 4
>  
> +/*
> + * ACPI spec 1.0b
> + * 5.2.3 System Description Table Header
> + */
> +#define ACPI_DESC_HEADER_OFFSET     36
> +
>  /*
>   * The total size of Generic Error Data Entry
>   * ACPI 6.1/6.2: 18.3.2.7.1 Generic Error Data,
> @@ -61,6 +67,25 @@
>   */
>  #define ACPI_GHES_GESB_SIZE                 20
>  
> +/*
> + * Offsets with regards to the start of the HEST table stored at
> + * ags->hest_addr_le,

If I read this literary, then offsets above are not what
declared later in this patch.
I'd really drop this comment altogether as it's confusing,
and rather get variables/macro naming right

> according with the memory layout map at
> + * docs/specs/acpi_hest_ghes.rst.
> + */

what we need is update to above doc, describing new and old ways.
a separate patch.

> +
> +/*
> + * ACPI 6.2: 18.3.2.8 Generic Hardware Error Source version 2
      ^^^^^^^^ - wrt version, I see it in 6.1.
      our req is to point to the earliest doc where it has appeared.
      it it must point to a later version for some justified reason
      the explanation 'why' should be mentioned in comment message.

please check all versioning/chapters you are touching/adding in this series.


> + * Table 18-382 Generic Hardware Error Source version 2 (GHESv2) Structure
> + */
> +#define HEST_GHES_V2_TABLE_SIZE  92

it's not table but rather an GHES_V2 entry in HEST
and should be named as such (emph on _entry_)

> +#define GHES_READ_ACK_ADDR_OFF          64
please, add a comment like below but for 'Read Ack Register'


> +/*
> + * ACPI 6.2: 18.3.2.7: Generic Hardware Error Source
> + * Table 18-380: 'Error Status Address' field
> + */
> +#define GHES_ERR_STATUS_ADDR_OFF  20
> +
>  /*
>   * Values for error_severity field
>   */
> @@ -212,14 +237,6 @@ static void build_ghes_error_table(GArray *hardware_errors, BIOSLinker *linker,
>  {
>      int i, error_status_block_offset;
>  
> -    /*
> -     * TODO: Current version supports only one source.
> -     * A further patch will drop this check, after adding a proper migration
> -     * code, as, for the code to work, we need to store a bios pointer to the
> -     * HEST table.
> -     */
> -    assert(num_sources == 1);
> -
>      /* Build error_block_address */
>      for (i = 0; i < num_sources; i++) {
>          build_append_int_noprefix(hardware_errors, 0, sizeof(uint64_t));
> @@ -352,6 +369,14 @@ void acpi_build_hest(GArray *table_data, GArray *hardware_errors,
>                          .oem_id = oem_id, .oem_table_id = oem_table_id };
>      uint32_t hest_offset;
>      int i;
> +    AcpiGedState *acpi_ged_state;
> +    AcpiGhesState *ags = NULL;
> +
> +    acpi_ged_state = ACPI_GED(object_resolve_path_type("", TYPE_ACPI_GED,
> +                                                       NULL));
> +    if (acpi_ged_state) {
> +        ags = &acpi_ged_state->ghes_state;
> +    }
>  
>      hest_offset = table_data->len;
>  
> @@ -371,10 +396,12 @@ void acpi_build_hest(GArray *table_data, GArray *hardware_errors,
>       * Tell firmware to write into GPA the address of HEST via fw_cfg,
>       * once initialized.
>       */
> -    bios_linker_loader_write_pointer(linker,
> -                                     ACPI_HEST_ADDR_FW_CFG_FILE, 0,
> -                                     sizeof(uint64_t),
> -                                     ACPI_BUILD_TABLE_FILE, hest_offset);
> +    if (ags->use_hest_addr) {
> +        bios_linker_loader_write_pointer(linker,
> +                                         ACPI_HEST_ADDR_FW_CFG_FILE, 0,
> +                                         sizeof(uint64_t),
> +                                         ACPI_BUILD_TABLE_FILE, hest_offset);
> +    }

I'd move this patch before 2/14, to avoid issues during bisection.

Also legacy variant is hidden in build_ghes_error_table()

    /*                                                                           
     * tell firmware to write hardware_errors GPA into                           
     * hardware_errors_addr fw_cfg, once the former has been initialized.        
     */
    bios_linker_loader_write_pointer()

 and after this patch we end up with scattered code that should pick
 only one them (but doesn't).

As prereq, I'd move legacy into acpi_build_hest() as separate patch,
then do this patch adds above 'if' gate,
and followup patch [2/14 currently] adds  bios_linker_loader_write_pointer(ACPI_HEST_ADDR_FW_CFG_FILE)

>  }
>  
>  void acpi_ghes_add_fw_cfg(AcpiGhesState *ags, FWCfgState *s,

shouldn't we do the same for fw_cfg_add_file_callback() hunk added
in previous patch and related 'fw_cfg_add_file_callback(s, ACPI_HW_ERROR_ADDR_FW_CFG_FILE'
we need only one of them.

> @@ -420,6 +447,78 @@ static void get_hw_error_offsets(uint64_t ghes_addr,
>      *read_ack_register_addr = ghes_addr + sizeof(uint64_t);
>  }
>  
> +static void get_ghes_source_offsets(uint16_t source_id,
> +                                    uint64_t hest_addr,
> +                                    uint64_t *cper_addr,
> +                                    uint64_t *read_ack_start_addr,
> +                                    Error **errp)
> +{
> +    uint64_t hest_err_block_addr, hest_read_ack_addr;
> +    uint64_t err_source_entry, error_block_addr;
> +    uint32_t num_sources, i;
> +
> +    hest_addr += ACPI_DESC_HEADER_OFFSET;
> +
> +    cpu_physical_memory_read(hest_addr, &num_sources,
> +                             sizeof(num_sources));
> +    num_sources = le32_to_cpu(num_sources);
> +
> +    err_source_entry = hest_addr + sizeof(num_sources);
> +
> +    /*
> +     * Currently, HEST Error source navigates only for GHESv2 tables
> +     */

> +
not needed newline

> +    for (i = 0; i < num_sources; i++) {
> +        uint64_t addr = err_source_entry;
> +        uint16_t type, src_id;
> +
> +        cpu_physical_memory_read(addr, &type, sizeof(type));
> +        type = le16_to_cpu(type);
> +
> +        /* For now, we only know the size of GHESv2 table */
> +        if (type != ACPI_GHES_SOURCE_GENERIC_ERROR_V2) {
> +            error_setg(errp, "HEST: type %d not supported.", type);
> +            return;
> +        }
> +
> +        /* Compare CPER source address at the GHESv2 structure */
> +        addr += sizeof(type);
> +        cpu_physical_memory_read(addr, &src_id, sizeof(src_id));
> +        if (le16_to_cpu(src_id) == source_id) {
> +            break;
> +        }
> +
> +        err_source_entry += HEST_GHES_V2_TABLE_SIZE;
> +    }
> +    if (i == num_sources) {
> +        error_setg(errp, "HEST: Source %d not found.", source_id);
> +        return;
> +    }
> +
> +    /* Navigate though table address pointers */
> +    hest_err_block_addr = err_source_entry + GHES_ERR_STATUS_ADDR_OFF +
> +                          GAS_ADDR_OFFSET;
> +
> +    cpu_physical_memory_read(hest_err_block_addr, &error_block_addr,
> +                             sizeof(error_block_addr));
> +
I'd drop newlines for related read/processing

> +    error_block_addr = le64_to_cpu(error_block_addr);

> +
> +    cpu_physical_memory_read(error_block_addr, cper_addr,
> +                             sizeof(*cper_addr));
> +
ditto

> +    *cper_addr = le64_to_cpu(*cper_addr);
> +
> +    hest_read_ack_addr = err_source_entry + GHES_READ_ACK_ADDR_OFF +
> +                         GAS_ADDR_OFFSET;
> +
> +    cpu_physical_memory_read(hest_read_ack_addr, read_ack_start_addr,
> +                             sizeof(*read_ack_start_addr));
> +
ditto

> +    *read_ack_start_addr = le64_to_cpu(*read_ack_start_addr);
> +}
> +
>  void ghes_record_cper_errors(const void *cper, size_t len,
>                               uint16_t source_id, Error **errp)
>  {
> @@ -440,8 +539,13 @@ void ghes_record_cper_errors(const void *cper, size_t len,
>      }
>      ags = &acpi_ged_state->ghes_state;
>  
> -    get_hw_error_offsets(le64_to_cpu(ags->hw_error_le),
> -                         &cper_addr, &read_ack_register_addr);
> +    if (!ags->hest_addr_le) {
> +        get_hw_error_offsets(le64_to_cpu(ags->hw_error_le),
> +                             &cper_addr, &read_ack_register_addr);
> +    } else {
> +        get_ghes_source_offsets(source_id, le64_to_cpu(ags->hest_addr_le),
> +                                &cper_addr, &read_ack_register_addr, errp);
> +    }
>  
>      if (!cper_addr) {
>          error_setg(errp, "can not find Generic Error Status Block");
> diff --git a/include/hw/acpi/ghes.h b/include/hw/acpi/ghes.h
> index 237721fec0a2..6c2e57af0456 100644
> --- a/include/hw/acpi/ghes.h
> +++ b/include/hw/acpi/ghes.h
> @@ -61,6 +61,7 @@ typedef struct AcpiGhesState {
>      uint64_t hest_addr_le;
>      uint64_t hw_error_le;
>      bool present; /* True if GHES is present at all on this board */
> +    bool use_hest_addr; /* True if HEST address is present */
>  } AcpiGhesState;
>  
>  /*

an