Igt-dev Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
To: Lucas De Marchi <lucas.demarchi@intel.com>,
	<igt-dev@lists.freedesktop.org>
Subject: Re: [igt-dev] [i-g-t] intel-gfx-fw-info: Make it compatible with HuC via GSC
Date: Tue, 25 Apr 2023 18:44:05 +0530	[thread overview]
Message-ID: <ZEfSHauxt7nIOC5X@bvivekan-mobl> (raw)
In-Reply-To: <20230414173408.3584070-1-lucas.demarchi@intel.com>

On 14.04.2023 10:34, Lucas De Marchi wrote:
> When HuC is loaded via GSC, the firmware format is different and there
> is no information for kernel to parse except the version, that is in
> a different location.  Check for the magic field as the first dword and
> parse the blob differently based on that.
> 
> Tesetd with
> https://git.kernel.org/pub/scm/linux/kernel/git/firmware/linux-firmware.git/tree/i915/dg2_huc_gsc.bin?id=8f86b5ab3e051170ea240fc409d457e16e24bc21,
> with output "version: 7.10.3" as expected. Also checked with a couple
> of GuC firmware blobs and CSS-based HuC firmware blobs to guarantee
> this doesn't regress.
> 
> Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
> ---
>  tools/intel-gfx-fw-info | 85 +++++++++++++++++++++++++++++++----------
>  1 file changed, 65 insertions(+), 20 deletions(-)
> 
> diff --git a/tools/intel-gfx-fw-info b/tools/intel-gfx-fw-info
> index 77903bbb7..fef310a41 100755
> --- a/tools/intel-gfx-fw-info
> +++ b/tools/intel-gfx-fw-info
> @@ -59,6 +59,23 @@ struct uc_css_header {
>  	} rsvd;
>  	u32 header_info;
>  };
> +
> +#define HUC_GSC_VERSION_HI_DW		44
> +#define   HUC_GSC_MAJOR_VER_HI_MASK	(0xFF << 0)
> +#define   HUC_GSC_MINOR_VER_HI_MASK	(0xFF << 16)
> +#define HUC_GSC_VERSION_LO_DW		45
> +#define   HUC_GSC_PATCH_VER_LO_MASK	(0xFF << 0)
> +
> +// Add a fake definition for the GSC's header so this script can still
> +// check the version
> +
> +struct uc_huc_gsc_header {
> +	u32 raw[HUC_GSC_VERSION_LO_DW + 1];

I haven't tested how the raw dump looks like, but just want to check if
changing the struct as following would do any help in improving the raw dump
output?

struct uc_huc_gsc_header {
	u32 reserved[HUC_GSC_VERSION_HI_DW];
	u32 huc_ver_hi;
	u32 huc_ver_low;
};

Other than that, patch looks good.

Reviewed-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>

Regards,
Bala

> +};
> +
> +struct magic {
> +	char data[4];
> +};
>  """
>  
>  logging.basicConfig(format="%(levelname)s: %(message)s")
> @@ -83,26 +100,49 @@ def FIELD_GET(mask: int, value: int) -> int:
>      return (value & mask) >> ffs(mask)
>  
>  
> -def decode(fw) -> str:
> -    data = []
> +class Fw:
> +    def __init__(self, fw):
> +        self.fw = fw
> +
> +
> +class FwCss(Fw):
> +    def decode(self):
> +        data = []
> +
> +        CSS_SW_VERSION_UC_MAJOR = 0xFF << 16
> +        CSS_SW_VERSION_UC_MINOR = 0xFF << 8
> +        CSS_SW_VERSION_UC_PATCH = 0xFF
> +        major = FIELD_GET(CSS_SW_VERSION_UC_MAJOR, self.fw.sw_version)
> +        minor = FIELD_GET(CSS_SW_VERSION_UC_MINOR, self.fw.sw_version)
> +        patch = FIELD_GET(CSS_SW_VERSION_UC_PATCH, self.fw.sw_version)
> +        data += [f"version: {major}.{minor}.{patch}"]
> +
> +        CSS_DATE_DAY = 0xFF
> +        CSS_DATE_MONTH = 0xFF << 8
> +        CSS_DATE_YEAR = 0xFFFF << 16
> +        day = FIELD_GET(CSS_DATE_DAY, self.fw.date)
> +        month = FIELD_GET(CSS_DATE_MONTH, self.fw.date)
> +        year = FIELD_GET(CSS_DATE_YEAR, self.fw.date)
> +        data += [f"date: {year:02x}-{month:02x}-{day:02x}"]
> +
> +        return data
> +
>  
> -    CSS_SW_VERSION_UC_MAJOR = 0xFF << 16
> -    CSS_SW_VERSION_UC_MINOR = 0xFF << 8
> -    CSS_SW_VERSION_UC_PATCH = 0xFF
> -    major = FIELD_GET(CSS_SW_VERSION_UC_MAJOR, fw.sw_version)
> -    minor = FIELD_GET(CSS_SW_VERSION_UC_MINOR, fw.sw_version)
> -    patch = FIELD_GET(CSS_SW_VERSION_UC_PATCH, fw.sw_version)
> -    data += [f"version: {major}.{minor}.{patch}"]
> +class FwGsc(Fw):
> +    def decode(self):
> +        data = []
>  
> -    CSS_DATE_DAY = 0xFF
> -    CSS_DATE_MONTH = 0xFF << 8
> -    CSS_DATE_YEAR = 0xFFFF << 16
> -    day = FIELD_GET(CSS_DATE_DAY, fw.date)
> -    month = FIELD_GET(CSS_DATE_MONTH, fw.date)
> -    year = FIELD_GET(CSS_DATE_YEAR, fw.date)
> -    data += [f"date: {year:02x}-{month:02x}-{day:02x}"]
> +        HUC_GSC_VERSION_HI_DW = 44
> +        HUC_GSC_MINOR_VER_HI_MASK = 0xFF << 16
> +        HUC_GSC_MAJOR_VER_HI_MASK = 0xFF
> +        HUC_GSC_VERSION_LO_DW = 45
> +        HUC_GSC_PATCH_VER_LO_MASK = 0xFF
> +        major = FIELD_GET(HUC_GSC_MAJOR_VER_HI_MASK, self.fw.raw[HUC_GSC_VERSION_HI_DW])
> +        minor = FIELD_GET(HUC_GSC_MINOR_VER_HI_MASK, self.fw.raw[HUC_GSC_VERSION_HI_DW])
> +        patch = FIELD_GET(HUC_GSC_PATCH_VER_LO_MASK, self.fw.raw[HUC_GSC_VERSION_LO_DW])
> +        data += [f"version: {major}.{minor}.{patch}"]
>  
> -    return data
> +        return data
>  
>  
>  def parse_args(argv: typing.List[str]) -> argparse.Namespace:
> @@ -122,14 +162,19 @@ def main(argv: typing.List[str]) -> int:
>  
>      try:
>          with open(args.filename, mode="rb") as f:
> -            fw = cparser.uc_css_header(f)
> +            magic = cparser.magic(f)
> +            f.seek(0, 0)
> +            if magic.data == b"$CPD":
> +                fw = FwGsc(cparser.uc_huc_gsc_header(f))
> +            else:
> +                fw = FwCss(cparser.uc_css_header(f))
>      except FileNotFoundError as e:
>          logging.fatal(e)
>          return 1
>  
> -    print(*decode(fw), sep="\n")
> +    print(*fw.decode(), sep="\n")
>      print("raw dump:", end="")
> -    cstruct.dumpstruct(fw, color=sys.stdout.isatty())
> +    cstruct.dumpstruct(fw.fw, color=sys.stdout.isatty())
>  
>      return 0
>  

  parent reply	other threads:[~2023-04-25 13:14 UTC|newest]

Thread overview: 8+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-04-14 17:34 [igt-dev] [PATCH i-g-t] intel-gfx-fw-info: Make it compatible with HuC via GSC Lucas De Marchi
2023-04-14 18:11 ` [igt-dev] ✓ Fi.CI.BAT: success for " Patchwork
2023-04-15  1:22 ` [igt-dev] ✗ Fi.CI.IGT: failure " Patchwork
2023-04-25 13:14 ` Balasubramani Vivekanandan [this message]
2023-04-25 13:24   ` [igt-dev] [i-g-t] " Lucas De Marchi
2023-04-28  2:38     ` Ceraolo Spurio, Daniele
2023-04-25 16:12 ` [igt-dev] [PATCH i-g-t] " Kamil Konieczny
2023-04-25 19:26   ` Lucas De Marchi

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=ZEfSHauxt7nIOC5X@bvivekan-mobl \
    --to=balasubramani.vivekanandan@intel.com \
    --cc=igt-dev@lists.freedesktop.org \
    --cc=lucas.demarchi@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox