Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: "Summers, Stuart" <stuart.summers@intel.com>
To: "intel-xe@lists.freedesktop.org" <intel-xe@lists.freedesktop.org>,
	"Vivekanandan,
	Balasubramani" <balasubramani.vivekanandan@intel.com>
Cc: "Roper, Matthew D" <matthew.d.roper@intel.com>
Subject: Re: [PATCH] drm/xe/device: Discard check for lmem_init
Date: Thu, 18 Dec 2025 20:57:34 +0000	[thread overview]
Message-ID: <a531d0f7bd469d3681c208e1e070afd73cd5df05.camel@intel.com> (raw)
In-Reply-To: <20251217125143.2430557-2-balasubramani.vivekanandan@intel.com>

On Wed, 2025-12-17 at 18:21 +0530, Balasubramani Vivekanandan wrote:
> Prior to lmem init check, driver is waiting for the pcode uncore_init
> status. uncore_init status will be asserted after the complete boot
> and
> initialization of the SoC by the pcode. uncore_init confirms that
> lmem
> init and mmio unblock has been already completed.
> It makes no sense to check for lmem init after the pcode uncore_init
> check. So it can be removed.

Matt has some good comments here. I'm also a little worried about
firmware bugs in early bring-up where we might go through the pcode
init but be missing the PCI config updates associated with lmem_init.
This at least guarantees that we're checking that here.

Thanks,
Stuart

> 
> Signed-off-by: Balasubramani Vivekanandan
> <balasubramani.vivekanandan@intel.com>
> ---
>  drivers/gpu/drm/xe/xe_device.c | 67 +++-----------------------------
> --
>  1 file changed, 5 insertions(+), 62 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_device.c
> b/drivers/gpu/drm/xe/xe_device.c
> index 1197f914ef77..3818d0cccb0e 100644
> --- a/drivers/gpu/drm/xe/xe_device.c
> +++ b/drivers/gpu/drm/xe/xe_device.c
> @@ -8,7 +8,6 @@
>  #include <linux/aperture.h>
>  #include <linux/delay.h>
>  #include <linux/fault-inject.h>
> -#include <linux/iopoll.h>
>  #include <linux/units.h>
>  
>  #include <drm/drm_atomic_helper.h>
> @@ -630,63 +629,6 @@ static int xe_set_dma_info(struct xe_device *xe)
>         return err;
>  }
>  
> -static int lmem_initializing(struct xe_device *xe)
> -{
> -       if (xe_mmio_read32(xe_root_tile_mmio(xe), GU_CNTL) &
> LMEM_INIT)
> -               return 0;
> -
> -       if (signal_pending(current))
> -               return -EINTR;
> -
> -       return 1;
> -}
> -
> -static int wait_for_lmem_ready(struct xe_device *xe)
> -{
> -       const unsigned long TIMEOUT_SEC = 60;
> -       unsigned long prev_jiffies;
> -       int initializing;
> -
> -       if (!IS_DGFX(xe))
> -               return 0;
> -
> -       if (IS_SRIOV_VF(xe))
> -               return 0;
> -
> -       if (!lmem_initializing(xe))
> -               return 0;
> -
> -       drm_dbg(&xe->drm, "Waiting for lmem initialization\n");
> -       prev_jiffies = jiffies;
> -
> -       /*
> -        * The boot firmware initializes local memory and
> -        * assesses its health. If memory training fails,
> -        * the punit will have been instructed to keep the GT powered
> -        * down.we won't be able to communicate with it
> -        *
> -        * If the status check is done before punit updates the
> register,
> -        * it can lead to the system being unusable.
> -        * use a timeout and defer the probe to prevent this.
> -        */
> -       poll_timeout_us(initializing = lmem_initializing(xe),
> -                       initializing <= 0,
> -                       20 * USEC_PER_MSEC, TIMEOUT_SEC *
> USEC_PER_SEC, true);
> -       if (initializing < 0)
> -               return initializing;
> -
> -       if (initializing) {
> -               drm_dbg(&xe->drm, "lmem not initialized by
> firmware\n");
> -               return -EPROBE_DEFER;
> -       }
> -
> -       drm_dbg(&xe->drm, "lmem ready after %ums",
> -               jiffies_to_msecs(jiffies - prev_jiffies));
> -
> -       return 0;
> -}
> -ALLOW_ERROR_INJECTION(wait_for_lmem_ready, ERRNO); /* See
> xe_pci_probe() */
> -
>  static void vf_update_device_info(struct xe_device *xe)
>  {
>         xe_assert(xe, IS_SRIOV_VF(xe));
> @@ -740,6 +682,11 @@ int xe_device_probe_early(struct xe_device *xe)
>         if (IS_SRIOV_VF(xe))
>                 vf_update_device_info(xe);
>  
> +       /*
> +        * Check for pcode uncore_init status to confirm if the SoC
> +        * initialization is complete. Until done, any MMIO or lmem
> access from
> +        * the driver will be blocked
> +        */
>         err = xe_pcode_probe_early(xe);
>         if (err || xe_survivability_mode_is_requested(xe)) {
>                 int save_err = err;
> @@ -756,10 +703,6 @@ int xe_device_probe_early(struct xe_device *xe)
>                 return save_err;
>         }
>  
> -       err = wait_for_lmem_ready(xe);
> -       if (err)
> -               return err;
> -
>         xe->wedged.mode = xe_modparam.wedged_mode;
>  
>         err = xe_device_vram_alloc(xe);


      parent reply	other threads:[~2025-12-18 20:57 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-12-17 12:51 [PATCH] drm/xe/device: Discard check for lmem_init Balasubramani Vivekanandan
2025-12-17 13:54 ` ✓ CI.KUnit: success for " Patchwork
2025-12-17 14:53 ` ✓ Xe.CI.BAT: " Patchwork
2025-12-17 23:12 ` [PATCH] " Matt Roper
2025-12-18 15:57   ` Ville Syrjälä
2025-12-19 14:38     ` Vivekanandan, Balasubramani
2025-12-19 14:27   ` Vivekanandan, Balasubramani
2025-12-19 16:42     ` Matt Roper
2025-12-18 11:42 ` ✗ Xe.CI.Full: failure for " Patchwork
2025-12-18 20:57 ` Summers, Stuart [this message]

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=a531d0f7bd469d3681c208e1e070afd73cd5df05.camel@intel.com \
    --to=stuart.summers@intel.com \
    --cc=balasubramani.vivekanandan@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=matthew.d.roper@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