From: Rodrigo Vivi <rodrigo.vivi@intel.com>
To: Mahesh Kumar <mahesh1.kumar@intel.com>
Cc: intel-gfx@lists.freedesktop.org
Subject: Re: [PATCH 4/5] drm/i915/kbl+: Enable IPC only for symmetric memory configurations
Date: Fri, 17 Aug 2018 11:20:20 -0700 [thread overview]
Message-ID: <20180817182020.GC8880@intel.com> (raw)
In-Reply-To: <20180726141410.2185-5-mahesh1.kumar@intel.com>
On Thu, Jul 26, 2018 at 07:44:09PM +0530, Mahesh Kumar wrote:
> IPC may cause underflows if not used with dual channel symmetric
> memory configuration. Disable IPC for non symmetric configurations in
> affected platforms.
> Display WA #1141
>
> Signed-off-by: Mahesh Kumar <mahesh1.kumar@intel.com>
> ---
> drivers/gpu/drm/i915/i915_drv.c | 43 ++++++++++++++++++++++++++++++++++++-----
> drivers/gpu/drm/i915/i915_drv.h | 1 +
> drivers/gpu/drm/i915/intel_pm.c | 2 +-
> 3 files changed, 40 insertions(+), 6 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
> index 86bc2e685522..2273664166bc 100644
> --- a/drivers/gpu/drm/i915/i915_drv.c
> +++ b/drivers/gpu/drm/i915/i915_drv.c
> @@ -1141,21 +1141,47 @@ skl_dram_get_channel_info(struct dram_channel_info *ch, u32 val)
> return 0;
> }
>
> +static bool
> +intel_is_dram_ipc_capable(struct drm_i915_private *dev_priv,
> + u32 val_ch0, u32 val_ch1,
> + struct dram_channel_info *ch0)
what about
intel_is_dram_symmetric() ?
> +{
> + /* Display WA #1141: SKL:all KBL:all CNL:A CNL:B */
move this to the wa call, not the the function check...
> + if (INTEL_GEN(dev_priv) > 9 &&
> + !IS_CNL_REVID(dev_priv, CNL_REVID_A0, CNL_REVID_B0))
please don't add CNL pre-prod wa
> + return true;
> +
> + if (!IS_KABYLAKE(dev_priv) && !IS_SKYLAKE(dev_priv))
> + return true;
actually remove all platforms checks here...
> +
> + if (val_ch0 != val_ch1)
> + return false;
> +
> + if (ch0->s_info.size == 0)
> + return true;
> + if (ch0->l_info.size == ch0->s_info.size &&
> + ch0->l_info.width == ch0->s_info.width &&
> + ch0->l_info.rank == ch0->s_info.rank)
> + return true;
> +
> + return false;
return (val_ch0 == val_ch1 && (ch0->s_info.size == 0 ||
(ch0->l_info.size == ch0->s_info.size &&
ch0->l_info.width == ch0->s_info.width &&
ch0->l_info.rank == ch0->s_info.rank)))
> +}
> +
> static int
> skl_dram_get_channels_info(struct drm_i915_private *dev_priv)
> {
> struct dram_info *dram_info = &dev_priv->dram_info;
> struct dram_channel_info ch0, ch1;
> - u32 val;
> + u32 val_ch0, val_ch1;
> int ret;
>
> - val = I915_READ(SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
> - ret = skl_dram_get_channel_info(&ch0, val);
> + val_ch0 = I915_READ(SKL_MAD_DIMM_CH0_0_0_0_MCHBAR_MCMAIN);
> + ret = skl_dram_get_channel_info(&ch0, val_ch0);
> if (ret == 0)
> dram_info->num_channels++;
>
> - val = I915_READ(SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
> - ret = skl_dram_get_channel_info(&ch1, val);
> + val_ch1 = I915_READ(SKL_MAD_DIMM_CH1_0_0_0_MCHBAR_MCMAIN);
> + ret = skl_dram_get_channel_info(&ch1, val_ch1);
> if (ret == 0)
> dram_info->num_channels++;
>
> @@ -1185,6 +1211,13 @@ skl_dram_get_channels_info(struct drm_i915_private *dev_priv)
> if (ch0.is_16gb_dimm || ch1.is_16gb_dimm)
> dram_info->is_16gb_dimm = true;
>
> + if (intel_is_dram_ipc_capable(dev_priv, val_ch0, val_ch1, &ch0))
> + dev_priv->ipc_capable_mem = true;
> + else
> + dev_priv->ipc_capable_mem = false;
> +
> + DRM_DEBUG_KMS("memory configuration is %sIPC capable\n",
> + dev_priv->ipc_capable_mem ? "" : "not ");
> return 0;
> }
>
> diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
> index 854f3c828e01..036d6554c017 100644
> --- a/drivers/gpu/drm/i915/i915_drv.h
> +++ b/drivers/gpu/drm/i915/i915_drv.h
> @@ -2112,6 +2112,7 @@ struct drm_i915_private {
> bool chv_phy_assert[2];
>
> bool ipc_enabled;
> + bool ipc_capable_mem;
I don't think we need to stage this...
>
> /* Used to save the pipe-to-encoder mapping for audio */
> struct intel_encoder *av_enc_map[I915_MAX_PIPES];
> diff --git a/drivers/gpu/drm/i915/intel_pm.c b/drivers/gpu/drm/i915/intel_pm.c
> index 2446f53adf21..39e400d5f555 100644
> --- a/drivers/gpu/drm/i915/intel_pm.c
> +++ b/drivers/gpu/drm/i915/intel_pm.c
> @@ -6097,7 +6097,7 @@ void intel_enable_ipc(struct drm_i915_private *dev_priv)
> u32 val;
>
> /* Display WA #0477 WaDisableIPC: skl */
> - if (IS_SKYLAKE(dev_priv)) {
> + if (IS_SKYLAKE(dev_priv) || !dev_priv->ipc_capable_mem) {
> dev_priv->ipc_enabled = false;
This is not the WA 1141 for other platforms than SKL. Please only keep skl here.
For the other WA add 4us across all watermark levels
/* Display WA #1141: skl,kbl,cfl */
if ((IS_KABYLAKE(dev_priv) || IS_COFFEELAKE(dev_priv)) &&
intel_is_dram_symmetric())
levels += 4;
> return;
> }
> --
> 2.16.2
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2018-08-17 18:20 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-07-26 14:14 [PATCH 0/5] Decode memdev info and bandwidth and implemnt latency WA Mahesh Kumar
2018-07-26 14:14 ` [PATCH 1/5] drm/i915/bxt: Decode memory bandwidth and parameters Mahesh Kumar
2018-08-16 22:29 ` Rodrigo Vivi
2018-08-17 17:43 ` Rodrigo Vivi
2018-08-21 9:53 ` Kumar, Mahesh
2018-08-21 15:12 ` Rodrigo Vivi
2018-07-26 14:14 ` [PATCH 2/5] drm/i915/skl+: " Mahesh Kumar
2018-08-16 22:35 ` Rodrigo Vivi
2018-08-21 10:21 ` Kumar, Mahesh
2018-07-26 14:14 ` [PATCH 3/5] drm/i915: Implement 16GB dimm wa for latency level-0 Mahesh Kumar
2018-07-27 3:51 ` Matt Turner
2018-07-27 6:10 ` Kumar, Mahesh
2018-07-28 5:48 ` Rodrigo Vivi
2018-07-31 14:18 ` Kumar, Mahesh
2018-08-17 17:57 ` Rodrigo Vivi
2018-07-26 14:14 ` [PATCH 4/5] drm/i915/kbl+: Enable IPC only for symmetric memory configurations Mahesh Kumar
2018-08-17 18:20 ` Rodrigo Vivi [this message]
2018-08-21 14:57 ` Kumar, Mahesh
2018-08-21 16:00 ` Kumar, Mahesh
2018-08-21 18:56 ` Rodrigo Vivi
2018-08-22 13:02 ` Kumar, Mahesh
2018-08-22 15:39 ` Rodrigo Vivi
2018-07-26 14:14 ` [PATCH 5/5] drm/i915/skl+: don't trust IPC value set by BIOS Mahesh Kumar
2018-07-26 14:29 ` ✗ Fi.CI.SPARSE: warning for Decode memdev info and bandwidth and implemnt latency WA (rev2) Patchwork
2018-07-26 14:48 ` ✓ Fi.CI.BAT: success " Patchwork
2018-07-26 16:09 ` ✓ Fi.CI.IGT: " Patchwork
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=20180817182020.GC8880@intel.com \
--to=rodrigo.vivi@intel.com \
--cc=intel-gfx@lists.freedesktop.org \
--cc=mahesh1.kumar@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;
as well as URLs for NNTP newsgroup(s).