From: Tvrtko Ursulin <tursulin@ursulin.net>
To: Intel-gfx@lists.freedesktop.org
Cc: Jani Nikula <jani.nikula@intel.com>,
Michal Wajdeczko <michal.wajdeczko@linux.intel.com>
Subject: [PATCH v2] drm/i915: Number the platform enum strategically
Date: Thu, 8 Dec 2016 13:38:10 +0000 [thread overview]
Message-ID: <1481204290-13226-1-git-send-email-tvrtko.ursulin@linux.intel.com> (raw)
In-Reply-To: <20161208100401.GJ4815@nuc-i3427.alporthouse.com>
From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
If we use only power of two values for the platform enum
values we can let the compiler optimize some common
checks to a single conditional.
For example code like this:
if (IS_VALLEYVIEW(dev_priv) || IS_CHERRYVIEW(dev_priv))
Goes from this:
5c3c5: 8b 83 d8 06 00 00 mov 0x6d8(%rbx),%eax
5c3cb: 83 f8 12 cmp $0x12,%eax
5c3ce: 0f 84 f3 00 00 00 je 5c4c7 <fw_domain_init+0x1a7>
5c3d4: 83 f8 15 cmp $0x15,%eax
5c3d7: 0f 84 ea 00 00 00 je 5c4c7 <fw_domain_init+0x1a7>
To this:
5c1d5: f7 83 d8 06 00 00 00 testl $0x240000,0x6d8(%rbx)
5c1dc: 00 24 00
5c1df: 0f 85 da 00 00 00 jne 5c2bf <fw_domain_init+0x18f>
It is not much but there is value in this that as long as we
have to have conditions like the above sprinkled troughout the
code, we can at least have the generate binary a bit smarter.
Until we get to more than 32 platforms there is no downside
to this approach.
v2:
* Remove !! from the comparison. (Chris Wilson)
* Simplify the patch by moving the BIT() to the generator.
(Michal Wajdeczko)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@linux.intel.com>
---
drivers/gpu/drm/i915/i915_drv.h | 4 ++--
drivers/gpu/drm/i915/intel_device_info.c | 10 ++++++----
2 files changed, 8 insertions(+), 6 deletions(-)
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index fb8f4b7cd1ae..3be6a19f22eb 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -826,7 +826,7 @@ static inline unsigned int sseu_subslice_total(const struct sseu_dev_info *sseu)
}
enum intel_platform {
-#define i915_platform(name, value) INTEL_##name = value,
+#define i915_platform(name, value) INTEL_##name = BIT(value),
#include "i915_platforms.h"
#undef i915_platform
};
@@ -2590,7 +2590,7 @@ intel_info(const struct drm_i915_private *dev_priv)
#define i915_platform(name, id) \
static inline bool IS_##name(const struct drm_i915_private *dev_priv) \
{ \
- return (dev_priv)->info.platform == INTEL_##name; \
+ return (dev_priv)->info.platform & INTEL_##name; \
}
#include "i915_platforms.h"
#undef i915_platform
diff --git a/drivers/gpu/drm/i915/intel_device_info.c b/drivers/gpu/drm/i915/intel_device_info.c
index 5192d388d10e..26df6363e265 100644
--- a/drivers/gpu/drm/i915/intel_device_info.c
+++ b/drivers/gpu/drm/i915/intel_device_info.c
@@ -25,18 +25,20 @@
#include "i915_drv.h"
static const char * const platform_names[] = {
-#define i915_platform(name, id) [id] = #name,
+#define i915_platform(name, id) [__builtin_ffs(id)] = #name,
#include "i915_platforms.h"
#undef i915_platform
};
const char *intel_platform_name(enum intel_platform platform)
{
- if (WARN_ON_ONCE(platform >= ARRAY_SIZE(platform_names) ||
- platform_names[platform] == NULL))
+ unsigned int idx = ffs(platform);
+
+ if (WARN_ON_ONCE(idx >= ARRAY_SIZE(platform_names) ||
+ platform_names[idx] == NULL))
return "<unknown>";
- return platform_names[platform];
+ return platform_names[idx];
}
void intel_device_info_dump(struct drm_i915_private *dev_priv)
--
2.7.4
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx
next prev parent reply other threads:[~2016-12-08 13:38 UTC|newest]
Thread overview: 18+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-12-08 9:49 [PATCH 0/3] Claw back the IS_<platform> optimisation Tvrtko Ursulin
2016-12-08 9:49 ` [PATCH 1/3] drm/i915: Introduct i915_platforms.h Tvrtko Ursulin
2016-12-08 10:41 ` Jani Nikula
2016-12-08 13:16 ` Joonas Lahtinen
2016-12-08 13:21 ` Tvrtko Ursulin
2016-12-08 9:49 ` [PATCH 2/3] drm/i915: Generate all IS_<platform> macros Tvrtko Ursulin
2016-12-08 10:46 ` Jani Nikula
2016-12-08 13:26 ` Tvrtko Ursulin
2016-12-08 13:37 ` Jani Nikula
2016-12-08 13:40 ` Jani Nikula
2016-12-08 13:42 ` Tvrtko Ursulin
2016-12-08 14:00 ` Jani Nikula
2016-12-08 14:10 ` Tvrtko Ursulin
2016-12-08 9:49 ` [PATCH 3/3] drm/i915: Number the platform enum strategically Tvrtko Ursulin
2016-12-08 10:04 ` Chris Wilson
2016-12-08 13:38 ` Tvrtko Ursulin [this message]
2016-12-08 11:54 ` ✓ Fi.CI.BAT: success for Claw back the IS_<platform> optimisation Patchwork
2016-12-08 15:15 ` ✗ Fi.CI.BAT: warning for Claw back the IS_<platform> optimisation (rev2) 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=1481204290-13226-1-git-send-email-tvrtko.ursulin@linux.intel.com \
--to=tursulin@ursulin.net \
--cc=Intel-gfx@lists.freedesktop.org \
--cc=jani.nikula@intel.com \
--cc=michal.wajdeczko@linux.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