All of lore.kernel.org
 help / color / mirror / Atom feed
From: Thomas Gleixner <tglx@linutronix.de>
To: LKML <linux-kernel@vger.kernel.org>
Cc: x86@kernel.org, John Stultz <john.stultz@linaro.org>,
	Vincenzo Frascino <vincenzo.frascino@arm.com>,
	Andy Lutomirski <luto@kernel.org>,
	Christophe Leroy <christophe.leroy@c-s.fr>,
	Paolo Bonzini <pbonzini@redhat.com>,
	Juergen Gross <jgross@suse.com>,
	Michael Kelley <mikelley@microsoft.com>,
	Sasha Levin <sashal@kernel.org>,
	Ralf Baechle <ralf@linux-mips.org>,
	Paul Burton <paulburton@kernel.org>,
	James Hogan <jhogan@kernel.org>,
	Russell King <linux@armlinux.org.uk>,
	Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>,
	Mark Rutland <mark.rutland@arm.com>,
	Marc Zyngier <maz@kernel.org>, Andrei Vagin <avagin@gmail.com>
Subject: [patch V2 16/17] lib/vdso: Allow architectures to override the ns shift operation
Date: Fri, 07 Feb 2020 13:39:03 +0100	[thread overview]
Message-ID: <20200207124403.857649978@linutronix.de> (raw)
In-Reply-To: 20200207123847.339896630@linutronix.de

From: Christophe Leroy <christophe.leroy@c-s.fr>

On powerpc/32, GCC (8.1) generates pretty bad code for the ns >>= vd->shift
operation taking into account that the shift is always <= 32 and the upper
part of the result is likely to be zero. GCC makes reversed assumptions
considering the shift to be likely >= 32 and the upper part to be like not
zero.

unsigned long long shift(unsigned long long x, unsigned char s)
{
	return x >> s;
}

results in:

00000018 <shift>:
  18:	35 25 ff e0 	addic.  r9,r5,-32
  1c:	41 80 00 10 	blt     2c <shift+0x14>
  20:	7c 64 4c 30 	srw     r4,r3,r9
  24:	38 60 00 00 	li      r3,0
  28:	4e 80 00 20 	blr
  2c:	54 69 08 3c 	rlwinm  r9,r3,1,0,30
  30:	21 45 00 1f 	subfic  r10,r5,31
  34:	7c 84 2c 30 	srw     r4,r4,r5
  38:	7d 29 50 30 	slw     r9,r9,r10
  3c:	7c 63 2c 30 	srw     r3,r3,r5
  40:	7d 24 23 78 	or      r4,r9,r4
  44:	4e 80 00 20 	blr

Even when forcing the shift to be smaller than 32 with an &= 31, it still
considers the shift as likely >= 32.

Move the default shift implementation into an inline which can be redefined
in architecture code via a macro.

[ tglx: Made the shift argument u32 and removed the __arch prefix ]

Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lore.kernel.org/r/b3d449de856982ed060a71e6ace8eeca4654e685.1580399657.git.christophe.leroy@c-s.fr

---
 lib/vdso/gettimeofday.c |   11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

--- a/lib/vdso/gettimeofday.c
+++ b/lib/vdso/gettimeofday.c
@@ -39,6 +39,13 @@ u64 vdso_calc_delta(u64 cycles, u64 last
 }
 #endif
 
+#ifndef vdso_shift_ns
+static __always_inline u64 vdso_shift_ns(u64 ns, u32 shift)
+{
+	return ns >> shift;
+}
+#endif
+
 #ifndef __arch_vdso_hres_capable
 static inline bool __arch_vdso_hres_capable(void)
 {
@@ -80,7 +87,7 @@ static int do_hres_timens(const struct v
 		ns = vdso_ts->nsec;
 		last = vd->cycle_last;
 		ns += vdso_calc_delta(cycles, last, vd->mask, vd->mult);
-		ns >>= vd->shift;
+		ns = vdso_shift_ns(ns, vd->shift);
 		sec = vdso_ts->sec;
 	} while (unlikely(vdso_read_retry(vd, seq)));
 
@@ -148,7 +155,7 @@ static __always_inline int do_hres(const
 		ns = vdso_ts->nsec;
 		last = vd->cycle_last;
 		ns += vdso_calc_delta(cycles, last, vd->mask, vd->mult);
-		ns >>= vd->shift;
+		ns = vdso_shift_ns(ns, vd->shift);
 		sec = vdso_ts->sec;
 	} while (unlikely(vdso_read_retry(vd, seq)));
 


  parent reply	other threads:[~2020-02-07 13:26 UTC|newest]

Thread overview: 63+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-02-07 12:38 [patch V2 00/17] VDSO consolidation Thomas Gleixner
2020-02-07 12:38 ` [patch V2 01/17] x86/vdso: Mark the TSC clocksource path likely Thomas Gleixner
2020-02-14 12:00   ` Vincenzo Frascino
2020-02-17 15:12   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-07 12:38 ` [patch V2 02/17] ARM: vdso: Remove unused function Thomas Gleixner
2020-02-14 10:21   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-07 12:38 ` [patch V2 03/17] lib/vdso: Allow the high resolution parts to be compiled out Thomas Gleixner
2020-02-14 11:54   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-07 12:38 ` [patch V2 04/17] ARM: vdso: Compile high resolution parts conditionally Thomas Gleixner
2020-02-14 11:55   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-07 12:38 ` [patch V2 05/17] MIPS: " Thomas Gleixner
2020-02-14 11:55   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-07 12:38 ` [patch V2 06/17] clocksource: Cleanup struct clocksource and documentation Thomas Gleixner
2020-02-14 11:57   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-07 12:38 ` [patch V2 07/17] x86/vdso: Move VDSO clocksource state tracking to callback Thomas Gleixner
2020-02-14 11:58   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-07 12:38 ` [patch V2 08/17] clocksource: Add common vdso clock mode storage Thomas Gleixner
2020-02-17 10:36   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-07 12:38 ` [patch V2 09/17] x86/vdso: Use generic VDSO " Thomas Gleixner
2020-02-14 10:32   ` Paolo Bonzini
2020-02-17 10:57   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-07 12:38 ` [patch V2 10/17] mips: vdso: " Thomas Gleixner
2020-02-17 10:52   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-17 19:18   ` tip-bot2 for Thomas Gleixner
2020-02-07 12:38 ` [patch V2 11/17] ARM/arm64: vdso: Use common vdso " Thomas Gleixner
2020-02-17 10:43   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-17 19:18   ` tip-bot2 for Thomas Gleixner
     [not found]   ` <CGME20200221115643eucas1p12ecb95c6161853a0e7dfe9207db079be@eucas1p1.samsung.com>
2020-02-21 11:56     ` [patch V2 11/17] " Marek Szyprowski
2020-02-21 11:56       ` Marek Szyprowski
2020-02-21 13:08       ` Vincenzo Frascino
2020-02-21 13:08         ` Vincenzo Frascino
2020-02-21 18:24       ` Vincenzo Frascino
2020-02-21 18:24         ` Vincenzo Frascino
2020-02-07 12:38 ` [patch V2 12/17] lib/vdso: Cleanup clock mode storage leftovers Thomas Gleixner
2020-02-17 11:04   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-17 19:18   ` tip-bot2 for Thomas Gleixner
2020-02-07 12:39 ` [patch V2 13/17] lib/vdso: Avoid highres update if clocksource is not VDSO capable Thomas Gleixner
2020-02-17 11:07   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-17 19:18   ` tip-bot2 for Thomas Gleixner
2020-02-07 12:39 ` [patch V2 14/17] lib/vdso: Move VCLOCK_TIMENS to vdso_clock_modes Thomas Gleixner
2020-02-17 11:12   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Thomas Gleixner
2020-02-17 19:18   ` tip-bot2 for Thomas Gleixner
2020-02-07 12:39 ` [patch V2 15/17] lib/vdso: Allow fixed clock mode Thomas Gleixner
2020-02-17 11:14   ` Vincenzo Frascino
2020-02-17 15:11   ` [tip: timers/core] " tip-bot2 for Christophe Leroy
2020-02-17 19:18   ` tip-bot2 for Christophe Leroy
2020-02-07 12:39 ` Thomas Gleixner [this message]
2020-02-17 11:15   ` [patch V2 16/17] lib/vdso: Allow architectures to override the ns shift operation Vincenzo Frascino
2020-02-07 12:39 ` [patch V2 17/17] lib/vdso: Allow architectures to provide the vdso data pointer Thomas Gleixner
2020-02-17 12:09   ` Vincenzo Frascino

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=20200207124403.857649978@linutronix.de \
    --to=tglx@linutronix.de \
    --cc=avagin@gmail.com \
    --cc=catalin.marinas@arm.com \
    --cc=christophe.leroy@c-s.fr \
    --cc=jgross@suse.com \
    --cc=jhogan@kernel.org \
    --cc=john.stultz@linaro.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux@armlinux.org.uk \
    --cc=luto@kernel.org \
    --cc=mark.rutland@arm.com \
    --cc=maz@kernel.org \
    --cc=mikelley@microsoft.com \
    --cc=paulburton@kernel.org \
    --cc=pbonzini@redhat.com \
    --cc=ralf@linux-mips.org \
    --cc=sashal@kernel.org \
    --cc=vincenzo.frascino@arm.com \
    --cc=will@kernel.org \
    --cc=x86@kernel.org \
    /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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.