mirror of https://lore.kernel.org/lkml/
 help / color / mirror / Atom feed
From: Brian Gerst <brgerst@gmail.com>
To: Tejun Heo <tj@kernel.org>
Cc: Ingo Molnar <mingo@elte.hu>,
	linux-kernel@vger.kernel.org, Brian Gerst <brgerst@gmail.com>
Subject: [PATCH 3/5] x86-64: Rework __per_cpu_load adjustments
Date: Sun, 18 Jan 2009 19:52:23 -0500	[thread overview]
Message-ID: <1232326345-3534-3-git-send-email-brgerst@gmail.com> (raw)
In-Reply-To: <1232326345-3534-2-git-send-email-brgerst@gmail.com>

Use cpu_number to determine if the adjustment is necessary.

Signed-off-by: Brian Gerst <brgerst@gmail.com>
---
 arch/x86/kernel/head_64.S |   24 ++++++++----------------
 1 files changed, 8 insertions(+), 16 deletions(-)

diff --git a/arch/x86/kernel/head_64.S b/arch/x86/kernel/head_64.S
index c8ace88..98ea26a 100644
--- a/arch/x86/kernel/head_64.S
+++ b/arch/x86/kernel/head_64.S
@@ -207,19 +207,15 @@ ENTRY(secondary_startup_64)
 
 #ifdef CONFIG_SMP
 	/*
-	 * early_gdt_base should point to the gdt_page in static percpu init
-	 * data area.  Computing this requires two symbols - __per_cpu_load
-	 * and per_cpu__gdt_page.  As linker can't do no such relocation, do
-	 * it by hand.  As early_gdt_descr is manipulated by C code for
-	 * secondary CPUs, this should be done only once for the boot CPU
-	 * when early_gdt_descr_base contains zero.
+	 * Fix up static pointers that need __per_cpu_load added.  The assembler
+	 * is unable to do this directly.  This is only needed for the boot cpu.
+	 * These values are set up with the correct base addresses by C code for
+	 * secondary cpus.
 	 */
-	movq	early_gdt_descr_base(%rip), %rax
-	testq	%rax, %rax
-	jnz	1f
-	movq	$__per_cpu_load, %rax
-	addq	$per_cpu__gdt_page, %rax
-	movq	%rax, early_gdt_descr_base(%rip)
+	movq	initial_gs(%rip), %rax
+	cmpl	$0, per_cpu__cpu_number(%rax)
+	jne	1f
+	addq	%rax, early_gdt_descr_base(%rip)
 1:
 #endif
 	/*
@@ -431,12 +427,8 @@ NEXT_PAGE(level2_spare_pgt)
 	.globl early_gdt_descr
 early_gdt_descr:
 	.word	GDT_ENTRIES*8-1
-#ifdef CONFIG_SMP
 early_gdt_descr_base:
-	.quad   0x0000000000000000
-#else
 	.quad	per_cpu__gdt_page
-#endif
 
 ENTRY(phys_base)
 	/* This must match the first entry in level2_kernel_pgt */
-- 
1.6.1.rc1


  reply	other threads:[~2009-01-19  0:53 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-01-18 16:39 [PATCHSET x86:core/percpu] percpu: move PDA fields to percpu Tejun Heo
2009-01-18 16:39 ` [PATCH 01/12] x86-64: Move irq stats from PDA to per-cpu and consolidate with 32-bit Tejun Heo
2009-01-18 16:39 ` [PATCH 02/12] x86-64: Move TLB state " Tejun Heo
2009-01-18 16:39 ` [PATCH 03/12] x86-64: Convert irqstacks to per-cpu Tejun Heo
2009-01-18 18:16   ` Brian Gerst
2009-01-18 23:38     ` Tejun Heo
2009-01-18 23:43       ` Ingo Molnar
2009-01-19  0:52         ` Tejun Heo
2009-01-18 16:39 ` [PATCH 04/12] x86-64: Convert exception stacks " Tejun Heo
2009-01-18 16:39 ` [PATCH 05/12] x86-64: Move cpu number from PDA to per-cpu and consolidate with 32-bit Tejun Heo
2009-01-18 16:39 ` [PATCH 06/12] x86-64: Move current task " Tejun Heo
2009-01-18 16:39 ` [PATCH 07/12] x86-64: Move kernelstack from PDA to per-cpu Tejun Heo
2009-01-18 16:39 ` [PATCH 08/12] x86-64: Move oldrsp " Tejun Heo
2009-01-18 16:39 ` [PATCH 09/12] x86-64: Move irqcount " Tejun Heo
2009-01-18 16:39 ` [PATCH 10/12] x86-64: Move nodenumber " Tejun Heo
2009-01-18 16:39 ` [PATCH 11/12] x86-64: Move isidle " Tejun Heo
2009-01-18 16:39 ` [PATCH 12/12] x86-64: Use absolute displacements for per-cpu accesses Tejun Heo
2009-01-18 16:49 ` [PATCHSET x86:core/percpu] percpu: move PDA fields to percpu Ingo Molnar
2009-01-19  0:51 ` Brian Gerst
2009-01-19  0:52   ` [PATCH 1/5] x86-64: Remove pda_init() Brian Gerst
2009-01-19  0:52     ` [PATCH 2/5] percpu: Refactor percpu.h Brian Gerst
2009-01-19  0:52       ` Brian Gerst [this message]
2009-01-19  0:52         ` [PATCH 4/5] x86-64: Remove the PDA Brian Gerst
2009-01-19  0:52           ` [PATCH 5/5] x86-64: Remove pda.h Brian Gerst
2009-01-19  2:46             ` [PATCH 6/6] linker script: kill PERCPU_VADDR_PREALLOC() Tejun Heo
2009-01-19  2:18           ` [PATCH 4/5] x86-64: Remove the PDA Tejun Heo
2009-01-19  2:52             ` Brian Gerst
2009-01-19  3:05               ` Tejun Heo
2009-01-19  2:46           ` [PATCH UPDATED 4/5] x86-64: Move stack_canary into irq_stack Tejun Heo
2009-02-05 22:30       ` [PATCH 2/5] percpu: Refactor percpu.h Tony Luck
2009-02-06 11:11         ` Brian Gerst
2009-02-06 19:06           ` Luck, Tony
2009-01-19  9:53     ` [PATCH 1/5] x86-64: Remove pda_init() Peter Zijlstra
2009-01-19  2:45   ` [PATCHSET x86:core/percpu] percpu: move PDA fields to percpu Tejun Heo
2009-01-19 11:19     ` Ingo Molnar
2009-01-19 11:35       ` Ingo Molnar
2009-01-19 13:43         ` Tejun Heo
2009-01-20  3:34           ` Tejun Heo
2009-01-20  7:24             ` Ingo Molnar

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=1232326345-3534-3-git-send-email-brgerst@gmail.com \
    --to=brgerst@gmail.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@elte.hu \
    --cc=tj@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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox

Powered by JetHome