virtualization.lists.linux-foundation.org archive mirror
 help / color / mirror / Atom feed
From: Jeremy Fitzhardinge <jeremy@goop.org>
To: Andi Kleen <ak@suse.de>
Cc: Zachary Amsden <zach@vmware.com>,
	Xen-devel <xen-devel@lists.xensource.com>,
	Ian Pratt <ian.pratt@xensource.com>,
	lkml <linux-kernel@vger.kernel.org>,
	Chris Wright <chrisw@sous-sol.org>,
	virtualization@lists.osdl.org,
	Andrew Morton <akpm@linux-foundation.org>,
	Ulrich Drepper <drepper@redhat.com>,
	Roland McGrath <roland@redhat.com>,
	Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
Subject: [patch 02/28]xen: Add nosegneg capability to the vsyscall page notes
Date: Thu, 10 May 2007 17:06:45 -0700	[thread overview]
Message-ID: <20070511001156.003375000@goop.org> (raw)
In-Reply-To: 20070511000643.025196000@goop.org

[-- Attachment #1: xen-vsyscall-note.patch --]
[-- Type: text/plain, Size: 3061 bytes --]

Add the "nosegneg" fake capabilty to the vsyscall page notes. This is
used by the runtime linker to select a glibc version which then
disables negative-offset accesses to the thread-local segment via
%gs. These accesses require emulation in Xen (because segments are
truncated to protect the hypervisor address space) and avoiding them
provides a measurable performance boost.

Signed-off-by: Ian Pratt <ian.pratt@xensource.com>
Signed-off-by: Christian Limpach <Christian.Limpach@cl.cam.ac.uk>
Signed-off-by: Chris Wright <chrisw@sous-sol.org>
Signed-off-by: Jeremy Fitzhardinge <jeremy@xensource.com>
Acked-by: Zachary Amsden <zach@vmware.com>
Cc: Roland McGrath <roland@redhat.com>
Cc: Ulrich Drepper <drepper@redhat.com>

---
 arch/i386/kernel/vsyscall-note.S |   23 +++++++++++++++++++++++
 include/asm-i386/elf.h           |   14 ++++++++++----
 2 files changed, 33 insertions(+), 4 deletions(-)

===================================================================
--- a/arch/i386/kernel/vsyscall-note.S
+++ b/arch/i386/kernel/vsyscall-note.S
@@ -12,3 +12,26 @@ ELFNOTE_START(Linux, 0, "a")
 ELFNOTE_START(Linux, 0, "a")
 	.long LINUX_VERSION_CODE
 ELFNOTE_END
+
+#ifdef CONFIG_XEN
+#include <asm/elf.h>
+
+/*
+ * Add a special note telling glibc's dynamic linker a fake hardware
+ * flavor that it will use to choose the search path for libraries in the
+ * same way it uses real hardware capabilities like "mmx".
+ * We supply "nosegneg" as the fake capability, to indicate that we
+ * do not like negative offsets in instructions using segment overrides,
+ * since we implement those inefficiently.  This makes it possible to
+ * install libraries optimized to avoid those access patterns in someplace
+ * like /lib/i686/tls/nosegneg.  Note that an /etc/ld.so.conf.d/file
+ * corresponding to the bits here is needed to make ldconfig work right.
+ * It should contain:
+ *	hwcap 0 nosegneg
+ * to match the mapping of bit to name that we give here.
+ */
+ELFNOTE_START(GNU, 2, "a")
+	.long 1, 1<<VDSO_NOTE_NONEGSEG_BIT		/* ncaps, mask */
+	.byte VDSO_NOTE_NONEGSEG_BIT; .asciz "nosegneg"	/* bit, name */
+ELFNOTE_END
+#endif
===================================================================
--- a/include/asm-i386/elf.h
+++ b/include/asm-i386/elf.h
@@ -1,10 +1,12 @@
 #ifndef __ASMi386_ELF_H
 #define __ASMi386_ELF_H
 
-/*
- * ELF register definitions..
- */
+/* Bit used for the pseudo-hwcap for non-negative segments.  We use
+   bit 1 to avoid bugs in some versions of glibc when bit 0 is
+   used. */
+#define VDSO_NOTE_NONEGSEG_BIT	1
 
+#ifndef __ASSEMBLY__
 #include <asm/ptrace.h>
 #include <asm/user.h>
 #include <asm/auxvec.h>
@@ -24,6 +26,9 @@
 #define R_386_GOTPC	10
 #define R_386_NUM	11
 
+/*
+ * ELF register definitions..
+ */
 typedef unsigned long elf_greg_t;
 
 #define ELF_NGREG (sizeof (struct user_regs_struct) / sizeof(elf_greg_t))
@@ -160,6 +165,7 @@ do if (vdso_enabled) {							\
 		NEW_AUX_ENT(AT_SYSINFO_EHDR, VDSO_CURRENT_BASE);	\
 } while (0)
 
-#endif
+#endif	/* ARCH_DLINFO */
+#endif	/* __ASSEMBLY__ */
 
 #endif

-- 

  parent reply	other threads:[~2007-05-11  0:06 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-05-11  0:06 [patch 00/28]xen: Xen implementation for paravirt_ops Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 01/28]xen: Allocate and free vmalloc areas Jeremy Fitzhardinge
2007-05-11  0:06 ` Jeremy Fitzhardinge [this message]
2007-05-11 20:07   ` [patch 02/28]xen: Add nosegneg capability to the vsyscall page notes Roland McGrath
2007-05-11 20:25     ` Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 03/28]xen: Add Xen interface header files Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 04/28]xen: Core Xen implementation Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 05/28]xen: Xen virtual mmu Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 06/28]xen: xen event channels Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 07/28]xen: xen time implementation Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 08/28]xen: xen configuration Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 09/28]xen: Complete pagetable pinning for Xen Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 10/28]xen: ignore RW mapping of RO pages in pagetable_init Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 11/28]xen: fix multicall batching Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 12/28]xen: Account for time stolen by Xen Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 13/28]xen: Implement xen_sched_clock Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 14/28]xen: Xen SMP guest support Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 15/28]xen: Add support for preemption Jeremy Fitzhardinge
2007-05-11  0:06 ` [patch 16/28]xen: lazy-mmu operations Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 17/28]xen: deal with negative stolen time Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 18/28]xen: Use the hvc console infrastructure for Xen console Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 19/28]xen: Add early printk support via hvc console Jeremy Fitzhardinge
2007-05-12 19:08   ` [Xen-devel] " Bastian Blank
2007-05-12 19:57     ` Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 20/28]xen: Add Xen grant table support Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 21/28]xen: Add the Xenbus sysfs and virtual device hotplug driver Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 22/28]xen: Add Xen virtual block device driver Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 23/28]xen: rename xen netif_ structures to xen_netif_ Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 24/28]xen: Add the Xen virtual network device driver Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 25/28]xen: Xen machine operations Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 26/28]xen: handle external requests for shutdown, reboot and sysrq Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 27/28]xen: Place vcpu_info structure into per-cpu memory, if possible Jeremy Fitzhardinge
2007-05-11  0:07 ` [patch 28/28]xen: Attempt to patch inline versions of common operations Jeremy Fitzhardinge

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=20070511001156.003375000@goop.org \
    --to=jeremy@goop.org \
    --cc=Christian.Limpach@cl.cam.ac.uk \
    --cc=ak@suse.de \
    --cc=akpm@linux-foundation.org \
    --cc=chrisw@sous-sol.org \
    --cc=drepper@redhat.com \
    --cc=ian.pratt@xensource.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=roland@redhat.com \
    --cc=virtualization@lists.osdl.org \
    --cc=xen-devel@lists.xensource.com \
    --cc=zach@vmware.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).