Add nosegneg capability to the vsyscall page notes

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>

1 files changed, 28 insertions, 0 deletions

diff --git a/arch/i386/kernel/vsyscall-note.S b/arch/i386/kernel/vsyscall-note.S
index 52e0cbbac70c..271f16a8ca01 100644
--- a/arch/i386/kernel/vsyscall-note.S
+++ b/arch/i386/kernel/vsyscall-note.S
@@ -12,3 +12,31 @@
 ELFNOTE_START(Linux, 0, "a")
         .long LINUX_VERSION_CODE
 ELFNOTE_END
+
+#ifdef CONFIG_XEN
+
+/*
+ * 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 1 nosegneg
+ * to match the mapping of bit to name that we give here.
+ */
+
+/* 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; the choice is otherwise arbitrary. */
+#define VDSO_NOTE_NONEGSEG_BIT  1
+
+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