rpms/kernel
3
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

Add patch to fix loop in VDSO (rhbz 1178975)

Browse Source
This commit is contained in:
Josh Boyer 2015-01-12 08:17:01 -05:00
parent c60911c994
commit fe651573b7
2 changed files with 64 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
kernel.spec
View File

@ -614,6 +614,9 @@ Patch26092: xhci-Add-broken-streams-quirk-for-Fresco-Logic-FL100.patch
Patch26093: uas-Add-US_FL_NO_ATA_1X-for-Seagate-devices-with-usb.patch
Patch26094: uas-Add-US_FL_NO_REPORT_OPCODES-for-JMicron-JMS566-w.patch
#rhbz 1178975
Patch26125: x86-vdso-Use-asm-volatile-in-__getcpu.patch
# git clone ssh://git.fedorahosted.org/git/kernel-arm64.git, git diff master...devel
Patch30000: kernel-arm64.patch
@ -1336,6 +1339,9 @@ ApplyPatch xhci-Add-broken-streams-quirk-for-Fresco-Logic-FL100.patch
ApplyPatch uas-Add-US_FL_NO_ATA_1X-for-Seagate-devices-with-usb.patch
ApplyPatch uas-Add-US_FL_NO_REPORT_OPCODES-for-JMicron-JMS566-w.patch
#rhbz 1178975
ApplyPatch x86-vdso-Use-asm-volatile-in-__getcpu.patch
%if 0%{?aarch64patches}
ApplyPatch kernel-arm64.patch
%ifnarch aarch64 # this is stupid, but i want to notice before secondary koji does.
@ -2202,6 +2208,9 @@ fi
# ||----w |
# || ||
%changelog
* Mon Jan 12 2015 Josh Boyer <jwboyer@fedoraproject.org>
- Add patch to fix loop in VDSO (rhbz 1178975)
* Fri Jan 09 2015 Josh Boyer <jwboyer@fedoraproject.org> - 3.19.0-0.rc3.git2.1
- Linux v3.19-rc3-69-g11c8f01b423b

55
x86-vdso-Use-asm-volatile-in-__getcpu.patch Normal file
View File

@ -0,0 +1,55 @@
From 1ddf0b1b11aa8a90cef6706e935fc31c75c406ba Mon Sep 17 00:00:00 2001
From: Andy Lutomirski <luto@amacapital.net>
Date: Sun, 21 Dec 2014 08:57:46 -0800
Subject: x86, vdso: Use asm volatile in __getcpu
In Linux 3.18 and below, GCC hoists the lsl instructions in the
pvclock code all the way to the beginning of __vdso_clock_gettime,
slowing the non-paravirt case significantly. For unknown reasons,
presumably related to the removal of a branch, the performance issue
is gone as of
e76b027e6408 x86,vdso: Use LSL unconditionally for vgetcpu
but I don't trust GCC enough to expect the problem to stay fixed.
There should be no correctness issue, because the __getcpu calls in
__vdso_vlock_gettime were never necessary in the first place.
Note to stable maintainers: In 3.18 and below, depending on
configuration, gcc 4.9.2 generates code like this:
9c3: 44 0f 03 e8 lsl %ax,%r13d
9c7: 45 89 eb mov %r13d,%r11d
9ca: 0f 03 d8 lsl %ax,%ebx
This patch won't apply as is to any released kernel, but I'll send a
trivial backported version if needed.
Fixes: 51c19b4f5927 x86: vdso: pvclock gettime support
Cc: stable@vger.kernel.org # 3.8+
Cc: Marcelo Tosatti <mtosatti@redhat.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Andy Lutomirski <luto@amacapital.net>
diff --git a/arch/x86/include/asm/vgtod.h b/arch/x86/include/asm/vgtod.h
index e7e9682..f556c48 100644
--- a/arch/x86/include/asm/vgtod.h
+++ b/arch/x86/include/asm/vgtod.h
@@ -80,9 +80,11 @@ static inline unsigned int __getcpu(void)
/*
* Load per CPU data from GDT. LSL is faster than RDTSCP and
- * works on all CPUs.
+ * works on all CPUs. This is volatile so that it orders
+ * correctly wrt barrier() and to keep gcc from cleverly
+ * hoisting it out of the calling function.
*/
- asm("lsl %1,%0" : "=r" (p) : "r" (__PER_CPU_SEG));
+ asm volatile ("lsl %1,%0" : "=r" (p) : "r" (__PER_CPU_SEG));
return p;
}
--
cgit v0.10.2