From: Dave Hansen <dave@linux.vnet.ibm.com>
To: Pavel Machek <pavel@ucw.cz>
Cc: Greg KH <gregkh@suse.de>,
linux-kernel@vger.kernel.org, stable@kernel.org,
stable-review@kernel.org, torvalds@linux-foundation.org,
akpm@linux-foundation.org, alan@lxorguk.ukuu.org.uk,
Avi Kivity <avi@redhat.com>, Avi Kivity <avi@qumranet.com>
Subject: Re: [patch 14/48] KVM: Reduce kvm stack usage in kvm_arch_vm_ioctl()
Date: Sun, 06 Sep 2009 06:43:09 -0700 [thread overview]
Message-ID: <1252244589.14385.22402.camel@nimitz> (raw)
In-Reply-To: <20090906054749.GB1431@ucw.cz>
On Sun, 2009-09-06 at 07:47 +0200, Pavel Machek wrote:
> On Fri 2009-09-04 13:07:26, Greg KH wrote:
> > On my machine with gcc 3.4, kvm uses ~2k of stack in a few
> > select functions. This is mostly because gcc fails to
> > notice that the different case: statements could have their
> > stack usage combined. It overflows very nicely if interrupts
> > happen during one of these large uses.
> >
> > This patch uses two methods for reducing stack usage.
> > 1. dynamically allocate large objects instead of putting
> > on the stack.
> > 2. Use a union{} member for all of the case variables. This
> > tricks gcc into combining them all into a single stack
> > allocation. (There's also a comment on this)
>
> Are the 'reduce stack usage' patches suitable for stable? The rules
> said that fix must be for 'serious problem', not 'theoretical
> issue'...
I guess some context got dropped at some point. I was getting really
consistent oopses and goofy memory corruption when running KVM:
http://lkml.org/lkml/2008/3/25/340
Not theoretical at all. I think it cost me a few new gray hairs.
> > Signed-off-by: Dave Hansen <dave@linux.vnet.ibm.com>
> > Signed-off-by: Avi Kivity <avi@qumranet.com>
> > Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
> > ---
> > arch/x86/kvm/x86.c | 72 +++++++++++++++++++++++++++++++++--------------------
> > 1 file changed, 45 insertions(+), 27 deletions(-)
> >
> > --- a/arch/x86/kvm/x86.c
> > +++ b/arch/x86/kvm/x86.c
> > @@ -1630,6 +1630,15 @@ long kvm_arch_vm_ioctl(struct file *filp
> > struct kvm *kvm = filp->private_data;
> > void __user *argp = (void __user *)arg;
> > int r = -EINVAL;
> > + /*
> > + * This union makes it completely explicit to gcc-3.x
> > + * that these two variables' stack usage should be
> > + * combined, not added together.
> > + */
> > + union {
> > + struct kvm_pit_state ps;
> > + struct kvm_memory_alias alias;
> > + } u;
> >
> > switch (ioctl) {
> > case KVM_SET_TSS_ADDR:
>
> ...plus this is really ugly hack. Just declare the variable inside the
> case block that needs it?
Do we need to give that a better comment? It's explained a bit better
here:
http://lkml.org/lkml/2008/7/17/12
http://lkml.org/lkml/2008/7/17/16
Would this comment help?
/*
* gcc-3.x will sum the stack usage of two stack variables
* if they are declared in two different case blocks. This
* union makes it explicit that their stack space can be
* shared which greatly reduces stack usage.
*/
-- Dave
next prev parent reply other threads:[~2009-09-06 13:43 UTC|newest]
Thread overview: 56+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <20090904200712.724048145@mini.kroah.org>
2009-09-04 20:11 ` [patch 00/48] 2.6.27.32-stable review Greg KH
2009-09-04 20:07 ` [patch 01/48] ALSA: hda - Fix MacBookPro 3,1/4,1 quirk with ALC889A Greg KH
2009-09-04 20:07 ` [patch 02/48] clone(): fix race between copy_process() and de_thread() Greg KH
2009-09-04 20:07 ` [patch 03/48] ehea: Fix napi list corruption on ifconfig down Greg KH
2009-09-04 20:07 ` [patch 04/48] sound: pcm_lib: fix unsorted list constraint handling Greg KH
2009-09-04 20:07 ` [patch 05/48] SUNRPC: Fix rpc_task_force_reencode Greg KH
2009-09-04 20:07 ` [patch 06/48] KVM: VMX: Change cs reset state to be a data segment Greg KH
2009-09-04 20:07 ` [patch 07/48] KVM: VMX: Change segment dpl at reset to 3 Greg KH
2009-09-04 20:07 ` [patch 08/48] KVM: Load real mode segments correctly Greg KH
2009-09-04 20:07 ` [patch 09/48] KVM: Allocate guest memory as MAP_PRIVATE, not MAP_SHARED Greg KH
2009-09-04 20:07 ` [patch 10/48] KVM: Dont call get_user_pages(.force = 1) Greg KH
2009-09-04 20:07 ` [patch 11/48] KVM: MMU: Add locking around kvm_mmu_slot_remove_write_access() Greg KH
2009-09-04 20:07 ` [patch 12/48] KVM: MMU: Flush tlbs after clearing write permission when accessing dirty log Greg KH
2009-09-04 20:07 ` [patch 13/48] KVM: MMU: Fix setting the accessed bit on non-speculative sptes Greg KH
2009-09-04 20:07 ` [patch 14/48] KVM: Reduce kvm stack usage in kvm_arch_vm_ioctl() Greg KH
2009-09-06 5:47 ` Pavel Machek
2009-09-06 13:43 ` Dave Hansen [this message]
2009-09-04 20:07 ` [patch 15/48] KVM: Reduce stack usage in kvm_vcpu_ioctl() Greg KH
2009-09-04 20:07 ` [patch 16/48] KVM: Reduce stack usage in kvm_arch_vcpu_ioctl() Greg KH
2009-09-04 20:07 ` [patch 17/48] KVM: Reduce stack usage in kvm_pv_mmu_op() Greg KH
2009-09-04 20:07 ` [patch 18/48] KVM: add MC5_MISC msr read support Greg KH
2009-09-04 20:07 ` [patch 19/48] KVM: set debug registers after "schedulable" section Greg KH
2009-09-04 20:07 ` [patch 20/48] KVM: MMU: increase per-vcpu rmap cache alloc size Greg KH
2009-09-04 20:07 ` [patch 21/48] KVM: VMX: Set IGMT bit in EPT entry Greg KH
2009-09-04 20:07 ` [patch 22/48] KVM: Dont destroy vcpu in case vcpu_setup fails Greg KH
2009-09-04 20:07 ` [patch 23/48] KVM: VMX: Dont allow uninhibited access to EFER on i386 Greg KH
2009-09-04 20:07 ` [patch 24/48] KVM: SVM: Remove port 80 passthrough Greg KH
2009-09-04 20:07 ` [patch 25/48] KVM: Make EFER reads safe when EFER does not exist Greg KH
2009-09-04 20:07 ` [patch 26/48] KVM: VMX: Handle vmx instruction vmexits Greg KH
2009-09-04 20:07 ` [patch 27/48] KVM: Make paravirt tlb flush also reload the PAE PDPTRs Greg KH
2009-09-04 20:07 ` [patch 28/48] KVM: Fix PDPTR reloading on CR4 writes Greg KH
2009-09-04 20:07 ` [patch 29/48] KVM: MMU: do not free active mmu pages in free_mmu_pages() Greg KH
2009-09-04 20:07 ` [patch 30/48] KVM: Fix dirty bit tracking for slots with large pages Greg KH
2009-09-04 20:07 ` [patch 31/48] KVM: x86: check for cr3 validity in mmu_alloc_roots Greg KH
2009-09-04 20:07 ` [patch 32/48] KVM: MMU: protect kvm_mmu_change_mmu_pages with mmu_lock Greg KH
2009-09-04 20:07 ` [patch 33/48] appletalk: fix atalk_getname() leak Greg KH
2009-09-04 20:07 ` [patch 34/48] can: Fix raw_getname() leak Greg KH
2009-09-04 20:07 ` [patch 35/48] do_sigaltstack: avoid copying stack_t as a structure to user space Greg KH
2009-09-04 20:07 ` [patch 36/48] econet: Fix econet_getname() leak Greg KH
2009-09-04 20:07 ` [patch 37/48] irda: Fix irda_getname() leak Greg KH
2009-09-04 20:07 ` [patch 38/48] kthreads: fix kthread_create() vs kthread_stop() race Greg KH
2009-09-04 20:07 ` [patch 39/48] NET: llc, zero sockaddr_llc struct Greg KH
2009-09-04 20:07 ` [patch 40/48] netrom: Fix nr_getname() leak Greg KH
2009-09-04 20:07 ` [patch 41/48] ocfs2: Initialize the cluster were writing to in a non-sparse extend Greg KH
2009-09-04 21:00 ` Joel Becker
2009-09-06 8:32 ` Joel Becker
2009-09-06 18:17 ` Greg KH
2009-09-04 20:07 ` [patch 42/48] rose: Fix rose_getname() leak Greg KH
2009-09-04 20:07 ` [patch 43/48] ALSA: hda - Add missing vmaster initialization for ALC269 Greg KH
2009-09-04 20:07 ` [patch 44/48] parport: quickfix the proc registration bug Greg KH
2009-09-04 20:07 ` [patch 45/48] USB: removal of tty->low_latency hack dating back to the old serial code Greg KH
2009-09-04 20:07 ` [patch 46/48] Remove low_latency flag setting from nozomi and mxser drivers Greg KH
2009-09-04 20:07 ` [patch 47/48] SCSI: sr: report more accurate drive status after closing the tray Greg KH
2009-09-04 20:08 ` [patch 48/48] SUNRPC: Fix tcp reconnection Greg KH
2009-09-06 22:58 ` [patch 00/48] 2.6.27.32-stable review Tilman Schmidt
2009-09-10 22:33 ` Greg KH
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=1252244589.14385.22402.camel@nimitz \
--to=dave@linux.vnet.ibm.com \
--cc=akpm@linux-foundation.org \
--cc=alan@lxorguk.ukuu.org.uk \
--cc=avi@qumranet.com \
--cc=avi@redhat.com \
--cc=gregkh@suse.de \
--cc=linux-kernel@vger.kernel.org \
--cc=pavel@ucw.cz \
--cc=stable-review@kernel.org \
--cc=stable@kernel.org \
--cc=torvalds@linux-foundation.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