Don't conflate sysret and sysexit; they're different instructions with
different semantics, and may be in use at the same time (at least
within the same kernel, depending on whether its an Intel or AMD
system).
sysexit - just return to userspace, does no register restoration of
any kind; must explicitly atomically enable interrupts.
sysret - reloads flags from r11, so no need to explicitly enable
interrupts on 64-bit, responsible for restoring usermode %gs
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citirx.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
x86: use __KERNEL_DS as SS when returning to a kernel thread
This is needed when the kernel is running on RING3, such as under Xen.
x86_64 has a weird feature that makes it #GP on iret when SS is a null
descriptor.
This need to be tested on bare metal to make sure it doesn't cause any
problems. AMD specs say SS is always ignored (except on iret?).
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com> Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Because Xen doesn't support PSE mappings in guests, all code which
assumed the presence of PSE has been changed to fall back to smaller
mappings if necessary. As a result, PSE is optional rather than
required (though still used whereever possible).
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
x86, 64-bit: adjust mapping of physical pagetables to work with Xen
This makes a few of changes to the construction of the initial
pagetables to work better with paravirt_ops/Xen. The main areas
are:
1. Support non-PSE mapping of memory, since Xen doesn't currently
allow 2M pages to be mapped in guests.
2. Make sure that the ioremap alias of all pages are dropped before
attaching the new page to the pagetable. This avoids having
writable aliases of pagetable pages.
3. Preserve existing pagetable entries, rather than overwriting. Its
possible that a fair amount of pagetable has already been constructed,
so reuse what's already in place rather than ignoring and overwriting it.
The algorithm relies on the invariant that any page which is part of
the kernel pagetable is itself mapped in the linear memory area. This
way, it can avoid using ioremap on a pagetable page.
The invariant holds because it maps memory from low to high addresses,
and also allocates memory from low to high. Each allocated page can
map at least 2M of address space, so the mapped area will always
progress much faster than the allocated area. It relies on the early
boot code mapping enough pages to get started.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
The first essentially cleans up after head_64.S. It clears the
bss, zaps low identity mappings, sets up some early exception
handlers.
The second part preserves the boot data, reserves the kernel's
text/data/bss, pagetables and ramdisk, and then starts the kernel
proper.
This split is so that Xen can call the second part to do the set up it
needs done. It doesn't need any of the first part setups, because it
doesn't boot via head_64.S, and its redundant or actively damaging.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Eduardo Habkost [Wed, 25 Jun 2008 04:19:16 +0000 (00:19 -0400)]
paravirt/x86, 64-bit: move __PAGE_OFFSET to leave a space for hypervisor
Set __PAGE_OFFSET to the most negative possible address +
16*PGDIR_SIZE. The gap is to allow a space for a hypervisor to fit.
The gap is more or less arbitrary, but it's what Xen needs.
When booting native, kernel/head_64.S has a set of compile-time
generated pagetables used at boot time. This patch removes their
absolutely hard-coded layout, and makes it parameterised on
__PAGE_OFFSET (and __START_KERNEL_map).
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com> Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
x86/paravirt: define PARA_INDIRECT for indirect asm calls
On 32-bit it's best to use a %cs: prefix to access memory where the
other segments may not bet set up properly yet. On 64-bit it's best
to use a rip-relative addressing mode. Define PARA_INDIRECT() to
abstract this and generate the proper addressing mode in each case.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Jan Beulich points out that vmalloc_sync_all() assumes that the
kernel's pmd is always expected to be present in the pgd. The current
pgd construction code will add the pgd to the pgd_list before its pmds
have been pre-populated, thereby making it visible to
vmalloc_sync_all().
However, because pgd_prepopulate_pmd also does the allocation, it may
block and cannot be done under spinlock.
The solution is to preallocate the pmds out of the spinlock, then
populate them while holding the pgd_list lock.
This patch also pulls the pmd preallocation and mop-up functions out
to be common, assuming that the compiler will generate no code for
them when PREALLOCTED_PMDS is 0. Also, there's no need for pgd_ctor
to clear the pgd again, since it's allocated as a zeroed page.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu> Cc: Jan Beulich <jbeulich@novell.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Add hooks which are called at pgd_alloc/free time. The pgd_alloc hook
may return an error code, which if non-zero, causes the pgd allocation
to be failed. The hooks may be used to allocate/free auxillary
per-pgd information.
also fix:
> * Ingo Molnar <mingo@elte.hu> wrote:
>
> include/asm/pgalloc.h: In function ‘paravirt_pgd_free':
> include/asm/pgalloc.h:14: error: parameter name omitted
> arch/x86/kernel/entry_64.S: In file included from
> arch/x86/kernel/traps_64.c:51:include/asm/pgalloc.h: In function ‘paravirt_pgd_free':
> include/asm/pgalloc.h:14: error: parameter name omitted
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
vmalloc_sync_all() is only called from register_die_notifier and
alloc_vm_area. Neither is on any performance-critical paths, so
vmalloc_sync_all() itself is not on any hot paths.
Given that the optimisations in vmalloc_sync_all add a fair amount of
code and complexity, and are fairly hard to evaluate for correctness,
it's better to just remove them to simplify the code rather than worry
about its absolute performance.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Ingo Molnar [Thu, 26 Jun 2008 10:40:35 +0000 (12:40 +0200)]
x86: build fix
fix:
In file included from arch/x86/kernel/setup.c:118:
include/asm/highmem.h:64: error: expected identifier or ‘(' before ‘do'
include/asm/highmem.h:64: error: expected identifier or ‘(' before ‘while'
include/asm/highmem.h:67: error: expected identifier or ‘(' before ‘do'
include/asm/highmem.h:67: error: expected identifier or ‘(' before ‘while'
Bernhard Walle [Thu, 26 Jun 2008 07:23:21 +0000 (09:23 +0200)]
x86, crashdump, /proc/vmcore: remove CONFIG_EXPERIMENTAL from kdump
I would suggest to remove the "experimental" status from Kdump.
Kdump is now in the kernel since a long time and used by Enterprise
distributions. I don't think that "experimental" is true any more.
Signed-off-by: Bernhard Walle <bwalle@suse.de> Cc: vgoyal@redhat.com Cc: kexec@lists.infradead.org Cc: Bernhard Walle <bwalle@suse.de> Cc: akpm@linux-foundation.org Signed-off-by: Ingo Molnar <mingo@elte.hu>
Paul Jackson [Wed, 25 Jun 2008 12:44:46 +0000 (05:44 -0700)]
x86 boot: only pick up additional EFI memmap if add_efi_memmap flag
Applies on top of the previous patch:
x86 boot: add code to add BIOS provided EFI memory entries to kernel
Instead of always adding EFI memory map entries (if present) to the
memory map after initially finding either E820 BIOS memory map entries
and/or kernel command line memmap entries, -instead- only add such
additional EFI memory map entries if the kernel boot option:
add_efi_memmap
is specified.
Requiring this 'add_efi_memmap' option is backward compatible with
kernels that didn't load such additional EFI memory map entries in
the first place, and it doesn't override a configuration that tries
to replace all E820 or EFI BIOS memory map entries with ones given
entirely on the kernel command line.
Signed-off-by: Paul Jackson <pj@sgi.com> Cc: "Yinghai Lu" <yhlu.kernel@gmail.com> Cc: "Jack Steiner" <steiner@sgi.com> Cc: "Mike Travis" <travis@sgi.com> Cc: "Huang Cc: Ying" <ying.huang@intel.com> Cc: "Andi Kleen" <andi@firstfloor.org> Cc: "Andrew Morton" <akpm@linux-foundation.org> Cc: Paul Jackson <pj@sgi.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Paul Jackson [Wed, 25 Jun 2008 12:44:40 +0000 (05:44 -0700)]
mm, generic, x86 boot: more tweaks to hex prints of some pfn addresses
Fix some problems with (and applies on top of) a previous patch:
x86 boot: show pfn addresses in hex not decimal in some kernel info printks
Primarily change "0x%8lx" format, which displays with a right aligned
space filled hex number (spaces between the "0x" prefix and the number),
into "%0#10lx" format, which zero fills instead of space fills, and
which uses the printf flag '#' to request the "0x" prefix instead of
hard coding it.
Also replace some other "0x%lx" formats with "%#lx", making use of the
'#' printf flag again.
Signed-off-by: Paul Jackson <pj@sgi.com> Cc: "Yinghai Lu" <yhlu.kernel@gmail.com> Cc: "Jack Steiner" <steiner@sgi.com> Cc: "Mike Travis" <travis@sgi.com> Cc: "Huang Cc: Ying" <ying.huang@intel.com> Cc: "Andi Kleen" <andi@firstfloor.org> Cc: "Andrew Morton" <akpm@linux-foundation.org> Cc: Paul Jackson <pj@sgi.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
This is a preparatory patch for the next patch in series.
Moves some code from e820_setup_gap to a new function e820_search_gap.
This patch is a part of a bug fix where we walk the ACPI table to calculate
a gap for PCI optional devices.
v1->v2: Patch on top of tip/master.
Fixes a bug introduced in the last patch about the typeof "last".
Also the new function e820_search_gap now returns if we found a gap in
e820_map.
Yinghai Lu [Tue, 24 Jun 2008 21:57:29 +0000 (14:57 -0700)]
x86: add table_top check for alloc_low_page in 64 bit
that range is from find_e820_area, so don't try to use end_pfn to see
if out of boundary...use table_top instead to avoid possible strange
result while cross the boundary...
also change early_printk to printk, because init_memory_mapping is after
early param parsing, and console=uart8250 already working at that time.
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
wrmsr is a special instruction which can have arbitrary system-wide
effects. We don't want the compiler to reorder it with respect to
memory operations, so make it a memory barrier.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Add "memory" clobbers to savesegment and loadsegment, since they can
affect memory accesses and we never want the compiler to reorder them
with respect to memory references.
Signed-off-by: Jeremy Fitzhardinge <jeremy.fitzhardinge@citrix.com> Cc: xen-devel <xen-devel@lists.xensource.com> Cc: Stephen Tweedie <sct@redhat.com> Cc: Eduardo Habkost <ehabkost@redhat.com> Cc: Mark McLoughlin <markmc@redhat.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Cyrill Gorcunov [Tue, 24 Jun 2008 20:52:05 +0000 (22:52 +0200)]
x86: nmi_watchdog - use NMI_NONE by default
There is no need to keep NMI_DISABLED definition and use it
for nmi_watchdog by default. Here is the point why:
- IO-APIC and APIC chips are programmed for nmi_watchdog support at very
early stage of kernel booting and not having nmi_watchdog specified as
boot option lead only to nmi_watchdog becomes to NMI_NONE anyway
- enable nmi_watchdog thru /proc/sys/kernel/nmi if it was not specified at
boot is not possible too (even having this sysfs entry)
Paul Jackson [Sun, 22 Jun 2008 14:22:17 +0000 (07:22 -0700)]
x86 boot: more consistently use type int for node ids
Everywhere I look, node id's are of type 'int', except in this one
case, which has 'unsigned long'. Change this one to 'int' as well.
There is nothing special about the way this variable 'nid' is used in
this routine to justify using an unusual type here.
Signed-off-by: Paul Jackson <pj@sgi.com> Cc: "Yinghai Lu" <yhlu.kernel@gmail.com> Cc: "Jack Steiner" <steiner@sgi.com> Cc: "Mike Travis" <travis@sgi.com> Cc: "Huang Cc: Ying" <ying.huang@intel.com> Cc: "Andi Kleen" <andi@firstfloor.org> Cc: "Andrew Morton" <akpm@linux-foundation.org> Cc: Paul Jackson <pj@sgi.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Paul Jackson [Sun, 22 Jun 2008 14:22:12 +0000 (07:22 -0700)]
x86 boot: show pfn addresses in hex not decimal in some kernel info printks
Page frame numbers (the portion of physical addresses above the low
order page offsets) are displayed in several kernel debug and info
prints in decimal, not hex. Decimal addresse are unreadable. Use hex.
Signed-off-by: Paul Jackson <pj@sgi.com> Cc: "Yinghai Lu" <yhlu.kernel@gmail.com> Cc: "Jack Steiner" <steiner@sgi.com> Cc: "Mike Travis" <travis@sgi.com> Cc: "Huang Cc: Ying" <ying.huang@intel.com> Cc: "Andi Kleen" <andi@firstfloor.org> Cc: "Andrew Morton" <akpm@linux-foundation.org> Cc: Paul Jackson <pj@sgi.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Paul Jackson [Sun, 22 Jun 2008 14:22:07 +0000 (07:22 -0700)]
x86 boot: allow overlapping early reserve memory ranges
Add support for overlapping early memory reservations.
In general, they still can't overlap, and will panic
with "Overlapping early reservations" if they do overlap.
But if a memory range is reserved with the new call:
reserve_early_overlap_ok()
rather than with the usual call:
reserve_early()
then subsequent early reservations are allowed to overlap.
This new reserve_early_overlap_ok() call is only used in one
place so far, which is the "BIOS reserved" reservation for the
the EBDA region, which out of Paranoia reserves more than what
the BIOS might have specified, and which thus might overlap with
another legitimate early memory reservation (such as, perhaps,
the EFI memmap.)
Signed-off-by: Paul Jackson <pj@sgi.com> Cc: "Yinghai Lu" <yhlu.kernel@gmail.com> Cc: "Jack Steiner" <steiner@sgi.com> Cc: "Mike Travis" <travis@sgi.com> Cc: "Huang Cc: Ying" <ying.huang@intel.com> Cc: "Andi Kleen" <andi@firstfloor.org> Cc: "Andrew Morton" <akpm@linux-foundation.org> Cc: Paul Jackson <pj@sgi.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Paul Jackson [Sun, 22 Jun 2008 14:22:02 +0000 (07:22 -0700)]
x86 boot: x86_64 efi compiler warning fix
Fix a compiler warning. Rather than always casting a u32 to a pointer
(which generates a warning on x86_64 systems) instead separate the
x86_32 and x86_64 assignments entirely with ifdefs. Until other recent
changes to this code, it used to have x86_64 separated like this.
Signed-off-by: Paul Jackson <pj@sgi.com> Cc: "Yinghai Lu" <yhlu.kernel@gmail.com> Cc: "Jack Steiner" <steiner@sgi.com> Cc: "Mike Travis" <travis@sgi.com> Cc: "Huang Cc: Ying" <ying.huang@intel.com> Cc: "Andi Kleen" <andi@firstfloor.org> Cc: "Andrew Morton" <akpm@linux-foundation.org> Cc: Paul Jackson <pj@sgi.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
Ingo Molnar [Mon, 23 Jun 2008 20:19:22 +0000 (22:19 +0200)]
x86: build fix
fix:
arch/x86/kernel/setup_32.c:409: error: 'enable_local_apic' undeclared (first use in this function)
arch/x86/kernel/setup_32.c:409: error: (Each undeclared identifier is reported only once
arch/x86/kernel/setup_32.c:409: error: for each function it appears in.)
Yinghai Lu [Sat, 21 Jun 2008 08:14:27 +0000 (01:14 -0700)]
x86: clean up init_amd()
1. move out calling of check_enable_amd_mmconf_dmi out of setup_64.c
put it into init_amd(), so don't need to make extra dmi check for
system with other cpus.
2. 15 --> 0xf
Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com> Signed-off-by: Ingo Molnar <mingo@elte.hu>
projects / linux-2.6 / log