Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
/*-
|
|
|
|
* Copyright (c) 2013 The FreeBSD Foundation
|
|
|
|
* All rights reserved.
|
|
|
|
*
|
|
|
|
* This software was developed by Konstantin Belousov <kib@FreeBSD.org>
|
|
|
|
* under sponsorship from the FreeBSD Foundation.
|
|
|
|
*
|
|
|
|
* Redistribution and use in source and binary forms, with or without
|
|
|
|
* modification, are permitted provided that the following conditions
|
|
|
|
* are met:
|
|
|
|
* 1. Redistributions of source code must retain the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer.
|
|
|
|
* 2. Redistributions in binary form must reproduce the above copyright
|
|
|
|
* notice, this list of conditions and the following disclaimer in the
|
|
|
|
* documentation and/or other materials provided with the distribution.
|
|
|
|
*
|
|
|
|
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
|
|
|
|
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
|
|
|
|
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
|
|
|
|
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
|
|
|
|
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
|
|
|
|
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
|
|
|
|
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
|
|
|
|
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
|
|
|
|
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
|
|
|
|
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
|
|
|
|
* SUCH DAMAGE.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <sys/cdefs.h>
|
|
|
|
__FBSDID("$FreeBSD$");
|
|
|
|
|
|
|
|
#define RB_AUGMENT(entry) dmar_gas_augment_entry(entry)
|
|
|
|
|
|
|
|
#include <sys/param.h>
|
|
|
|
#include <sys/systm.h>
|
|
|
|
#include <sys/malloc.h>
|
|
|
|
#include <sys/bus.h>
|
|
|
|
#include <sys/interrupt.h>
|
|
|
|
#include <sys/kernel.h>
|
|
|
|
#include <sys/ktr.h>
|
|
|
|
#include <sys/lock.h>
|
|
|
|
#include <sys/proc.h>
|
|
|
|
#include <sys/rwlock.h>
|
|
|
|
#include <sys/memdesc.h>
|
|
|
|
#include <sys/mutex.h>
|
|
|
|
#include <sys/sysctl.h>
|
|
|
|
#include <sys/rman.h>
|
|
|
|
#include <sys/taskqueue.h>
|
|
|
|
#include <sys/tree.h>
|
|
|
|
#include <sys/uio.h>
|
Use VT-d interrupt remapping block (IR) to perform FSB messages
translation. In particular, despite IO-APICs only take 8bit apic id,
IR translation structures accept 32bit APIC Id, which allows x2APIC
mode to function properly. Extend msi_cpu of struct msi_intrsrc and
io_cpu of ioapic_intsrc to full int from one byte.
KPI of IR is isolated into the x86/iommu/iommu_intrmap.h, to avoid
bringing all dmar headers into interrupt code. The non-PCI(e) devices
which generate message interrupts on FSB require special handling. The
HPET FSB interrupts are remapped, while DMAR interrupts are not.
For each msi and ioapic interrupt source, the iommu cookie is added,
which is in fact index of the IRE (interrupt remap entry) in the IR
table. Cookie is made at the source allocation time, and then used at
the map time to fill both IRE and device registers. The MSI
address/data registers and IO-APIC redirection registers are
programmed with the special values which are recognized by IR and used
to restore the IRE index, to find proper delivery mode and target.
Map all MSI interrupts in the block when msi_map() is called.
Since an interrupt source setup and dismantle code are done in the
non-sleepable context, flushing interrupt entries cache in the IR
hardware, which is done async and ideally waits for the interrupt,
requires busy-wait for queue to drain. The dmar_qi_wait_for_seq() is
modified to take a boolean argument requesting busy-wait for the
written sequence number instead of waiting for interrupt.
Some interrupts are configured before IR is initialized, e.g. ACPI
SCI. Add intr_reprogram() function to reprogram all already
configured interrupts, and call it immediately before an IR unit is
enabled. There is still a small window after the IO-APIC redirection
entry is reprogrammed with cookie but before the unit is enabled, but
to fix this properly, IR must be started much earlier.
Add workarounds for 5500 and X58 northbridges, some revisions of which
have severe flaws in handling IR. Use the same identification methods
as employed by Linux.
Review: https://reviews.freebsd.org/D1892
Reviewed by: neel
Discussed with: jhb
Tested by: glebius, pho (previous versions)
Sponsored by: The FreeBSD Foundation
MFC after: 3 weeks
2015-03-19 13:57:47 +00:00
|
|
|
#include <sys/vmem.h>
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
#include <dev/pci/pcivar.h>
|
|
|
|
#include <vm/vm.h>
|
|
|
|
#include <vm/vm_extern.h>
|
|
|
|
#include <vm/vm_kern.h>
|
|
|
|
#include <vm/vm_object.h>
|
|
|
|
#include <vm/vm_page.h>
|
|
|
|
#include <vm/vm_map.h>
|
|
|
|
#include <vm/uma.h>
|
|
|
|
#include <machine/atomic.h>
|
|
|
|
#include <machine/bus.h>
|
|
|
|
#include <machine/md_var.h>
|
|
|
|
#include <machine/specialreg.h>
|
|
|
|
#include <x86/include/busdma_impl.h>
|
|
|
|
#include <x86/iommu/intel_reg.h>
|
|
|
|
#include <x86/iommu/busdma_dmar.h>
|
|
|
|
#include <x86/iommu/intel_dmar.h>
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Guest Address Space management.
|
|
|
|
*/
|
|
|
|
|
|
|
|
static uma_zone_t dmar_map_entry_zone;
|
|
|
|
|
|
|
|
static void
|
|
|
|
intel_gas_init(void)
|
|
|
|
{
|
|
|
|
|
|
|
|
dmar_map_entry_zone = uma_zcreate("DMAR_MAP_ENTRY",
|
|
|
|
sizeof(struct dmar_map_entry), NULL, NULL,
|
|
|
|
NULL, NULL, UMA_ALIGN_PTR, 0);
|
|
|
|
}
|
|
|
|
SYSINIT(intel_gas, SI_SUB_DRIVERS, SI_ORDER_FIRST, intel_gas_init, NULL);
|
|
|
|
|
|
|
|
struct dmar_map_entry *
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_alloc_entry(struct dmar_domain *domain, u_int flags)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
struct dmar_map_entry *res;
|
|
|
|
|
|
|
|
KASSERT((flags & ~(DMAR_PGF_WAITOK)) == 0,
|
|
|
|
("unsupported flags %x", flags));
|
|
|
|
|
|
|
|
res = uma_zalloc(dmar_map_entry_zone, ((flags & DMAR_PGF_WAITOK) !=
|
|
|
|
0 ? M_WAITOK : M_NOWAIT) | M_ZERO);
|
2013-11-01 17:38:52 +00:00
|
|
|
if (res != NULL) {
|
2015-06-26 07:01:29 +00:00
|
|
|
res->domain = domain;
|
|
|
|
atomic_add_int(&domain->entries_cnt, 1);
|
2013-11-01 17:38:52 +00:00
|
|
|
}
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (res);
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_free_entry(struct dmar_domain *domain, struct dmar_map_entry *entry)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
KASSERT(domain == entry->domain,
|
|
|
|
("mismatched free domain %p entry %p entry->domain %p", domain,
|
|
|
|
entry, entry->domain));
|
|
|
|
atomic_subtract_int(&domain->entries_cnt, 1);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
uma_zfree(dmar_map_entry_zone, entry);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
dmar_gas_cmp_entries(struct dmar_map_entry *a, struct dmar_map_entry *b)
|
|
|
|
{
|
|
|
|
|
|
|
|
/* Last entry have zero size, so <= */
|
|
|
|
KASSERT(a->start <= a->end, ("inverted entry %p (%jx, %jx)",
|
|
|
|
a, (uintmax_t)a->start, (uintmax_t)a->end));
|
|
|
|
KASSERT(b->start <= b->end, ("inverted entry %p (%jx, %jx)",
|
|
|
|
b, (uintmax_t)b->start, (uintmax_t)b->end));
|
|
|
|
KASSERT(a->end <= b->start || b->end <= a->start ||
|
|
|
|
a->end == a->start || b->end == b->start,
|
|
|
|
("overlapping entries %p (%jx, %jx) %p (%jx, %jx)",
|
|
|
|
a, (uintmax_t)a->start, (uintmax_t)a->end,
|
|
|
|
b, (uintmax_t)b->start, (uintmax_t)b->end));
|
|
|
|
|
|
|
|
if (a->end < b->end)
|
|
|
|
return (-1);
|
|
|
|
else if (b->end < a->end)
|
|
|
|
return (1);
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
dmar_gas_augment_entry(struct dmar_map_entry *entry)
|
|
|
|
{
|
|
|
|
struct dmar_map_entry *l, *r;
|
|
|
|
|
|
|
|
for (; entry != NULL; entry = RB_PARENT(entry, rb_entry)) {
|
|
|
|
l = RB_LEFT(entry, rb_entry);
|
|
|
|
r = RB_RIGHT(entry, rb_entry);
|
|
|
|
if (l == NULL && r == NULL) {
|
|
|
|
entry->free_down = entry->free_after;
|
|
|
|
} else if (l == NULL && r != NULL) {
|
|
|
|
entry->free_down = MAX(entry->free_after, r->free_down);
|
|
|
|
} else if (/*l != NULL && */ r == NULL) {
|
|
|
|
entry->free_down = MAX(entry->free_after, l->free_down);
|
|
|
|
} else /* if (l != NULL && r != NULL) */ {
|
|
|
|
entry->free_down = MAX(entry->free_after, l->free_down);
|
|
|
|
entry->free_down = MAX(entry->free_down, r->free_down);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
RB_GENERATE(dmar_gas_entries_tree, dmar_map_entry, rb_entry,
|
|
|
|
dmar_gas_cmp_entries);
|
|
|
|
|
|
|
|
static void
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_fix_free(struct dmar_domain *domain, struct dmar_map_entry *entry)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
struct dmar_map_entry *next;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
next = RB_NEXT(dmar_gas_entries_tree, &domain->rb_root, entry);
|
|
|
|
entry->free_after = (next != NULL ? next->start : domain->end) -
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
entry->end;
|
|
|
|
dmar_gas_augment_entry(entry);
|
|
|
|
}
|
|
|
|
|
|
|
|
#ifdef INVARIANTS
|
|
|
|
static void
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_check_free(struct dmar_domain *domain)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
struct dmar_map_entry *entry, *next, *l, *r;
|
|
|
|
dmar_gaddr_t v;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
RB_FOREACH(entry, dmar_gas_entries_tree, &domain->rb_root) {
|
|
|
|
KASSERT(domain == entry->domain,
|
|
|
|
("mismatched free domain %p entry %p entry->domain %p",
|
|
|
|
domain, entry, entry->domain));
|
|
|
|
next = RB_NEXT(dmar_gas_entries_tree, &domain->rb_root, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (next == NULL) {
|
2015-06-26 07:01:29 +00:00
|
|
|
MPASS(entry->free_after == domain->end - entry->end);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
} else {
|
|
|
|
MPASS(entry->free_after = next->start - entry->end);
|
|
|
|
MPASS(entry->end <= next->start);
|
|
|
|
}
|
|
|
|
l = RB_LEFT(entry, rb_entry);
|
|
|
|
r = RB_RIGHT(entry, rb_entry);
|
|
|
|
if (l == NULL && r == NULL) {
|
|
|
|
MPASS(entry->free_down == entry->free_after);
|
|
|
|
} else if (l == NULL && r != NULL) {
|
|
|
|
MPASS(entry->free_down = MAX(entry->free_after,
|
|
|
|
r->free_down));
|
|
|
|
} else if (r == NULL) {
|
|
|
|
MPASS(entry->free_down = MAX(entry->free_after,
|
|
|
|
l->free_down));
|
|
|
|
} else {
|
|
|
|
v = MAX(entry->free_after, l->free_down);
|
2017-04-14 15:16:41 +00:00
|
|
|
v = MAX(v, r->free_down);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
MPASS(entry->free_down == v);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
#endif
|
|
|
|
|
|
|
|
static bool
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_rb_insert(struct dmar_domain *domain, struct dmar_map_entry *entry)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
struct dmar_map_entry *prev, *found;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
found = RB_INSERT(dmar_gas_entries_tree, &domain->rb_root, entry);
|
|
|
|
dmar_gas_fix_free(domain, entry);
|
|
|
|
prev = RB_PREV(dmar_gas_entries_tree, &domain->rb_root, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (prev != NULL)
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_fix_free(domain, prev);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (found == NULL);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_rb_remove(struct dmar_domain *domain, struct dmar_map_entry *entry)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
struct dmar_map_entry *prev;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
prev = RB_PREV(dmar_gas_entries_tree, &domain->rb_root, entry);
|
|
|
|
RB_REMOVE(dmar_gas_entries_tree, &domain->rb_root, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (prev != NULL)
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_fix_free(domain, prev);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_init_domain(struct dmar_domain *domain)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
struct dmar_map_entry *begin, *end;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
begin = dmar_gas_alloc_entry(domain, DMAR_PGF_WAITOK);
|
|
|
|
end = dmar_gas_alloc_entry(domain, DMAR_PGF_WAITOK);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_LOCK(domain);
|
|
|
|
KASSERT(domain->entries_cnt == 2, ("dirty domain %p", domain));
|
|
|
|
KASSERT(RB_EMPTY(&domain->rb_root), ("non-empty entries %p", domain));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
|
|
|
begin->start = 0;
|
|
|
|
begin->end = DMAR_PAGE_SIZE;
|
2015-06-26 07:01:29 +00:00
|
|
|
begin->free_after = domain->end - begin->end;
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
begin->flags = DMAR_MAP_ENTRY_PLACE | DMAR_MAP_ENTRY_UNMAPPED;
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_rb_insert(domain, begin);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
end->start = domain->end;
|
|
|
|
end->end = domain->end;
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
end->free_after = 0;
|
|
|
|
end->flags = DMAR_MAP_ENTRY_PLACE | DMAR_MAP_ENTRY_UNMAPPED;
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_rb_insert(domain, end);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
domain->first_place = begin;
|
|
|
|
domain->last_place = end;
|
|
|
|
domain->flags |= DMAR_DOMAIN_GAS_INITED;
|
|
|
|
DMAR_DOMAIN_UNLOCK(domain);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
void
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_fini_domain(struct dmar_domain *domain)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
struct dmar_map_entry *entry, *entry1;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_ASSERT_LOCKED(domain);
|
|
|
|
KASSERT(domain->entries_cnt == 2, ("domain still in use %p", domain));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
entry = RB_MIN(dmar_gas_entries_tree, &domain->rb_root);
|
|
|
|
KASSERT(entry->start == 0, ("start entry start %p", domain));
|
|
|
|
KASSERT(entry->end == DMAR_PAGE_SIZE, ("start entry end %p", domain));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
KASSERT(entry->flags == DMAR_MAP_ENTRY_PLACE,
|
2015-06-26 07:01:29 +00:00
|
|
|
("start entry flags %p", domain));
|
|
|
|
RB_REMOVE(dmar_gas_entries_tree, &domain->rb_root, entry);
|
|
|
|
dmar_gas_free_entry(domain, entry);
|
|
|
|
|
|
|
|
entry = RB_MAX(dmar_gas_entries_tree, &domain->rb_root);
|
|
|
|
KASSERT(entry->start == domain->end, ("end entry start %p", domain));
|
|
|
|
KASSERT(entry->end == domain->end, ("end entry end %p", domain));
|
|
|
|
KASSERT(entry->free_after == 0, ("end entry free_after %p", domain));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
KASSERT(entry->flags == DMAR_MAP_ENTRY_PLACE,
|
2015-06-26 07:01:29 +00:00
|
|
|
("end entry flags %p", domain));
|
|
|
|
RB_REMOVE(dmar_gas_entries_tree, &domain->rb_root, entry);
|
|
|
|
dmar_gas_free_entry(domain, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
RB_FOREACH_SAFE(entry, dmar_gas_entries_tree, &domain->rb_root,
|
|
|
|
entry1) {
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
KASSERT((entry->flags & DMAR_MAP_ENTRY_RMRR) != 0,
|
2015-06-26 07:01:29 +00:00
|
|
|
("non-RMRR entry left %p", domain));
|
|
|
|
RB_REMOVE(dmar_gas_entries_tree, &domain->rb_root, entry);
|
|
|
|
dmar_gas_free_entry(domain, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
struct dmar_gas_match_args {
|
2015-06-26 07:01:29 +00:00
|
|
|
struct dmar_domain *domain;
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
dmar_gaddr_t size;
|
2015-04-08 01:55:22 +00:00
|
|
|
int offset;
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
const struct bus_dma_tag_common *common;
|
|
|
|
u_int gas_flags;
|
|
|
|
struct dmar_map_entry *entry;
|
|
|
|
};
|
|
|
|
|
|
|
|
static bool
|
|
|
|
dmar_gas_match_one(struct dmar_gas_match_args *a, struct dmar_map_entry *prev,
|
|
|
|
dmar_gaddr_t end)
|
|
|
|
{
|
|
|
|
dmar_gaddr_t bs, start;
|
|
|
|
|
|
|
|
if (a->entry->start + a->size > end)
|
|
|
|
return (false);
|
|
|
|
|
|
|
|
/* DMAR_PAGE_SIZE to create gap after new entry. */
|
|
|
|
if (a->entry->start < prev->end + DMAR_PAGE_SIZE ||
|
2015-04-08 01:55:22 +00:00
|
|
|
a->entry->start + a->size + a->offset + DMAR_PAGE_SIZE >
|
|
|
|
prev->end + prev->free_after)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (false);
|
|
|
|
|
|
|
|
/* No boundary crossing. */
|
2015-04-08 01:55:22 +00:00
|
|
|
if (dmar_test_boundary(a->entry->start + a->offset, a->size,
|
|
|
|
a->common->boundary))
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (true);
|
|
|
|
|
|
|
|
/*
|
2015-04-08 01:55:22 +00:00
|
|
|
* The start + offset to start + offset + size region crosses
|
|
|
|
* the boundary. Check if there is enough space after the
|
|
|
|
* next boundary after the prev->end.
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
*/
|
2016-04-21 19:57:40 +00:00
|
|
|
bs = rounddown2(a->entry->start + a->offset + a->common->boundary,
|
|
|
|
a->common->boundary);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
start = roundup2(bs, a->common->alignment);
|
|
|
|
/* DMAR_PAGE_SIZE to create gap after new entry. */
|
2015-04-08 01:55:22 +00:00
|
|
|
if (start + a->offset + a->size + DMAR_PAGE_SIZE <=
|
|
|
|
prev->end + prev->free_after &&
|
|
|
|
start + a->offset + a->size <= end &&
|
|
|
|
dmar_test_boundary(start + a->offset, a->size,
|
2015-03-17 22:00:11 +00:00
|
|
|
a->common->boundary)) {
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
a->entry->start = start;
|
|
|
|
return (true);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
2015-03-17 22:00:11 +00:00
|
|
|
* Not enough space to align at the requested boundary, or
|
|
|
|
* boundary is smaller than the size, but allowed to split.
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
* We already checked that start + size does not overlap end.
|
|
|
|
*
|
|
|
|
* XXXKIB. It is possible that bs is exactly at the start of
|
|
|
|
* the next entry, then we do not have gap. Ignore for now.
|
|
|
|
*/
|
|
|
|
if ((a->gas_flags & DMAR_GM_CANSPLIT) != 0) {
|
|
|
|
a->size = bs - a->entry->start;
|
|
|
|
return (true);
|
|
|
|
}
|
|
|
|
|
|
|
|
return (false);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void
|
|
|
|
dmar_gas_match_insert(struct dmar_gas_match_args *a,
|
|
|
|
struct dmar_map_entry *prev)
|
|
|
|
{
|
|
|
|
struct dmar_map_entry *next;
|
|
|
|
bool found;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The prev->end is always aligned on the page size, which
|
|
|
|
* causes page alignment for the entry->start too. The size
|
|
|
|
* is checked to be multiple of the page size.
|
|
|
|
*
|
|
|
|
* The page sized gap is created between consequent
|
|
|
|
* allocations to ensure that out-of-bounds accesses fault.
|
|
|
|
*/
|
|
|
|
a->entry->end = a->entry->start + a->size;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
next = RB_NEXT(dmar_gas_entries_tree, &a->domain->rb_root, prev);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
KASSERT(next->start >= a->entry->end &&
|
2015-03-17 21:55:33 +00:00
|
|
|
next->start - a->entry->start >= a->size &&
|
|
|
|
prev->end <= a->entry->end,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
("dmar_gas_match_insert hole failed %p prev (%jx, %jx) "
|
2015-06-26 07:01:29 +00:00
|
|
|
"free_after %jx next (%jx, %jx) entry (%jx, %jx)", a->domain,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
(uintmax_t)prev->start, (uintmax_t)prev->end,
|
|
|
|
(uintmax_t)prev->free_after,
|
|
|
|
(uintmax_t)next->start, (uintmax_t)next->end,
|
|
|
|
(uintmax_t)a->entry->start, (uintmax_t)a->entry->end));
|
|
|
|
|
|
|
|
prev->free_after = a->entry->start - prev->end;
|
|
|
|
a->entry->free_after = next->start - a->entry->end;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
found = dmar_gas_rb_insert(a->domain, a->entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
KASSERT(found, ("found dup %p start %jx size %jx",
|
2015-06-26 07:01:29 +00:00
|
|
|
a->domain, (uintmax_t)a->entry->start, (uintmax_t)a->size));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
a->entry->flags = DMAR_MAP_ENTRY_MAP;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
KASSERT(RB_PREV(dmar_gas_entries_tree, &a->domain->rb_root,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
a->entry) == prev,
|
|
|
|
("entry %p prev %p inserted prev %p", a->entry, prev,
|
2015-06-26 07:01:29 +00:00
|
|
|
RB_PREV(dmar_gas_entries_tree, &a->domain->rb_root, a->entry)));
|
|
|
|
KASSERT(RB_NEXT(dmar_gas_entries_tree, &a->domain->rb_root,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
a->entry) == next,
|
|
|
|
("entry %p next %p inserted next %p", a->entry, next,
|
2015-06-26 07:01:29 +00:00
|
|
|
RB_NEXT(dmar_gas_entries_tree, &a->domain->rb_root, a->entry)));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
dmar_gas_lowermatch(struct dmar_gas_match_args *a, struct dmar_map_entry *prev)
|
|
|
|
{
|
|
|
|
struct dmar_map_entry *l;
|
|
|
|
int ret;
|
|
|
|
|
|
|
|
if (prev->end < a->common->lowaddr) {
|
|
|
|
a->entry->start = roundup2(prev->end + DMAR_PAGE_SIZE,
|
|
|
|
a->common->alignment);
|
|
|
|
if (dmar_gas_match_one(a, prev, a->common->lowaddr)) {
|
|
|
|
dmar_gas_match_insert(a, prev);
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
}
|
2015-04-08 01:55:22 +00:00
|
|
|
if (prev->free_down < a->size + a->offset + DMAR_PAGE_SIZE)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (ENOMEM);
|
|
|
|
l = RB_LEFT(prev, rb_entry);
|
|
|
|
if (l != NULL) {
|
|
|
|
ret = dmar_gas_lowermatch(a, l);
|
|
|
|
if (ret == 0)
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
l = RB_RIGHT(prev, rb_entry);
|
|
|
|
if (l != NULL)
|
|
|
|
return (dmar_gas_lowermatch(a, l));
|
|
|
|
return (ENOMEM);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
|
|
|
dmar_gas_uppermatch(struct dmar_gas_match_args *a)
|
|
|
|
{
|
|
|
|
struct dmar_map_entry *next, *prev, find_entry;
|
|
|
|
|
|
|
|
find_entry.start = a->common->highaddr;
|
2015-06-26 07:01:29 +00:00
|
|
|
next = RB_NFIND(dmar_gas_entries_tree, &a->domain->rb_root,
|
|
|
|
&find_entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (next == NULL)
|
|
|
|
return (ENOMEM);
|
2015-06-26 07:01:29 +00:00
|
|
|
prev = RB_PREV(dmar_gas_entries_tree, &a->domain->rb_root, next);
|
|
|
|
KASSERT(prev != NULL, ("no prev %p %jx", a->domain,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
(uintmax_t)find_entry.start));
|
|
|
|
for (;;) {
|
|
|
|
a->entry->start = prev->start + DMAR_PAGE_SIZE;
|
|
|
|
if (a->entry->start < a->common->highaddr)
|
|
|
|
a->entry->start = a->common->highaddr;
|
|
|
|
a->entry->start = roundup2(a->entry->start,
|
|
|
|
a->common->alignment);
|
2015-06-26 07:01:29 +00:00
|
|
|
if (dmar_gas_match_one(a, prev, a->domain->end)) {
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
dmar_gas_match_insert(a, prev);
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* XXXKIB. This falls back to linear iteration over
|
|
|
|
* the free space in the high region. But high
|
|
|
|
* regions are almost unused, the code should be
|
|
|
|
* enough to cover the case, although in the
|
|
|
|
* non-optimal way.
|
|
|
|
*/
|
|
|
|
prev = next;
|
2015-06-26 07:01:29 +00:00
|
|
|
next = RB_NEXT(dmar_gas_entries_tree, &a->domain->rb_root,
|
|
|
|
prev);
|
|
|
|
KASSERT(next != NULL, ("no next %p %jx", a->domain,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
(uintmax_t)find_entry.start));
|
2015-06-26 07:01:29 +00:00
|
|
|
if (next->end >= a->domain->end)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (ENOMEM);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_find_space(struct dmar_domain *domain,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
const struct bus_dma_tag_common *common, dmar_gaddr_t size,
|
2015-04-08 01:55:22 +00:00
|
|
|
int offset, u_int flags, struct dmar_map_entry *entry)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
struct dmar_gas_match_args a;
|
|
|
|
int error;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_ASSERT_LOCKED(domain);
|
|
|
|
KASSERT(entry->flags == 0, ("dirty entry %p %p", domain, entry));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
KASSERT((size & DMAR_PAGE_MASK) == 0, ("size %jx", (uintmax_t)size));
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
a.domain = domain;
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
a.size = size;
|
2015-04-08 01:55:22 +00:00
|
|
|
a.offset = offset;
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
a.common = common;
|
|
|
|
a.gas_flags = flags;
|
|
|
|
a.entry = entry;
|
|
|
|
|
|
|
|
/* Handle lower region. */
|
|
|
|
if (common->lowaddr > 0) {
|
2015-06-26 07:01:29 +00:00
|
|
|
error = dmar_gas_lowermatch(&a, RB_ROOT(&domain->rb_root));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (error == 0)
|
|
|
|
return (0);
|
|
|
|
KASSERT(error == ENOMEM,
|
|
|
|
("error %d from dmar_gas_lowermatch", error));
|
|
|
|
}
|
|
|
|
/* Handle upper region. */
|
2015-06-26 07:01:29 +00:00
|
|
|
if (common->highaddr >= domain->end)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (ENOMEM);
|
|
|
|
error = dmar_gas_uppermatch(&a);
|
|
|
|
KASSERT(error == ENOMEM,
|
|
|
|
("error %d from dmar_gas_uppermatch", error));
|
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_alloc_region(struct dmar_domain *domain, struct dmar_map_entry *entry,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
u_int flags)
|
|
|
|
{
|
|
|
|
struct dmar_map_entry *next, *prev;
|
|
|
|
bool found;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_ASSERT_LOCKED(domain);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
|
|
|
if ((entry->start & DMAR_PAGE_MASK) != 0 ||
|
|
|
|
(entry->end & DMAR_PAGE_MASK) != 0)
|
|
|
|
return (EINVAL);
|
|
|
|
if (entry->start >= entry->end)
|
|
|
|
return (EINVAL);
|
2015-06-26 07:01:29 +00:00
|
|
|
if (entry->end >= domain->end)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (EINVAL);
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
next = RB_NFIND(dmar_gas_entries_tree, &domain->rb_root, entry);
|
|
|
|
KASSERT(next != NULL, ("next must be non-null %p %jx", domain,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
(uintmax_t)entry->start));
|
2015-06-26 07:01:29 +00:00
|
|
|
prev = RB_PREV(dmar_gas_entries_tree, &domain->rb_root, next);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
/* prev could be NULL */
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Adapt to broken BIOSes which specify overlapping RMRR
|
|
|
|
* entries.
|
|
|
|
*
|
|
|
|
* XXXKIB: this does not handle a case when prev or next
|
|
|
|
* entries are completely covered by the current one, which
|
|
|
|
* extends both ways.
|
|
|
|
*/
|
|
|
|
if (prev != NULL && prev->end > entry->start &&
|
|
|
|
(prev->flags & DMAR_MAP_ENTRY_PLACE) == 0) {
|
|
|
|
if ((prev->flags & DMAR_MAP_ENTRY_RMRR) == 0)
|
|
|
|
return (EBUSY);
|
|
|
|
entry->start = prev->end;
|
|
|
|
}
|
|
|
|
if (next != NULL && next->start < entry->end &&
|
|
|
|
(next->flags & DMAR_MAP_ENTRY_PLACE) == 0) {
|
|
|
|
if ((next->flags & DMAR_MAP_ENTRY_RMRR) == 0)
|
|
|
|
return (EBUSY);
|
|
|
|
entry->end = next->start;
|
|
|
|
}
|
|
|
|
if (entry->end == entry->start)
|
|
|
|
return (0);
|
|
|
|
|
|
|
|
if (prev != NULL && prev->end > entry->start) {
|
|
|
|
/* This assumes that prev is the placeholder entry. */
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_rb_remove(domain, prev);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
prev = NULL;
|
|
|
|
}
|
|
|
|
if (next != NULL && next->start < entry->end) {
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_rb_remove(domain, next);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
next = NULL;
|
|
|
|
}
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
found = dmar_gas_rb_insert(domain, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
KASSERT(found, ("found RMRR dup %p start %jx end %jx",
|
2015-06-26 07:01:29 +00:00
|
|
|
domain, (uintmax_t)entry->start, (uintmax_t)entry->end));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
entry->flags = DMAR_MAP_ENTRY_RMRR;
|
|
|
|
|
|
|
|
#ifdef INVARIANTS
|
|
|
|
struct dmar_map_entry *ip, *in;
|
2015-06-26 07:01:29 +00:00
|
|
|
ip = RB_PREV(dmar_gas_entries_tree, &domain->rb_root, entry);
|
|
|
|
in = RB_NEXT(dmar_gas_entries_tree, &domain->rb_root, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
KASSERT(prev == NULL || ip == prev,
|
|
|
|
("RMRR %p (%jx %jx) prev %p (%jx %jx) ins prev %p (%jx %jx)",
|
|
|
|
entry, entry->start, entry->end, prev,
|
|
|
|
prev == NULL ? 0 : prev->start, prev == NULL ? 0 : prev->end,
|
|
|
|
ip, ip == NULL ? 0 : ip->start, ip == NULL ? 0 : ip->end));
|
|
|
|
KASSERT(next == NULL || in == next,
|
|
|
|
("RMRR %p (%jx %jx) next %p (%jx %jx) ins next %p (%jx %jx)",
|
|
|
|
entry, entry->start, entry->end, next,
|
|
|
|
next == NULL ? 0 : next->start, next == NULL ? 0 : next->end,
|
|
|
|
in, in == NULL ? 0 : in->start, in == NULL ? 0 : in->end));
|
|
|
|
#endif
|
|
|
|
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
void
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_free_space(struct dmar_domain *domain, struct dmar_map_entry *entry)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_ASSERT_LOCKED(domain);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
KASSERT((entry->flags & (DMAR_MAP_ENTRY_PLACE | DMAR_MAP_ENTRY_RMRR |
|
|
|
|
DMAR_MAP_ENTRY_MAP)) == DMAR_MAP_ENTRY_MAP,
|
2015-06-26 07:01:29 +00:00
|
|
|
("permanent entry %p %p", domain, entry));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_rb_remove(domain, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
entry->flags &= ~DMAR_MAP_ENTRY_MAP;
|
|
|
|
#ifdef INVARIANTS
|
|
|
|
if (dmar_check_free)
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_check_free(domain);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
#endif
|
|
|
|
}
|
|
|
|
|
2013-11-01 17:38:52 +00:00
|
|
|
void
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_free_region(struct dmar_domain *domain, struct dmar_map_entry *entry)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
struct dmar_map_entry *next, *prev;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_ASSERT_LOCKED(domain);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
KASSERT((entry->flags & (DMAR_MAP_ENTRY_PLACE | DMAR_MAP_ENTRY_RMRR |
|
|
|
|
DMAR_MAP_ENTRY_MAP)) == DMAR_MAP_ENTRY_RMRR,
|
2015-06-26 07:01:29 +00:00
|
|
|
("non-RMRR entry %p %p", domain, entry));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
prev = RB_PREV(dmar_gas_entries_tree, &domain->rb_root, entry);
|
|
|
|
next = RB_NEXT(dmar_gas_entries_tree, &domain->rb_root, entry);
|
|
|
|
dmar_gas_rb_remove(domain, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
entry->flags &= ~DMAR_MAP_ENTRY_RMRR;
|
|
|
|
|
|
|
|
if (prev == NULL)
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_rb_insert(domain, domain->first_place);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (next == NULL)
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_rb_insert(domain, domain->last_place);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
int
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_map(struct dmar_domain *domain,
|
|
|
|
const struct bus_dma_tag_common *common, dmar_gaddr_t size, int offset,
|
|
|
|
u_int eflags, u_int flags, vm_page_t *ma, struct dmar_map_entry **res)
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
{
|
|
|
|
struct dmar_map_entry *entry;
|
|
|
|
int error;
|
|
|
|
|
|
|
|
KASSERT((flags & ~(DMAR_GM_CANWAIT | DMAR_GM_CANSPLIT)) == 0,
|
|
|
|
("invalid flags 0x%x", flags));
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
entry = dmar_gas_alloc_entry(domain, (flags & DMAR_GM_CANWAIT) != 0 ?
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
DMAR_PGF_WAITOK : 0);
|
|
|
|
if (entry == NULL)
|
|
|
|
return (ENOMEM);
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_LOCK(domain);
|
|
|
|
error = dmar_gas_find_space(domain, common, size, offset, flags,
|
|
|
|
entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (error == ENOMEM) {
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_UNLOCK(domain);
|
|
|
|
dmar_gas_free_entry(domain, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
#ifdef INVARIANTS
|
|
|
|
if (dmar_check_free)
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_check_free(domain);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
#endif
|
|
|
|
KASSERT(error == 0,
|
|
|
|
("unexpected error %d from dmar_gas_find_entry", error));
|
2015-06-26 07:01:29 +00:00
|
|
|
KASSERT(entry->end < domain->end, ("allocated GPA %jx, max GPA %jx",
|
|
|
|
(uintmax_t)entry->end, (uintmax_t)domain->end));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
entry->flags |= eflags;
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_UNLOCK(domain);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
error = domain_map_buf(domain, entry->start, entry->end - entry->start,
|
|
|
|
ma,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
((eflags & DMAR_MAP_ENTRY_READ) != 0 ? DMAR_PTE_R : 0) |
|
|
|
|
((eflags & DMAR_MAP_ENTRY_WRITE) != 0 ? DMAR_PTE_W : 0) |
|
|
|
|
((eflags & DMAR_MAP_ENTRY_SNOOP) != 0 ? DMAR_PTE_SNP : 0) |
|
|
|
|
((eflags & DMAR_MAP_ENTRY_TM) != 0 ? DMAR_PTE_TM : 0),
|
|
|
|
(flags & DMAR_GM_CANWAIT) != 0 ? DMAR_PGF_WAITOK : 0);
|
|
|
|
if (error == ENOMEM) {
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_domain_unload_entry(entry, true);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
KASSERT(error == 0,
|
2015-06-26 07:01:29 +00:00
|
|
|
("unexpected error %d from domain_map_buf", error));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
|
|
|
*res = entry;
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_map_region(struct dmar_domain *domain, struct dmar_map_entry *entry,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
u_int eflags, u_int flags, vm_page_t *ma)
|
|
|
|
{
|
|
|
|
dmar_gaddr_t start;
|
|
|
|
int error;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
KASSERT(entry->flags == 0, ("used RMRR entry %p %p %x", domain,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
entry, entry->flags));
|
|
|
|
KASSERT((flags & ~(DMAR_GM_CANWAIT)) == 0,
|
|
|
|
("invalid flags 0x%x", flags));
|
|
|
|
|
|
|
|
start = entry->start;
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_LOCK(domain);
|
|
|
|
error = dmar_gas_alloc_region(domain, entry, flags);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (error != 0) {
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_UNLOCK(domain);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
entry->flags |= eflags;
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_UNLOCK(domain);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (entry->end == entry->start)
|
|
|
|
return (0);
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
error = domain_map_buf(domain, entry->start, entry->end - entry->start,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
ma + OFF_TO_IDX(start - entry->start),
|
|
|
|
((eflags & DMAR_MAP_ENTRY_READ) != 0 ? DMAR_PTE_R : 0) |
|
|
|
|
((eflags & DMAR_MAP_ENTRY_WRITE) != 0 ? DMAR_PTE_W : 0) |
|
|
|
|
((eflags & DMAR_MAP_ENTRY_SNOOP) != 0 ? DMAR_PTE_SNP : 0) |
|
|
|
|
((eflags & DMAR_MAP_ENTRY_TM) != 0 ? DMAR_PTE_TM : 0),
|
|
|
|
(flags & DMAR_GM_CANWAIT) != 0 ? DMAR_PGF_WAITOK : 0);
|
|
|
|
if (error == ENOMEM) {
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_domain_unload_entry(entry, false);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (error);
|
|
|
|
}
|
|
|
|
KASSERT(error == 0,
|
2015-06-26 07:01:29 +00:00
|
|
|
("unexpected error %d from domain_map_buf", error));
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
|
|
|
|
return (0);
|
|
|
|
}
|
|
|
|
|
|
|
|
int
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_reserve_region(struct dmar_domain *domain, dmar_gaddr_t start,
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
dmar_gaddr_t end)
|
|
|
|
{
|
|
|
|
struct dmar_map_entry *entry;
|
|
|
|
int error;
|
|
|
|
|
2015-06-26 07:01:29 +00:00
|
|
|
entry = dmar_gas_alloc_entry(domain, DMAR_PGF_WAITOK);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
entry->start = start;
|
|
|
|
entry->end = end;
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_LOCK(domain);
|
|
|
|
error = dmar_gas_alloc_region(domain, entry, DMAR_GM_CANWAIT);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (error == 0)
|
|
|
|
entry->flags |= DMAR_MAP_ENTRY_UNMAPPED;
|
2015-06-26 07:01:29 +00:00
|
|
|
DMAR_DOMAIN_UNLOCK(domain);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
if (error != 0)
|
2015-06-26 07:01:29 +00:00
|
|
|
dmar_gas_free_entry(domain, entry);
|
Import the driver for VT-d DMAR hardware, as specified in the revision
1.3 of Intelб╝ Virtualization Technology for Directed I/O Architecture
Specification. The Extended Context and PASIDs from the rev. 2.2 are
not supported, but I am not aware of any released hardware which
implements them. Code does not use queued invalidation, see comments
for the reason, and does not provide interrupt remapping services.
Code implements the management of the guest address space per domain
and allows to establish and tear down arbitrary mappings, but not
partial unmapping. The superpages are created as needed, but not
promoted. Faults are recorded, fault records could be obtained
programmatically, and printed on the console.
Implement the busdma(9) using DMARs. This busdma backend avoids
bouncing and provides security against misbehaving hardware and driver
bad programming, preventing leaks and corruption of the memory by wild
DMA accesses.
By default, the implementation is compiled into amd64 GENERIC kernel
but disabled; to enable, set hw.dmar.enable=1 loader tunable. Code is
written to work on i386, but testing there was low priority, and
driver is not enabled in GENERIC. Even with the DMAR turned on,
individual devices could be directed to use the bounce busdma with the
hw.busdma.pci<domain>:<bus>:<device>:<function>.bounce=1 tunable. If
DMARs are capable of the pass-through translations, it is used,
otherwise, an identity-mapping page table is constructed.
The driver was tested on Xeon 5400/5500 chipset legacy machine,
Haswell desktop and E5 SandyBridge dual-socket boxes, with ahci(4),
ata(4), bce(4), ehci(4), mfi(4), uhci(4), xhci(4) devices. It also
works with em(4) and igb(4), but there some fixes are needed for
drivers, which are not committed yet. Intel GPUs do not work with
DMAR (yet).
Many thanks to John Baldwin, who explained me the newbus integration;
Peter Holm, who did all testing and helped me to discover and
understand several incredible bugs; and to Jim Harris for the access
to the EDS and BWG and for listening when I have to explain my
findings to somebody.
Sponsored by: The FreeBSD Foundation
MFC after: 1 month
2013-10-28 13:33:29 +00:00
|
|
|
return (error);
|
|
|
|
}
|