The faulting instruction needs to be restarted when the exception handler
is done handling the fault. bhyve now does this correctly by setting
'vmexit[vcpu].inst_length' to zero so the %rip is not advanced.
A minor complication is that the fault injection APIs are used by instruction
emulation code that is shared by vmm.ko and bhyve. Thus the argument that
refers to 'struct vm *' in kernel or 'struct vmctx *' in userspace needs to
be loosely typed as a 'void *'.
A nested exception condition arises when a second exception is triggered while
delivering the first exception. Most nested exceptions can be handled serially
but some are converted into a double fault. If an exception is generated during
delivery of a double fault then the virtual machine shuts down as a result of
a triple fault.
vm_exit_intinfo() is used to record that a VM-exit happened while an event was
being delivered through the IDT. If an exception is triggered while handling
the VM-exit it will be treated like a nested exception.
vm_entry_intinfo() is used by processor-specific code to get the event to be
injected into the guest on the next VM-entry. This function is responsible for
deciding the disposition of nested exceptions.
similar to -g.)
- Document -U to set the SMBIOS UUID.
- Add missing options to the usage output and to the manpage Synopsis.
- Don't claim that bvmdebug is amd64-only (it is also a device, not an
option).
Previously the sizes were inferred indirectly based on the size of the mappings
at 0 and 4GB respectively. This works fine as long as size of the allocation is
identical to the size of the mapping in the guest's address space. However, if
the mapping is disjoint then this assumption falls apart (e.g., due to the
legacy BIOS hole between 640KB and 1MB).
register pairs where two 32-bit registers make up a larger logical
size. Support those access by splitting the quad-word into two
double-words.
Reviewed by: grehan
to 16. This is arbitrary and is used to ensure that a vcpu goes back into
the vm_run() loop to process interrupts or rendezvous events in a timely
fashion.
Found with: Coverity Scan
CID: 1216436
it implicitly in vmm.ko.
Add ioctl VM_GET_CPUS to get the current set of 'active' and 'suspended' cpus
and display them via /usr/sbin/bhyvectl using the "--get-active-cpus" and
"--get-suspended-cpus" options.
This is in preparation for being able to reset virtual machine state without
having to destroy and recreate it.
fault on the destination buffer.
Prior to this change a page fault would be detected in vm_copyout(). This
was done after the I/O port access was done. If the I/O port access had
side-effects (e.g. reading the uart FIFO) then restarting the instruction
would result in incorrect behavior.
Fix this by validating the guest linear address before doing the I/O port
emulation. If the validation results in a page fault exception being injected
into the guest then the instruction can now be restarted without any
side-effects.
API function 'vie_calculate_gla()'.
While the current implementation is simplistic it forms the basis of doing
segmentation checks if the guest is in 32-bit protected mode.
of the guest linear address space. These APIs in turn use a new ioctl
'VM_GLA2GPA' to convert the guest linear address to guest physical.
Use the new copyin/copyout APIs when emulating ins/outs instruction in
bhyve(8).
'struct vm_guest_paging'.
Check for canonical addressing in vmm_gla2gpa() and inject a protection
fault into the guest if a violation is detected.
If the page table walk is restarted in vmm_gla2gpa() then reset 'ptpphys' to
point to the root of the page tables.
the UART FIFO.
The emulation is constrained in a number of ways: 64-bit only, doesn't check
for all exception conditions, limited to i/o ports emulated in userspace.
Some of these constraints will be relaxed in followup commits.
Requested by: grehan
Reviewed by: tychon (partially and a much earlier version)
the legacy 8259A PICs.
- Implement an ICH-comptabile PCI interrupt router on the lpc device with
8 steerable pins configured via config space access to byte-wide
registers at 0x60-63 and 0x68-6b.
- For each configured PCI INTx interrupt, route it to both an I/O APIC
pin and a PCI interrupt router pin. When a PCI INTx interrupt is
asserted, ensure that both pins are asserted.
- Provide an initial routing of PCI interrupt router (PIRQ) pins to
8259A pins (ISA IRQs) and initialize the interrupt line config register
for the corresponding PCI function with the ISA IRQ as this matches
existing hardware.
- Add a global _PIC method for OSPM to select the desired interrupt routing
configuration.
- Update the _PRT methods for PCI bridges to provide both APIC and legacy
PRT tables and return the appropriate table based on the configured
routing configuration. Note that if the lpc device is not configured, no
routing information is provided.
- When the lpc device is enabled, provide ACPI PCI link devices corresponding
to each PIRQ pin.
- Add a VMM ioctl to adjust the trigger mode (edge vs level) for 8259A
pins via the ELCR.
- Mark the power management SCI as level triggered.
- Don't hardcode the number of elements in Packages in the source for
the DSDT. iasl(8) will fill in the actual number of elements, and
this makes it simpler to generate a Package with a variable number of
elements.
Reviewed by: tycho
This has not added a lot of value when debugging bhyve issues while greatly
increasing the time and space required to store the core file.
Passing the "-C" option to bhyve(8) will change the default and dump guest
memory in the core dump.
Requested by: grehan
Reviewed by: grehan
Failing to do this will cause the kevent(2) notification to trigger
continuously and the bhyve(8) mevent thread will hog the cpu until the
characters on the backend tty device are drained.
Also, make the uart backend file descriptor non-blocking to avoid a
select(2) before every byte read from that backend.
Reviewed by: grehan
a 'hostcpu'. The new format of the argument string is "vcpu:hostcpu".
This allows pinning a subset of the vcpus if desired.
It also allows pinning a vcpu to more than a single 'hostcpu'.
Submitted by: novel (initial version)
However, if the original knote had been disabled then it is not automatically
re-enabled.
Fix this by using EV_ADD to create an mevent and EV_ENABLE to enable it.
Adding a kevent for the first time implicitly enables it so existing callers
of mevent_add() don't need to change.
Reviewed by: grehan
because there isn't a standard way to relay this information to the guest OS.
Add a command line option "-Y" to bhyve(8) to inhibit MPtable generation.
If the virtual machine is using PCI devices on buses other than 0 then it can
still use ACPI tables to convey this information to the guest.
Discussed with: grehan@
to sleep permanently by executing a HLT with interrupts disabled.
When this condition is detected the guest with be suspended with a reason of
VM_SUSPEND_HALT and the bhyve(8) process will exit.
Tested by executing "halt" inside a RHEL7-beta guest.
Discussed with: grehan@
Reviewed by: jhb@, tychon@
the 'HLT' instruction. This condition was detected by 'vm_handle_hlt()' and
converted into the SPINDOWN_CPU exitcode . The bhyve(8) process would exit
the vcpu thread in response to a SPINDOWN_CPU and when the last vcpu was
spun down it would reset the virtual machine via vm_suspend(VM_SUSPEND_RESET).
This functionality was broken in r263780 in a way that made it impossible
to kill the bhyve(8) process because it would loop forever in
vm_handle_suspend().
Unbreak this by removing the code to spindown vcpus. Thus a 'halt' from
a Linux guest will appear to be hung but this is consistent with the
behavior on bare metal. The guest can be rebooted by using the bhyvectl
options '--force-reset' or '--force-poweroff'.
Reviewed by: grehan@
by adding an argument to the VM_SUSPEND ioctl that specifies how the virtual
machine should be suspended, viz. VM_SUSPEND_RESET or VM_SUSPEND_POWEROFF.
The disposition of VM_SUSPEND is also made available to the exit handler
via the 'u.suspended' member of 'struct vm_exit'.
This capability is exposed via the '--force-reset' and '--force-poweroff'
arguments to /usr/sbin/bhyvectl.
Discussed with: grehan@
according to the method outlined in the AHCI spec.
Tested with FreeBSD 9/10/11 with MSI disabled,
and also NetBSD/amd64 (lightly).
Reviewed by: neel, tychon
MFC after: 3 weeks
Ignore writes, and return 0xff's, on config accesses when not set.
Behaviour now matches that seen on h/w.
Found with a NetBSD/amd64 guest.
Reviewed by: tychon
MFC after: 3 weeks
Status and Control register at port 0x61.
Be more conservative about "catching up" callouts that were supposed
to fire in the past by skipping an interrupt if it was
scheduled too far in the past.
Restore the PIT ACPI DSDT entries and add an entry for NMISC too.
Approved by: neel (co-mentor)
0xff. Some guests may attempt to read from this port to identify
psuedo-PNP ISA devices. (The ie(4) driver in FreeBSD/i386 is one
example.)
Reviewed by: grehan
Call through to /dev/random synchronously to fill
virtio buffers with RNG data.
Tested with FreeBSD-CURRENT and Ubuntu guests.
Submitted by: Leon Dang
Discussed with: markm
MFC after: 3 weeks
Sponsored by: Nahanni Systems
from any context i.e., it is not required to be called from a vcpu thread. The
ioctl simply sets a state variable 'vm->suspend' to '1' and returns.
The vcpus inspect 'vm->suspend' in the run loop and if it is set to '1' the
vcpu breaks out of the loop with a reason of 'VM_EXITCODE_SUSPENDED'. The
suspend handler waits until all 'vm->active_cpus' have transitioned to
'vm->suspended_cpus' before returning to userspace.
Discussed with: grehan
New ioctls VM_ISA_ASSERT_IRQ, VM_ISA_DEASSERT_IRQ and VM_ISA_PULSE_IRQ
can be used to manipulate the pic, and optionally the ioapic, pin state.
Reviewed by: jhb, neel
Approved by: neel (co-mentor)
This fixes the issue of bhyve appearing to halt when using
nmdm ports for the console, until a connection is made to
the other end.
bhyveload already does this.
Reported by: Many.
MFC after: 3 weeks.
processor-specific VMCS or VMCB. The pending exception will be delivered right
before entering the guest.
The order of event injection into the guest is:
- hardware exception
- NMI
- maskable interrupt
In the Intel VT-x case, a pending NMI or interrupt will enable the interrupt
window-exiting and inject it as soon as possible after the hardware exception
is injected. Also since interrupts are inherently asynchronous, injecting
them after the hardware exception should not affect correctness from the
guest perspective.
Rename the unused ioctl VM_INJECT_EVENT to VM_INJECT_EXCEPTION and restrict
it to only deliver x86 hardware exceptions. This new ioctl is now used to
inject a protection fault when the guest accesses an unimplemented MSR.
Discussed with: grehan, jhb
Reviewed by: jhb
simplify the implementation of the x2APIC virtualization assist in VT-x.
Prior to this change the vlapic allowed the guest to change its mode from
xAPIC to x2APIC. We don't allow that any more and the vlapic mode is locked
when the virtual machine is created. This is not very constraining because
operating systems already have to deal with BIOS setting up the APIC in
x2APIC mode at boot.
Fix a bug in the CPUID emulation where the x2APIC capability was leaking
from the host to the guest.
Ignore MMIO reads and writes to the vlapic in x2APIC mode. Similarly, ignore
MSR accesses to the vlapic when it is in xAPIC mode.
The default configuration of the vlapic is xAPIC. The "-x" option to bhyve(8)
can be used to change the mode to x2APIC instead.
Discussed with: grehan@
the non-standard zero capability list terminator. Instead, track
the start and end of the most recently added capability and use that
to adjust the previous capability's next pointer when a capability is
added and to determine the range of config registers belonging to
PCI capability registers.
Reviewed by: neel
This is done by representing each bus as root PCI device in ACPI. The device
implements the _BBN method to return the PCI bus number to the guest OS.
Each PCI bus keeps track of the resources that is decodes for devices
configured on the bus: i/o, mmio (32-bit) and mmio (64-bit). These windows
are advertised to the guest via the _CRS object of the root device.
Bus 0 is treated specially since it consumes the I/O ports to access the
PCI config space [0xcf8-0xcff]. It also decodes the legacy I/O ports that
are consumed by devices on the LPC bus. For this reason the LPC bridge can
be configured only on bus 0.
The bus number can be specified using the following command line option
to bhyve(8): "-s <bus>:<slot>:<func>,<emul>[,<config>]"
Discussed with: grehan@
Reviewed by: jhb@
the IDENTIFY DEVICE and IDENTIFY PACKET DEVICE commands.
Also, provide an indication a "D2H Register FIS" occurred during a SET FEATURES
command.
Approved by: grehan (co-mentor)
the virtio backends.
- Add a new ioctl to export the count of pins on the I/O APIC from vmm
to the hypervisor.
- Use pins on the I/O APIC >= 16 for PCI interrupts leaving 0-15 for
ISA interrupts.
- Populate the MP Table with I/O interrupt entries for any PCI INTx
interrupts.
- Create a _PRT table under the PCI root bridge in ACPI to route any
PCI INTx interrupts appropriately.
- Track which INTx interrupts are in use per-slot so that functions
that share a slot attempt to distribute their INTx interrupts across
the four available pins.
- Implicitly mask INTx interrupts if either MSI or MSI-X is enabled
and when the INTx DIS bit is set in a function's PCI command register.
Either assert or deassert the associated I/O APIC pin when the
state of one of those conditions changes.
- Add INTx support to the virtio backends.
- Always advertise the MSI capability in the virtio backends.
Submitted by: neel (7)
Reviewed by: neel
MFC after: 2 weeks
if it was above 4GB. This was seen with CentOS 6.5 guests with
large RAM, since the block drivers are loaded late in the
boot sequence and end up allocating descriptor memory from
high addresses.
Reported by: Michael Dexter
MFC after: 3 days
LPC devices. Among other things, the LPC serial ports now appear as
ACPI devices.
- Move the info for the top-level PCI bus into the PCI emulation code and
add ResourceProducer entries for the memory ranges decoded by the bus
for memory BARs.
- Add a framework to allow each PCI emulation driver to optionally write
an entry into the DSDT under the \_SB_.PCI0 namespace. The LPC driver
uses this to write a node for the LPC bus (\_SB_.PCI0.ISA).
- Add a linker set to allow any LPC devices to write entries into the
DSDT below the LPC node.
- Move the existing DSDT block for the RTC to the RTC driver.
- Add DSDT nodes for the AT PIC, the 8254 ISA timer, and the LPC UART
devices.
- Add a "SuperIO" device under the LPC node to claim "system resources"
aling with a linker set to allow various drivers to add IO or memory
ranges that should be claimed as a system resource.
- Add system resource entries for the extended RTC IO range, the registers
used for ACPI power management, the ELCR, PCI interrupt routing register,
and post data register.
- Add various helper routines for generating DSDT entries.
Reviewed by: neel (earlier version)
hides the setjmp/longjmp semantics of VM enter/exit. vmx_enter_guest() is used
to enter guest context and vmx_exit_guest() is used to transition back into
host context.
Fix a longstanding race where a vcpu interrupt notification might be ignored
if it happens after vmx_inject_interrupts() but before host interrupts are
disabled in vmx_resume/vmx_launch. We now called vmx_inject_interrupts() with
host interrupts disabled to prevent this.
Suggested by: grehan@
to SIGTERM when ACPI is enabled. Sending SIGTERM to the hypervisor when an
ACPI-aware OS is running will now trigger a soft-off allowing for a graceful
shutdown of the guest.
- Move constants for ACPI-related registers to acpi.h.
- Implement an SMI_CMD register with commands to enable and disable ACPI.
Currently the only change when ACPI is enabled is to enable the virtual
power button via SIGTERM.
- Implement a fixed-feature power button when ACPI is enabled by asserting
PWRBTN_STS in PM1_EVT when SIGTERM is received.
- Add support for EVFILT_SIGNAL events to mevent.
- Implement support for the ACPI system command interrupt (SCI) and assert
it when needed based on the values in PM1_EVT. Mark the SCI as active-low
and level triggered in the MADT and MP Table.
- Mark PCI interrupts in the MP Table as active-low in addition to level
triggered.
Reviewed by: neel
- Implement the PM1_EVT and PM1_CTL registers required by ACPI.
The PM1_EVT register is mostly a dummy as bhyve doesn't support any
of the hardware-initiated events. The only bit of PM1_CNT that is
implemented are the sleep request bits (SPL_EN and SLP_TYP) which
request a graceful power off for S5. In particular, for S5, bhyve
exits with a non-zero value which terminates the loop in vmrun.sh.
- Emulate the Reset Control register at I/O port 0xcf9 and advertise
it as the reset register via ACPI.
- Advertise an _S5 package.
- Extend the in/out interface to allow an in/out handler to request
that the hypervisor trigger a reset or power-off.
- While here, note that all vCPUs in a guest support C1 ("hlt").
Reviewed by: neel (earlier version)
- Add a generic routine to trigger an LVT interrupt that supports both
fixed and NMI delivery modes.
- Add an ioctl and bhyvectl command to trigger local interrupts inside a
guest. In particular, a global NMI similar to that raised by SERR# or
PERR# can be simulated by asserting LINT1 on all vCPUs.
- Extend the LVT table in the vCPU local APIC to support CMCI.
- Flesh out the local APIC error reporting a bit to cache errors and
report them via ESR when ESR is written to. Add support for asserting
the error LVT when an error occurs. Raise illegal vector errors when
attempting to signal an invalid vector for an interrupt or when sending
an IPI.
- Ignore writes to reserved bits in LVT entries.
- Export table entries the MADT and MP Table advertising the stock x86
config of LINT0 set to ExtInt and LINT1 wired to NMI.
Reviewed by: neel (earlier version)
state before the requested state transition. This guarantees that there is
exactly one ioctl() operating on a vcpu at any point in time and prevents
unintended state transitions.
More details available here:
http://lists.freebsd.org/pipermail/freebsd-virtualization/2013-December/001825.html
Reviewed by: grehan
Reported by: Markiyan Kushnir (markiyan.kushnir at gmail.com)
MFC after: 3 days
callers treat the MSI 'addr' and 'data' fields as opaque and also lets
bhyve implement multiple destination modes: physical, flat and clustered.
Submitted by: Tycho Nightingale (tycho.nightingale@pluribusnetworks.com)
Reviewed by: grehan@
vcpu and destroy its thread context. Also modify the 'HLT' processing to ignore
pending interrupts in the IRR if interrupts have been disabled by the guest.
The interrupt cannot be injected into the guest in any case so resuming it
is futile.
With this change "halt" from a Linux guest works correctly.
Reviewed by: grehan@
Tested by: Tycho Nightingale (tycho.nightingale@pluribusnetworks.com)
commit level triggered interrupts would work as long as the pin was not shared
among multiple interrupt sources.
The vlapic now keeps track of level triggered interrupts in the trigger mode
register and will forward the EOI for a level triggered interrupt to the
vioapic. The vioapic in turn uses the EOI to sample the level on the pin and
re-inject the vector if the pin is still asserted.
The vhpet is the first consumer of level triggered interrupts and advertises
that it can generate interrupts on pins 20 through 23 of the vioapic.
Discussed with: grehan@
bhyveload: introduce the -c <device> parameter
to select a tty for output (or "stdio")
bhyve: allow the puc and lpc-com backends to
accept a tty in addition to "stdio"
When used in conjunction with the null-modem device,
nmdm(4), this allows attach/detach to the guest console
and multiple concurrent serial ports. kgdb on a serial
port is now functional.
Reviewed by: neel
Requested by: Almost everyone that has used bhyve
MFC after: 10.0
Table is 22 bits, with the bit 31 being the interrupt-on-completion
bit.
OpenBSD and UEFI set this bit, resulting in large block i/o lengths
being sent to bhyve and coredumping the process. Fix by masking off
the relevant 22 bits when using the DBC field as a length.
Reviewed by: Zhixiang Yu
Discussed with: Tycho Nightingale (tycho.nightingale@pluribusnetworks.com)
MFC after: 10.0
actual value read by the guest from the device. The IOAPIC ID is now set to
zero in both MPtable/ACPI tables as well as in the ioapic device emulation.
Pointed out by: grehan@
bhyve supports a single timer block with 8 timers. The timers are all 32-bit
and capable of being operated in periodic mode. All timers support interrupt
delivery using MSI. Timers 0 and 1 also support legacy interrupt routing.
At the moment the timers are not connected to any ioapic pins but that will
be addressed in a subsequent commit.
This change is based on a patch from Tycho Nightingale (tycho.nightingale@pluribusnetworks.com).
to inject edge triggered legacy interrupts into the guest.
Start using the new API in device models that use edge triggered interrupts:
viz. the 8254 timer and the LPC/uart device emulation.
Submitted by: Tycho Nightingale (tycho.nightingale@pluribusnetworks.com)
upcoming in-kernel device emulations like the HPET.
The ioctls VM_IOAPIC_ASSERT_IRQ and VM_IOAPIC_DEASSERT_IRQ are used to
manipulate the ioapic pin state.
Discussed with: grehan@
Submitted by: Tycho Nightingale (tycho.nightingale@pluribusnetworks.com)
pin 2 of the IOAPIC.
Add an 'Interrupt Source Override' entry to the MADT to describe this
and start asserting interrupts on pin 2 in the 8254 device model.
Submitted by: Tycho Nightingale (tycho.nightingale@pluribusnetworks.com)
Remove the VM name from some of the thread-naming calls
since it is now in the proc title.
Slightly modify the thread-naming for the net and block
threads.
This improves readability when using top/ps with the -a
and -H options on a system with a large number of bhyve VMs.
Requested by: Michael Dexter
Reviewed by: neel
MFC after: 4 weeks
to a virtual machine then we implicitly create COM1 and COM2 ISA devices.
Prior to this change the only way of attaching a COM port to the virtual
machine was by presenting it as a PCI device that is mapped at the legacy
I/O address 0x3F8 or 0x2F8.
There were some issues with the original approach:
- It did not work at all with UEFI because UEFI will reprogram the PCI device
BARs and remap the COM1/COM2 ports at non-legacy addresses.
- OpenBSD GENERIC kernel does not create a /dev/console because it expects
the uart device at the legacy 0x3F8/0x2F8 address to be an ISA device.
- It was functional with a FreeBSD guest but caused the console to appear
on /dev/ttyu2 which was not intuitive.
The uart emulation is now independent of the bus on which it resides. Thus it
is possible to have uart devices on the PCI bus in addition to the legacy
COM1/COM2 devices behind the LPC bus.
The command line option to attach ISA COM1/COM2 ports to a virtual machine is
"-s <bus>,lpc -l com1,stdio".
The command line option to create a PCI-attached uart device is:
"-s <bus>,uart[,stdio]"
The command line option to create PCI-attached COM1/COM2 device is:
"-S <bus>,uart[,stdio]". This style of creating COM ports is deprecated.
Discussed with: grehan
Reviewed by: grehan
Submitted by: Tycho Nightingale (tycho.nightingale@pluribusnetworks.com)
M share/examples/bhyve/vmrun.sh
AM usr.sbin/bhyve/legacy_irq.c
AM usr.sbin/bhyve/legacy_irq.h
M usr.sbin/bhyve/Makefile
AM usr.sbin/bhyve/uart_emul.c
M usr.sbin/bhyve/bhyverun.c
AM usr.sbin/bhyve/uart_emul.h
M usr.sbin/bhyve/pci_uart.c
M usr.sbin/bhyve/pci_emul.c
M usr.sbin/bhyve/inout.c
M usr.sbin/bhyve/pci_emul.h
M usr.sbin/bhyve/inout.h
AM usr.sbin/bhyve/pci_lpc.c
AM usr.sbin/bhyve/pci_lpc.h
where a pin assertion while a source was masked would result in
the interrupt being lost, with the symptom being a console hang.
The condition is now recorded, and the interrupt generated when
the source is unmasked.
Discovered by: OpenBSD 5.4 MP
Reviewed by: neel
MFC after: 3 days
- Use #defines for capability bits
- Export the VTBLK_F_BLK_SIZE capability
- Fix bug in calculating capacity: it is in
512-byte units, not the underlying sector size
This allows virtio-blk to have backing devices
with non 512-byte sector sizes e.g. /dev/cd0, and
4K-block harddrives.
Reviewed by: neel
MFC after: 3 days
- remove assumption that the backing file/device had
512-byte sectors
- fix incorrect iovec size variable that would result
in a buffer overrun when an o/s issued an i/o request
with more s/g elements than the blockif api
Reviewed by: Zhixiang Yu (zxyu.core@gmail.com)
MFC after: 3 days
- Allow a hostbridge to be created with AMD as a vendor.
This passes the OpenBSD check to allow the use of MSI
on a PCI bus.
- Enable the i/o interrupt section of the mptable, and
populate it with unity ISA mappings. This allows the
'legacy' IRQ mappings of the PCI serial port to be
set up. Delete unused print routine that was obscuring code.
- Use the '-W' option to enable virtio single-vector MSI
rather than an environment variable. Update the virtio
net/block drivers to query this flag when setting up
interrupts.: bhyverun.c
- Fix the arithmetic used to derive the century byte in
RTC CMOS, as well as encoding it in BCD.
Reviewed by: neel
MFC after: 3 days
Linux writes to these nominally read-only registers,
so avoid having bhyve write warning messages to stdout
when the reg writes can be safely ignored. Change the
WPRINTF to DPRINTF which is conditional.
Reviewed by: mav
Discussed with: mav, Zhixiang Yu
MFC after: 3 days
'invpcid' instruction to the guest. Currently bhyve will try to enable this
capability unconditionally if it is available.
Consolidate code in bhyve to set the capabilities so it is no longer
duplicated in BSP and AP bringup.
Add a sysctl 'vm.pmap.invpcid_works' to display whether the 'invpcid'
instruction is available.
Reviewed by: grehan
MFC after: 3 days
annoying verbose boot error of the form
g_handleattr: vtbd0 bio_length 24 len 28 -> EFAULT
The ident returned by bhyve is a text string 'BHYVE-XXXX-XXXX', where
the X's are the first bytes of the md5 hash of the backing filename.
Reviewed by: neel
Approved by: re (gjb)
the 2 read-only bytes at the start of a PCI capability.
This is the sequence that OpenBSD uses when enabling
MSI interrupts, and works fine on real h/w.
In bhyve, convert the 4 byte write to a 2-byte write to
the r/w area past the first 2 r/o bytes of a capability.
Reviewed by: neel
Approved by: re@ (blanket)
"assignment makes pointer from integer without a cast", by changing the
cmd_lst and rbis members of struct ahci_port from integers to pointers.
Also surround a pow-of-2 test expression with parentheses to clarify it,
and avoid another gcc warning.
Approved by: re (glebius)
Reviewed by: grehan, mav
rather than opt-out.
Prior to this change if the "-g" option was not specified then a listening
socket for tunneling gdb packets would be opened at port 6466. If a second
virtual machine is fired up, also without the "-g" option, then that would
fail because there is already a listener on port 6466.
After this change if a gdb tunnel port needs to be created it needs to be
explicitly specified with a "-g <portnum>" command line option.
Reviewed by: grehan@
Approved by: re@ (blanket)
Make the amd64/pmap code aware of nested page table mappings used by bhyve
guests. This allows bhyve to associate each guest with its own vmspace and
deal with nested page faults in the context of that vmspace. This also
enables features like accessed/dirty bit tracking, swapping to disk and
transparent superpage promotions of guest memory.
Guest vmspace:
Each bhyve guest has a unique vmspace to represent the physical memory
allocated to the guest. Each memory segment allocated by the guest is
mapped into the guest's address space via the 'vmspace->vm_map' and is
backed by an object of type OBJT_DEFAULT.
pmap types:
The amd64/pmap now understands two types of pmaps: PT_X86 and PT_EPT.
The PT_X86 pmap type is used by the vmspace associated with the host kernel
as well as user processes executing on the host. The PT_EPT pmap is used by
the vmspace associated with a bhyve guest.
Page Table Entries:
The EPT page table entries as mostly similar in functionality to regular
page table entries although there are some differences in terms of what
bits are used to express that functionality. For e.g. the dirty bit is
represented by bit 9 in the nested PTE as opposed to bit 6 in the regular
x86 PTE. Therefore the bitmask representing the dirty bit is now computed
at runtime based on the type of the pmap. Thus PG_M that was previously a
macro now becomes a local variable that is initialized at runtime using
'pmap_modified_bit(pmap)'.
An additional wrinkle associated with EPT mappings is that older Intel
processors don't have hardware support for tracking accessed/dirty bits in
the PTE. This means that the amd64/pmap code needs to emulate these bits to
provide proper accounting to the VM subsystem. This is achieved by using
the following mapping for EPT entries that need emulation of A/D bits:
Bit Position Interpreted By
PG_V 52 software (accessed bit emulation handler)
PG_RW 53 software (dirty bit emulation handler)
PG_A 0 hardware (aka EPT_PG_RD)
PG_M 1 hardware (aka EPT_PG_WR)
The idea to use the mapping listed above for A/D bit emulation came from
Alan Cox (alc@).
The final difference with respect to x86 PTEs is that some EPT implementations
do not support superpage mappings. This is recorded in the 'pm_flags' field
of the pmap.
TLB invalidation:
The amd64/pmap code has a number of ways to do invalidation of mappings
that may be cached in the TLB: single page, multiple pages in a range or the
entire TLB. All of these funnel into a single EPT invalidation routine called
'pmap_invalidate_ept()'. This routine bumps up the EPT generation number and
sends an IPI to the host cpus that are executing the guest's vcpus. On a
subsequent entry into the guest it will detect that the EPT has changed and
invalidate the mappings from the TLB.
Guest memory access:
Since the guest memory is no longer wired we need to hold the host physical
page that backs the guest physical page before we can access it. The helper
functions 'vm_gpa_hold()/vm_gpa_release()' are available for this purpose.
PCI passthru:
Guest's with PCI passthru devices will wire the entire guest physical address
space. The MMIO BAR associated with the passthru device is backed by a
vm_object of type OBJT_SG. An IOMMU domain is created only for guest's that
have one or more PCI passthru devices attached to them.
Limitations:
There isn't a way to map a guest physical page without execute permissions.
This is because the amd64/pmap code interprets the guest physical mappings as
user mappings since they are numerically below VM_MAXUSER_ADDRESS. Since PG_U
shares the same bit position as EPT_PG_EXECUTE all guest mappings become
automatically executable.
Thanks to Alan Cox and Konstantin Belousov for their rigorous code reviews
as well as their support and encouragement.
Thanks for John Baldwin for reviewing the use of OBJT_SG as the backing
object for pci passthru mmio regions.
Special thanks to Peter Holm for testing the patch on short notice.
Approved by: re
Discussed with: grehan
Reviewed by: alc, kib
Tested by: pho
these.
The mux-vcpus option may return at some point, given it's utility
in finding bhyve (and FreeBSD) bugs.
Approved by: re@ (blanket)
Discussed with: neel@
https://wiki.freebsd.org/SummerOfCode2013/bhyveAHCI
This provides ICH8 SATA disk and ATAPI ports, selectable
via the bhyve slot command-line parameter:
SATA
-s <slot>,ahci-hd,<image-file>
ATAPI
-s <slot>,ahci-cd,<image-file>
Slight modifications by: grehan@
Approved by: re@ (blanket)
Obtained from: FreeBSD GSoC'13
timer support. This should be enough for the emulation of
h/w periodic timers (and no more) e.g. some of the 8254's
more esoteric modes that happen to be used by non-FreeBSD o/s's.
Approved by: re@ (blanket)
This should be sufficient for 10.0 and will do
until forthcoming work to avoid limitations
in this area is complete.
Thanks to Bela Lubkin at tidalscale for the
headsup on the apic/cpu id/io apic ASL parameters
that are actually hex values and broke when
written as decimal when 11 vCPUs were configured.
Approved by: re@
users to set the MAC address for a device.
Clean up some obsolete code in pci_virtio_net.c
Allow an error return from a PCI device emulation's init routine
to be propagated all the way back to the top-level and result in
the process exiting.
Submitted by: Dinakar Medavaram dinnu sun at gmail (original version)
If this capability is negotiated by the guest then the device will
generate an interrupt when it runs out of available tx/rx descriptors.
Reviewed by: grehan
Obtained from: NetApp
drop any frames that arrive while the device is starved for receive buffers.
This makes the receive path to only execute in context of the receive thread
and allows for further simplification.
Reviewed by: grehan
than blocking the vCPU thread. This improves bulk data performance
by ~30-40% and doesn't harm req/resp time for stock netperf runs.
Future work will use a thread pool rather than a thread per tx queue.
Submitted by: Dinakar Medavaram
Reviewed by: neel, grehan
Obtained from: NetApp
silently overwriting the previous assignment.
Gripe if the emulation is not recognized instead of silently ignoring the
emulated device.
If an error is detected by pci_parse_slot() then exit from the command line
parsing loop in main().
Submitted by (initial version): Chris Torek (chris.torek@gmail.com)
descriptors. Prior to this change the device would only work with guests
that chose to use indirect descriptors.
Modify the device reset callback to actually reset the device state.
Submitted by: Tycho Nightingale (tycho.nightingale@pluribusnetworks.com)
This was working by accident because:
- the RB_HEADs were being initialized to zero as part of BSS
- the pthread_rwlock functions were implicitly initializing the lock object
Obtained from: NetApp
- use clock_gettime(2) as the time base for the emulated ACPI timer instead
of directly using rdtsc().
- don't advertise the invariant TSC capability to the guest to discourage it
from using the TSC as its time base.
Discussed with: jhb@ (about making 'smp_tsc' a global)
Reported by: Dan Mack on freebsd-virtualization@
Obtained from: NetApp
- Respect the MEMEN and PORTEN bits in the command register
- Allow the guest to reprogram the address decoded by the BAR
Submitted by: Gopakumar T
Obtained from: NetApp
command line option "-m <memsize in MB>" to specify the memory size.
Prior to this change the user needed to explicitly specify the amount of
memory allocated below 4G (-m <lowmem>) and the amount above 4G (-M <highmem>).
The "-M" option is no longer supported by 'bhyveload' and 'bhyve'.
The start of the PCI hole is fixed at 3GB and cannot be directly changed
using command line options. However it is still possible to change this in
special circumstances via the 'vm_set_lowmem_limit()' API provided by
libvmmapi.
Submitted by: Dinakar Medavaram (initial version)
Reviewed by: grehan
Obtained from: NetApp
into the MSI-X table before using it to calculate the table index.
In the common case where the MSI-X table is located at the begining of the
BAR these two offsets are identical and thus the code was working by accident.
This change will fix the case where the MSI-X table is located in the middle
or at the end of the BAR that contains it.
Obtained from: NetApp
This seems prudent to do in its own right but it also opens up the possibility
of not having to mmap the entire guest address space in the 'bhyve' process
context.
Discussed with: grehan
Obtained from: NetApp
These set of ranges will be looked at if a standard memory
range isn't found, and won't be installed in the cache.
Use this to implement the memory behaviour of the PCI hole on
x86 systems, where writes are ignored and reads always return -1.
This allows breakpoints to be set when issuing a 'boot -d', which
has the side effect of accessing the PCI hole when changing the
PTE protection on kernel code, since the pmap layer hasn't been
initialized (a bug, but present in existing FreeBSD releases so
has to be handled).
Reviewed by: neel
Obtained from: NetApp
Prior to this change pinning was implemented via an ioctl (VM_SET_PINNING)
that called 'sched_bind()' on behalf of the user thread.
The ULE implementation of 'sched_bind()' bumps up 'td_pinned' which in turn
runs afoul of the assertion '(td_pinned == 0)' in userret().
Using the cpuset affinity to implement pinning of the vcpu threads works with
both 4BSD and ULE schedulers and has the happy side-effect of getting rid
of a bunch of code in vmm.ko.
Discussed with: grehan
the default.
The current behavior of advertising a single MSI vector can be requested by
setting the environment variable "BHYVE_USE_MSI" to "yes". The use of MSI
is not compliant with the virtio specification and will be eventually phased
out.
Submitted by: Gopakumar T
Obtained from: NetApp
can only be located at the beginning or the end of the BAR.
If the MSI-table is located in the middle of a BAR then we will split the
BAR into two and create two mappings - one before the table and one after
the table - leaving a hole in place of the table so accesses to it can be
trapped and emulated.
Obtained from: NetApp
devices are MSI-X capable. This in turn would lead it to treat bar 0 as
the MSI-X table bar even if the underlying device did not support MSI-X.
Fix this by providing an API to query the MSI-X table index of the emulated
device. If the underlying device does not support MSI-X then this API will
return -1.
Obtained from: NetApp
the default.
The current behavior of advertising a single MSI vector can be requested by
setting the environment variable "BHYVE_USE_MSI" to "true". The use of MSI
is not compliant with the virtio specification and will be eventually phased
out.
Submitted by: Gopakumar T
Obtained from: NetApp
statically. In most cases the number of table entries will be far less than
the maximum of 2048 allowed by the PCI specification.
Reuse macros from pcireg.h to interpret the MSI-X capability instead of rolling
our own.
Obtained from: NetApp
fill up to the uart's rx fifo size, and leave any remaining input
for when the rx fifo is read. This allows cut'n'paste of long lines
to be done into the bhyve console without truncation.
Also, introduce a mutex since the file input will run in the mevent
thread context and may corrupt state accessed by a vCPU thread.
Reviewed by: neel
Approved by: NetApp
With this change, dbench with >= 4 processes runs without getting
weird jumps forward in time when the APCI pmtimer is the default
timecounter.
Obtained from: NetApp
the guest to execute real or unpaged protected mode code - bhyve relies on
this feature to execute the AP bootstrap code.
Get rid of the hack that allowed bhyve to support SMP guests on processors
that do not have the "unrestricted guest" capability. This hack was entirely
FreeBSD-specific and would not work with any other guest OS.
Instead, limit the number of vcpus to 1 when executing on processors without
"unrestricted guest" capability.
Suggested by: grehan
Obtained from: NetApp
bhyve is intended to be a generic hypervisor, and not FreeBSD-specific.
(renaming internal routines will come later)
Reviewed by: neel
Obtained from: NetApp
On a nested page table fault the hypervisor will:
- fetch the instruction using the guest %rip and %cr3
- decode the instruction in 'struct vie'
- emulate the instruction in host kernel context for local apic accesses
- any other type of mmio access is punted up to user-space (e.g. ioapic)
The decoded instruction is passed as collateral to the user-space process
that is handling the PAGING exit.
The emulation code is fleshed out to include more addressing modes (e.g. SIB)
and more types of operands (e.g. imm8). The source code is unified into a
single file (vmm_instruction_emul.c) that is compiled into vmm.ko as well
as /usr/sbin/bhyve.
Reviewed by: grehan
Obtained from: NetApp
The -A option will create the minimal set of required ACPI tables in
guest memory. Since ACPI mandates an IOAPIC, the -I option must also
be used.
Template ASL files are created, and then passed to the iasl compiler
to generate AML files. These are then loaded into guest physical mem.
In support of this, the ACPI PM timer is implemented, in 32-bit mode.
Tested on 7.4/8.*/9.*/10-CURRENT.
Reviewed by: neel
Obtained from: NetApp
Discussed with: jhb (a long while back)
than waiting until AP bringup detects an out-of-range vCPU.
While here, fix all error output to use fprintf(stderr, ...
Reviewed by: neel
Reported by: @allanjude
Firmware tables require too much knowledge of system configuration,
and it's difficult to pass that information in general terms to a library.
The upcoming ACPI work exposed this - it will also livein bhyve.
Also, remove code specific to NetApp from the mptable name, and remove
the -n option from bhyve.
Reviewed by: neel
Obtained from: NetApp
- New memory region interface. An RB tree holds the regions,
with a last-found per-vCPU cache to deal with the common case
of repeated guest accesses to MMIO registers in the same page.
- Support memory-mapped BARs in PCI emulation.
mem.c/h - memory region interface
instruction_emul.c/h - remove old region interface.
Use gpa from EPT exit to avoid a tablewalk to
determine operand address. Determine operand size
and use when calling through to region handler.
fbsdrun.c - call into region interface on paging
exit. Distinguish between instruction emul error
and region not found
pci_emul.c/h - implement new BAR callback api.
Split BAR alloc routine into routines that
require/don't require the BAR phys address.
ioapic.c
pci_passthru.c
pci_virtio_block.c
pci_virtio_net.c
pci_uart.c - update to new BAR callback i/f
Reviewed by: neel
Obtained from: NetApp
AP needs to be activated by spinning up an execution context for it.
The local apic emulation is now completely done in the hypervisor and it will
detect writes to the ICR_LO register that try to bring up the AP. In response
to such writes it will return to userspace with an exit code of SPINUP_AP.
Reviewed by: grehan
the guest. Prior to the fix it was possible for such a bar to appear as a
32-bit bar as long as it was allocated from the region below 4GB.
This had the potential to confuse some drivers that were particular about
the size of the bars.
Obtained from: NetApp
These function number is specified by an optional [:<func>] after the slot
number: -s 1:0,virtio-net,tap0
Ditto for the mptable naming: -n 1:0,e0a
Obtained from: NetApp
or 32-bit signed integer.
Simplify the handling of indirect addressing with displacement by
unconditionally adding the 'instruction->disp' to the target address.
This is alright since 'instruction->disp' is non-zero only for the
addressing modes that specify a displacement.
Obtained from: NetApp
be activated as part of the slot config options.
The syntax is:
-s <slotnum>,uart[,stdio]
The stdio parameter instructs the code to perform i/o using
stdin/stdout. It can only be used for one instance.
To allow legacy i/o ports/irqs to be used, a new variant of
the slot command, -S, is introduced. When used to specify a
slot, the device will use legacy resources if it supports
them; otherwise it will be treated the same as the '-s' option.
Specifying the -S option with the uart will first use the 0x3f8/irq 4
config, and the second -S will use 0x2F8/irq 3.
Interrupt delivery is awaiting the arrival of the i/o apic code,
but this works fine in uart(4)'s polled mode.
This code was written by Cynthia Lu @ MIT while an intern at NetApp,
with further work from neel@ and grehan@.
Obtained from: NetApp
Includes instruction emulation for memory r/w access. This
opens the door for io-apic, local apic, hpet timer, and
legacy device emulation.
Submitted by: ryan dot berryhill at sandvine dot com
Reviewed by: grehan
Obtained from: Sandvine
run as a 1/2 CPU guest on an 8.1 bhyve host.
bhyve/inout.c
inout.h
fbsdrun.c
- Rather than exiting on accesses to unhandled i/o ports, emulate
hardware by returning -1 on reads and ignoring writes to unhandled
ports. Support the previous mode by allowing a 'strict' parameter
to be set from the command line.
The 8.1 guest kernel was vastly cut down from GENERIC and had no
ISA devices. Booting GENERIC exposes a massive amount of random
touching of i/o ports (hello syscons/vga/atkbdc).
bhyve/consport.c
dev/bvm/bvm_console.c
- implement a simplistic signature for the bvm console by returning
'bv' for an inw on the port. Also, set the priority of the console
to CN_REMOTE if the signature was returned. This works better in
an environment where multiple consoles are in the kernel (hello syscons)
bhyve/rtc.c
- return 0 for the access to RTC_EQUIPMENT (yes, you syscons)
amd64/vmm/x86.c
x86.h
- hide a bunch more CPUID leaf 1 bits from the guest to prevent
cpufreq drivers from probing.
The next step will be to move CPUID handling completely into
user-space. This will allow the full spectrum of changes from
presenting a lowest-common-denominator CPU type/feature set, to
exposing (almost) everything that the host can support.
Reviewed by: neel
Obtained from: NetApp
vmm.ko - kernel module for VT-x, VT-d and hypervisor control
bhyve - user-space sequencer and i/o emulation
vmmctl - dump of hypervisor register state
libvmm - front-end to vmm.ko chardev interface
bhyve was designed and implemented by Neel Natu.
Thanks to the following folk from NetApp who helped to make this available:
Joe CaraDonna
Peter Snyder
Jeff Heller
Sandeep Mann
Steve Miller
Brian Pawlowski
