EoIB is a fairly simple protocol in which raw Ethernet frames
(excluding the CRC) are encapsulated within Infiniband Unreliable
Datagrams, with a four-byte fixed EoIB header (which conveys no actual
information). The Ethernet broadcast domain is provided by a
multicast group, similar to the IPoIB IPv4 multicast group.
The mapping from Ethernet MAC addresses to Infiniband address vectors
is achieved by snooping incoming traffic and building a peer cache
which can then be used to map a MAC address into a port GID. The
address vector is completed using a path record lookup, as for IPoIB.
Note that this requires every packet to include a GRH.
Add basic support for EoIB devices. This driver is substantially
derived from the IPoIB driver. There is currently no mechanism for
automatically creating EoIB devices.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[infiniband] Make IPoIB support configurable at build time
Add a build configuration option VNIC_IPOIB to control whether or not
IPoIB support is included for Infiniband devices.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[ipoib] Resimplify test for received broadcast packets
Commit e62e52b ("[ipoib] Simplify test for received broadcast
packets") relies upon the multicast LID being present in the
destination address vector as passed to ipoib_complete_recv().
Unfortunately, this information is not present in many Infiniband
devices' completion queue entries.
Fix by testing instead for the presence of a multicast GID.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[pcbios] Restrict external memory allocations to the low 4GB
When running the 64-bit BIOS version of iPXE, restrict external memory
allocations to the low 4GB to ensure that allocations (such as for
initrds) fall within our identity-mapped memory region, and will be
accessible to the potentially 32-bit operating system.
Move largest_memblock() back to memtop_umalloc.c, since this change
imposes a restriction that applies only to BIOS builds.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[infiniband] Avoid multiple calls to ib_cmrc_shutdown()
When a CMRC connection is closed, the deferred shutdown process calls
ib_destroy_qp(). This will cause the receive work queue entries to
complete in error (since they are being cancelled), which will in turn
reschedule the deferred shutdown process. This eventually leads to
ib_destroy_conn() being called on a connection that has already been
freed.
Fix by explicitly cancelling any pending shutdown process after the
shutdown process has completed.
Ironically, this almost exactly reverts commit 019d4c1 ("[infiniband]
Use a one-shot process for CMRC shutdown"); prior to the introduction
of one-shot processes the only way to achieve a one-shot process was
for the process to cancel itself.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[librm] Add support for running in 64-bit long mode
Add support for running the BIOS version of iPXE in 64-bit long mode.
A 64-bit BIOS version of iPXE can be built using e.g.
make bin-x86_64-pcbios/ipxe.usb
make bin-x86_64-pcbios/8086100e.mrom
The 64-bit BIOS version should appear to function identically to the
normal 32-bit BIOS version. The physical memory layout is unaltered:
iPXE is still relocated to the top of the available 32-bit address
space. The code is linked to a virtual address of 0xffffffffeb000000
(in the negative 2GB as required by -mcmodel=kernel), with 4kB pages
created to cover the whole of .textdata. 2MB pages are created to
cover the whole of the 32-bit address space.
The 32-bit portions of the code run with VIRTUAL_CS and VIRTUAL_DS
configured such that truncating a 64-bit virtual address gives a
32-bit virtual address pointing to the same physical location.
The stack pointer remains as a physical address when running in long
mode (although the .stack section is accessible via the negative 2GB
virtual address); this is done in order to simplify the handling of
interrupts occurring while executing a portion of 32-bit code with
flat physical addressing via PHYS_CODE().
Interrupts may be enabled in either 64-bit long mode, 32-bit protected
mode with virtual addresses, 32-bit protected mode with physical
addresses, or 16-bit real mode. Interrupts occurring in any mode
other than real mode will be reflected down to real mode and handled
by whichever ISR is hooked into the BIOS interrupt vector table.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
In a 64-bit build, the entirety of the 32-bit address space is
identity-mapped and so any valid physical address may immediately be
used as a virtual address. Conversely, a virtual address that is
already within the 32-bit address space may immediately be used as a
physical address.
A valid virtual address that lies outside the 32-bit address space
must be an address within .textdata, and so can be converted to a
physical address by adding virt_offset.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[librm] Mark virt_offset, text16, data16, rm_cs, and rm_ds as constant
The physical locations of .textdata, .text16 and .data16 are constant
from the point of view of C code. Mark the relevant variables as
constant to allow gcc to optimise out redundant reads.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
No callers of prot_to_phys, phys_to_prot, or intr_to_prot require the
flags to be preserved. Remove the unnecessary pushfl/popfl pairs.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[librm] Add phys_call() wrapper for calling code with physical addressing
Add a phys_call() wrapper function (analogous to the existing
real_call() wrapper function) for calling code with flat physical
addressing, and use this wrapper within the PHYS_CODE() macro.
Move the relevant functionality inside librm.S, where it more
naturally belongs.
The COMBOOT code currently uses explicit calls to _virt_to_phys and
_phys_to_virt. These will need to be rewritten if our COMBOOT support
is ever generalised to be able to run in a 64-bit build.
Specifically:
- com32_exec_loop() should be restructured to use PHYS_CODE()
- com32_wrapper.S should be restructured to use an equivalent of
prot_call(), passing parameters via a struct i386_all_regs
- there appears to be no need for com32_wrapper.S to switch between
external and internal stacks; this could be omitted to simplify
the design.
For now, librm.S continues to expose _virt_to_phys and _phys_to_virt
for use by com32.c and com32_wrapper.S. Similarly, librm.S continues
to expose _intr_to_virt for use by gdbidt.S.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[build] Fix building on older versions of binutils
Some older versions of binutils have issues with both the use of
PROVIDE() and the interpretation of numeric literals within a section
description.
Work around these older versions by defining the required numeric
literals outside of any section description, and by automatically
determining whether or not to generate extra space for page tables
rather than relying on LDFLAGS.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
The bulk of the iPXE binary (the .textdata section) is physically
relocated at runtime to the top of the 32-bit address space in order
to allow space for an OS to be loaded. The relocation is achieved
with the assistance of segmentation: we adjust the code and data
segment bases so that the link-time addresses remain valid.
Segmentation is not available (for normal code and data segments) in
long mode. We choose to compile the C code with -mcmodel=kernel and
use a link-time address of 0xffffffffeb000000. This choice allows us
to identity-map the entirety of the 32-bit address space, and to alias
our chosen link-time address to the physical location of our .textdata
section. (This requires the .textdata section to always be aligned to
a page boundary.)
We simultaneously choose to set the 32-bit virtual address segment
bases such that the link-time addresses may simply be truncated to 32
bits in order to generate a valid 32-bit virtual address. This allows
symbols in .textdata to be trivially accessed by both 32-bit and
64-bit code.
There is no (sensible) way in 32-bit assembly code to generate the
required R_X86_64_32S relocation records for these truncated symbols.
However, subtracting the fixed constant 0xffffffff00000000 has the
same effect as truncation, and can be represented in a standard
R_X86_64_32 relocation record. We define the VIRTUAL() macro to
abstract away this truncation operation, and apply it to all
references by 32-bit (or 16-bit) assembly code to any symbols within
the .textdata section.
We define "virt_offset" for a 64-bit build as "the value to be added
to an address within .textdata in order to obtain its physical
address". With this definition, the low 32 bits of "virt_offset" can
be treated by 32-bit code as functionally equivalent to "virt_offset"
in a 32-bit build.
We define "text16" and "data16" for a 64-bit build as the physical
addresses of the .text16 and .data16 sections. Since a physical
address within the 32-bit address space may be used directly as a
64-bit virtual address (thanks to the identity map), this definition
provides the most natural access to variables in .text16 and .data16.
Note that this requires a minor adjustment in prot_to_real(), which
accesses .text16 using 32-bit virtual addresses.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[librm] Transition to protected mode within init_librm()
Long-mode operation will require page tables, which are too large to
sensibly fit in our .data16 segment in base memory.
Add a portion of init_librm() running in 32-bit protected mode to
provide access to high memory. Use this portion of init_librm() to
initialise the .textdata variables "virt_offset", "text16", and
"data16", eliminating the redundant (re)initialisation currently
performed on every mode transition as part of real_to_prot().
Signed-off-by: Michael Brown <mcb30@ipxe.org>
Use the standard "pushl $function ; pushw %cs ; call prot_call"
sequence everywhere that prot_call() is used.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[bios] Make uses of REAL_CODE() and PHYS_CODE() 64-bit clean
On a 64-bit CPU, any modification of a register by 32-bit or 16-bit
code will destroy the invisible upper 32 bits of the corresponding
64-bit register. For example: a 32-bit "pushl %eax" followed by a
"popl %eax" will zero the upper half of %rax. This differs from the
treatment of upper halves of 32-bit registers by 16-bit code: a
"pushw %ax" followed by a "popw %ax" will leave the upper 16 bits of
%eax unmodified.
Inline assembly generated using REAL_CODE() or PHYS_CODE() will
therefore have to preserve the upper halves of all registers, to avoid
clobbering registers that gcc expects to be preserved.
Output operands from REAL_CODE() and PHYS_CODE() assembly may
therefore contain undefined values in the upper 32 bits.
Fix by using explicit variable widths (e.g. uint32_t) for
non-discarded output operands, to ensure that undefined values in the
upper 32 bits of 64-bit registers are ignored.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
Michael Brown