[efi] Do not attempt EFI_USB_IO_PROTOCOL transfers during shutdown
On at least some platforms (observed with a Raspberry Pi), any attempt
to perform USB transfers via EFI_USB_IO_PROTOCOL during EFI shutdown
will lock up the system. This is quite probably due to the already
documented failure of all EFI timers when ExitBootServices() is
called: see e.g. commit 5cf5ffea2 "[efi] Work around temporal anomaly
encountered during ExitBootServices()".
Work around this problem by refusing to poll endpoints if shutdown is
in progress, and by immediately failing any attempts to enqueue new
transfers.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[efi] Report failed control transfers as expected by the USB core
The USB core reuses the I/O buffer space occupied by the USB setup
packet to hold the completion status for message transfers, assuming
that the message() method will always strip the setup packet before
returning. This assumption is correct for all of the hardware
controller drivers (XHCI, EHCI, and UHCI), since these drivers are
able to enqueue the transfer as a separate action from waiting for the
transfer to complete.
The EFI_USB_IO_PROTOCOL does not allow us to separate actions in this
way: there is only a single blocking method that both enqueues and
waits for completion. Our usbio driver therefore currently defers
stripping the setup packet until the control endpoint is polled.
This causes a bug if a message transfer is enqueued but never polled
and is subsequently cancelled, since the cancellation will be reported
with the I/O buffer still containing the setup packet. This breaks
the assumption that the setup packet has been stripped, and triggers
an assertion failure in usb_control_complete().
Fix by always stripping the setup packet in usbio_endpoint_message(),
and adjusting usbio_control_poll() to match.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[coverity] Override assumptions about wcrtomb() and hmac_init()
Newer versions of Coverity use built-in models for wcrtomb() and
hmac_init() that are capable of returning errors, and reports defects
due to code failing to check for these errors. The actual iPXE
implementations are simpler than Coverity's models and can never
return errors, so these defects are false positives.
Fix by overriding Coverity's built-in models for these functions.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[crypto] Drag in configured digestInfo prefixes for any use of RSA
Ensure that the configured RSA digestInfo prefixes are included in any
build that includes rsa.o (rather than relying on x509.o or tls.o also
being present in the final binary).
This allows the RSA self-tests to be run in isolation.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[tls] Add missing call to tls_tx_resume() when restarting negotiation
The restart of negotiation triggered by a HelloRequest currently does
not call tls_tx_resume() and so may end up leaving the connection in
an idle state in which the pending ClientHello is never sent.
Fix by calling tls_tx_resume() as part of tls_restart(), since the
call to tls_tx_resume() logically belongs alongside the code that sets
bits in tls->tx_pending.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[peerdist] Limit number of concurrent raw block downloads
Raw block downloads are expensive if the origin server uses HTTPS,
since each concurrent download will require local TLS resources
(including potentially large received encrypted data buffers).
Raw block downloads may also be prohibitively slow to initiate when
the origin server is using HTTPS and client certificates. Origin
servers for PeerDist downloads are likely to be running IIS, which has
a bug that breaks session resumption and requires each connection to
go through the full client certificate verification.
Limit the total number of concurrent raw block downloads to ameliorate
these problems.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[peerdist] Start block download timers from within opener methods
Move the responsibility for starting the block download timers from
peerblk_expired() to peerblk_raw_open() and peerblk_retrieval_open(),
in preparation for adding the ability to defer calls to
peerblk_raw_open() via a block download queue.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[build] Add predefined shortcut for Raspberry Pi builds
Add a build shortcut "rpi", allowing for e.g.
make CONFIG=rpi CROSS=aarch64-linux-gnu- bin-arm64-efi/rpi.efi
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[build] Move predefined all-drivers build shortcut to Makefile
The (very approximate) split between Makefile.housekeeping and
Makefile is that the former provides mechanism and the latter provides
policy.
Provide a section within Makefile as a home for predefined build
shortcuts such as the existing all-drivers build.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[build] Do not apply WORKAROUND_CFLAGS for host compiler
The WORKAROUND_CFLAGS list is constructed based on running tests on
the target compiler, and the results may not be valid for the host
compiler.
The only relevant workaround required for the host compiler is
-Wno-stringop-truncation, which is needed to avoid a spurious compiler
warning for a totally correct usage of strncpy() in util/elf2efi.c.
Duplicating the workaround tests for the host compiler is messy, as is
conditionally applying __attribute__((nonstring)). Fix instead by
disapplying WORKAROUND_CFLAGS for the host compiler, and using
memcpy() with an explicitly calculated length instead of strncpy() in
util/elf2efi.c.
Reported-by: Ignat Korchagin <ignat@cloudflare.com>
Reported-by: Christopher Clark <christopher.w.clark@gmail.com>
Signed-off-by: Michael Brown <mcb30@ipxe.org>
Compiling with gcc 9.1 generates lots of "taking address of packed
member of ... may result in an unaligned pointer value" warnings.
Some of these warnings are genuine, and indicate correctly that parts
of iPXE currently require the CPU (or runtime environment) to support
unaligned accesses. For example: the TCP/IP receive data path will
attempt to access 32-bit fields that may not be aligned to a 32-bit
boundary.
Other warnings are either spurious (such as when the pointer is to a
variable-length byte array, which can have no alignment requirement
anyway) or unhelpful (such as when the pointer is used solely to
provide a debug colour value for the DBGC() macro).
There appears to be no easy way to silence the spurious warnings.
Since the ability to perform unaligned accesses is already a
requirement for iPXE, work around the problem by silencing this class
of warnings.
Signed-off-by: Valentine Barshak <gvaxon@gmail.com>
Modified-by: Michael Brown <mcb30@ipxe.org>
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[build] Fix "'%s' directive argument is null" error
Use '%p' directive, and print handle's address if the address is null
and the handle doesn't have a name. This fixes the following
compilation error:
interface/efi/efi_debug.c:334:3: error: '%s' directive
argument is null [-Werror=format-overflow=]
Signed-off-by: Valentine Barshak <gvaxon@gmail.com>
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[fdt] Add ability to parse a MAC address from a flattened device tree
The Raspberry Pi NIC has no EEPROM to hold the MAC address. The
platform firmware (e.g. UEFI or U-Boot) will typically obtain the MAC
address from the VideoCore firmware and add it to the device tree,
which is then made available to subsequent programs such as iPXE or
the Linux kernel.
Add the ability to parse a flattened device tree and to extract the
MAC address.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[efi] Return only registered EFI devices from efidev_parent()
efidev_parent() currently assumes that any device with BUS_TYPE_EFI is
part of a struct efi_device. This assumption is not valid, since the
code in efi_device_info() may also create a device with BUS_TYPE_EFI.
Fix by searching through the list of registered EFI devices when
looking for a match, instead of relying on the bus type value.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[arm] Provide dummy implementations for {in,out}[s]{b,w,l}
It is currently not possible to build the all-drivers iPXE binaries
for ARM, since there is no implementation for inb(), outb(), etc.
There is no common standard for accessing I/O space on ARM platforms,
and there are almost no ARM-compatible peripherals that actually
require I/O space accesses.
Provide dummy implementations that behave as though no device is
present (i.e. ignore writes, return all bits high for reads). This is
sufficient to allow the all-drivers binaries to link, and should cause
drivers to behave as though no I/O space peripherals are present in
the system.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[build] Fix use of inline assembly on GCC 8 ARM64 builds
Commit 1a7746603 ("[build] Fix use of inline assembly on GCC 4.8 ARM64
builds") switched from using "%c0" to "%a0" in order to avoid an
"invalid operand prefix" error on the ARM64 version of GCC 4.8.
It appears that the ARM64 version of GCC 8 now produces an "invalid
address mode" error for the "%a0" form, but is happy with the original
"%c0" form.
Switch back to using the "%c0" form, on the assumption that the
requirement for "%a0" was a temporary aberration.
Originally-fixed-by: John L. Jolly <jjolly@suse.com>
Signed-off-by: Michael Brown <mcb30@ipxe.org>
The physical function defaults to operating in "PXE mode" after a
power-on reset. In this mode, receive descriptors are fetched and
written back as single descriptors. In normal (non-PXE mode)
operation, receive descriptors are fetched and written back only as
complete cachelines unless an interrupt is raised.
There is no way to return to PXE mode from non-PXE mode, and there is
no way for the virtual function driver to operate in PXE mode.
Choose to operate in non-PXE mode. This requires us to trick the
hardware into believing that it is raising an interrupt, so that it
will not defer writing back receive descriptors until a complete
cacheline (i.e. four packets) have been consumed. We do so by
configuring the hardware to use MSI-X with a dummy target location in
place of the usual APIC register.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[intelxl] Split out ring creation from context programming
The virtual function driver will use the same transmit and receive
descriptor ring structures, but will not itself construct and program
the ring context. Split out ring creation and destruction from the
programming of the ring context, to allow code to be shared between
physical and virtual function drivers.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[intelxl] Allow for arbitrary placement of ring tail registers
The virtual function transmit and receive ring tail register offsets
do not match those of the physical function. Allow the tail register
offsets to be specified separately.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
The physical function driver does not allow the virtual function to
request the use of 16-byte receive descriptors. Switch to using
32-byte receive descriptors.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[intelxl] Provide a mechanism for handling "send to VF" events
Provide a weak stub function for handling the "send to VF" event used
for communications between the physical and virtual function drivers.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[intelxl] Allow admin cookie to hold extended opcode and return code
The "send to PF" and "send to VF" admin queue descriptors (ab)use the
cookie field to hold the extended opcode and return code values.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
A virtual function reset is triggered via an admin queue command and
will reset the admin queue configuration registers. Allow the admin
queues to be reinitialised after such a reset, without requiring the
overhead (and potential failure paths) of freeing and reallocating the
queues.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[intelxl] Use one admin queue buffer per admin queue descriptor
We currently use a single data buffer shared between all admin queue
descriptors. This works for the physical function driver since we
have at most one command in progress and only a single event (which
does not use a data buffer).
The communication path between the physical and virtual function
drivers uses the event data buffer, and there is no way to prevent a
solicited event (i.e. a response to a request) from being overwritten
by an unsolicited event (e.g. a link status change).
Provide individual data buffers for each admin event queue descriptor
(and for each admin command queue descriptor, for the sake of
consistency).
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[intelxl] Allow for virtual function admin queue register maps
The register map for the virtual functions appears to have been
constructed using a random number generator.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[intelxl] Use VLAN tag in receive descriptor if present
The physical function driver does not allow the virtual function to
request that VLAN tags are left unstripped. Extract and use the VLAN
tag from the receive descriptor if present.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[vlan] Provide vlan_netdev_rx() and vlan_netdev_rx_err()
The Hermon driver uses vlan_find() to identify the appropriate VLAN
device for packets that are received with the VLAN tag already
stripped out by the hardware. Generalise this capability and expose
it for use by other network card drivers.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
The Intel 40 Gigabit Ethernet virtual functions support only MSI-X
interrupts, and will write back completed interrupt descriptors only
when the device attempts to raise an interrupt (or when a complete
cacheline of receive descriptors has been completed).
We cannot actually use MSI-X interrupts within iPXE, since we never
have ownership of the APIC. However, an MSI-X interrupt is
fundamentally just a DMA write of a single dword to an arbitrary
address. We can therefore configure the device to "raise" an
interrupt by writing a meaningless value to an otherwise unused memory
location: this is sufficient to trigger the receive descriptor
writeback logic.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[ocsp] Accept response certID with missing hashAlgorithm parameters
One of the design goals of ASN.1 DER is to provide a canonical
serialization of a data structure, thereby allowing for equality of
values to be tested by simply comparing the serialized bytes.
Some OCSP servers will modify the request certID to omit the optional
(and null) "parameters" portion of the hashAlgorithm. This is
arguably legal but breaks the ability to perform a straightforward
bitwise comparison on the entire certID field between request and
response.
Fix by comparing the OID-identified hashAlgorithm separately from the
remaining certID fields.
Originally-fixed-by: Thilo Fromm <Thilo@kinvolk.io>
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[tcp] Display "connecting" status until connection is established
Provide increased visibility into the progress of TCP connections by
displaying an explicit "connecting" status message while waiting for
the TCP handshake to complete.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
[tls] Display cross-certificate and OCSP status messages
TLS connections will almost always create background connections to
perform cross-signed certificate downloads and OCSP checks. There is
currently no direct visibility into which checks are taking place,
which makes troubleshooting difficult in the absence of either a
packet capture or a debug build.
Use the job progress message buffer to report the current cross-signed
certificate download or OCSP status check, where applicable.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
Record the session ID (if any) provided by the server and attempt to
reuse it for any concurrent connections to the same server.
If multiple connections are initiated concurrently (e.g. when using
PeerDist) then defer sending the ClientHello for all but the first
connection, to allow time for the first connection to potentially
obtain a session ID (and thereby speed up the negotiation for all
remaining connections).
Signed-off-by: Michael Brown <mcb30@ipxe.org>
On a Dell OptiPlex 7010, calling DisconnectController() on the LOM
device handle will lock up the system. Debugging shows that execution
is trapped in an infinite loop that is somehow trying to reconnect
drivers (without going via ConnectController()).
The problem can be reproduced in the UEFI shell with no iPXE code
present, by using the "disconnect" command. Experimentation shows
that the only fix is to unload (rather than just disconnect) the
"Ip4ConfigDxe" driver.
Add the concept of a blacklist of UEFI drivers that will be
automatically unloaded when iPXE runs as an application, and add the
Dell Ip4ConfigDxe driver to this blacklist.
Signed-off-by: Michael Brown <mcb30@ipxe.org>
Michael Brown