+2022-04-18 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.8.0.
+
+ * Filters can now match devices based on partially specified
+ class code and also on the programming interface.
+
+ * Reporting of link speeds, power limits, and virtual function tags
+ has been updated to the current PCIe specification.
+
+ * We decode the Data Object Exchange capability.
+
+ * Bus mapping mode works in non-zero domains.
+
+ * pci_fill_info() can fetch more fields: bridge bases, programming
+ interface, revision, subsystem vendor and device ID, OS driver,
+ and also parent bridge. Internally, the implementation was rewritten,
+ significantly reducing the number of corner cases to be handled.
+
+ * The Windows port was revived and greatly improved by Pali Rohár.
+ It requires less magic to compile. More importantly, it runs on both
+ old and recent Windows systems (see README.Windows for details).
+
+ * Added a new Windows back-end using the cfgmgr32 interface.
+ It does not provide direct access to the configuration space,
+ but basic information about the device is reported via pci_fill_info().
+ For back-ends of this type, we now provide an emulated read-only
+ config space.
+
+ * If the configuration space is not readable for some reason
+ (e.g., the cfgmgr32 back-end, but also badly implemented sleep mode
+ of some devices), lspci prints only information provided by the OS.
+
+ * The Hurd back-end was greatly improved thanks to Joan Lledó.
+
+ * Various minor bug fixes and improvements.
+
+ * We officially require a working C99 compiler. Sorry, MSVC.
+
+ * As usually, updated pci.ids to the current snapshot of the database.
+
+2020-05-31 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.7.0.
+
+ * Added or improved the following capabilities: Designated Vendor-Specific,
+ Compute eXpress Link, Resizable BARs, VF Resizable BARs, Link
+ Capabilities 2, Link Status 2.
+
+ * On Linux, lspci can show IOMMU groups.
+
+ * setpci can be asked to skip bus scan and operate on a device
+ completely specified by its domain/bus/dev/func address. This
+ involved major internal cleanup.
+
+ * The above feature of setpci uses the pci_get_dev() function,
+ which obtains a struct pci_dev without doing a bus scan. This was
+ always possible, but apparently little used, because back-ends
+ frequently choked when operating on such devices. Fixed a lot
+ of minor bugs related to this.
+
+ * Also, back-ends which do not support domains now correctly fail when
+ trying to access devices outside domain 0.
+
+ * Semantics of pci_fill_info() and pci_dev->known_fields was underspecified,
+ which lead to inconsistencies between back-ends. Improved documentation
+ to give a more precise definition and updated all back-ends to conform
+ to it. Most importantly, pci_dev->known_fields shows all fields requested
+ over the lifetime of the pci_dev, but never those which are not supported
+ by the back-end.
+
+ * As usually, updated pci.ids to the current snapshot of the database.
+
+2020-01-25 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.6.4.
+
+ * A new back-end for the GNU Hurd was contributed by Joan Lledó.
+
+ * When printing VPD item identifiers, non-ASCII characters are escaped.
+
+2020-01-22 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.6.3.
+
+ * `lspci -t' (tree mode) can be combined with `-s' to show a sub-tree.
+ We also fixed potential buffer overflows in the tree dumper.
+
+ * Cleaned messy code for dumping of I/O, memory, and ROM regions.
+ This helped fixing a bug, which caused some 64-bit regions to be
+ reported as virtual. All flags are now printed after the address
+ (previously, "[virtual]" and "[enhanced]" were before it for no good
+ reason).
+
+ * Added pci_find_cap_nr() to the library, which handles capabilities
+ which occur multiple times in a single device.
+
+ * Minor improvements in printing of PCIe capabilities.
+
+ * We now decode the Multicast and Secondary PCI Express extended
+ capabilities.
+
+ * The list of capability names available to setpci was updated.
+
+ * Minor bugs were fixed in FreeBSD and Solaris ports.
+
+ * We now prefer HTTPS URLs in all documentation
+
+ * The pci.ids file has a man page.
+
+ * As usually, updated pci.ids to the current snapshot of the database.
+
+2018-08-12 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.6.2.
+
+ * Added "-P" and "-PP" switches to lspci, which display the path
+ through bridges to each device.
+
+ * Fixed a couple of bugs in computation of bus topology. It was
+ previously used only for the tree display, but we re-use it for
+ computing the paths.
+
+ * As usually, updated pci.ids to the current snapshot of the database.
+
+2018-07-12 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.6.1.
+
+ * Fixed compilation issues on Linux systems with MUSL libc.
+
+2018-06-30 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.6.0.
+
+ * BARs reported by the OS, but not set on the device itself are
+ reliably marked with "[virtual]".
+
+ * Library: Introduced a generic mechanism of string properties. This
+ avoids lots of special cases and makes ABI compatibility easier.
+
+ * On Linux systems with OpenFirmware, report corresponding device tree nodes
+ as device properties.
+
+ * VPD decoder knows several non-standard extensions.
+
+ * When PCIe link speed is less than the maximum supported by the
+ device, it is explicitly marked as "downgraded".
+
+ * Several new capabilities are not decoded yet, but at least their
+ names are printed.
+
+ * The Null capability is easily decoded.
+
+ * Formatting of several capabilities was cleaned up.
+
+ * The VGA16 bit in the bridge control register is now supported.
+
+ * Added a port to SylixOS. Thanks to YuJian Gong for contribution.
+
+ * Added a port to DOS/DJGPP. Thanks to Rudolf Marek for contribution.
+
+ * The order in which back-ends are probed was decoupled from the
+ internal back-end IDs. This helps, because new back-ends must have
+ their ID allocated at the end to keep the ABI, but they might need
+ to be probed earlier.
+
+ * The fbsd-device back-end should work again.
+
+ * Fixed a couple of bugs. Most notably, DeviceName was not printed.
+
+ * As usually, updated pci.ids to the current snapshot of the database.
+
+2017-11-17 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.5.6.
+
+ * Improvements of the FreeBSD back-end: read-only access for non-root,
+ support DragonFly BSD, support extended config space.
+
+ * MN VPD keyword is decoded correctly.
+
+ * As usually, updated pci.ids to the current snapshot of the database.
+
+2017-07-05 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.5.5.
+
+ * Better decoding of AER capability.
+
+ * "Slot Implemented" flag is decoded for PCI/PCI-X to PCIe bridges.
+
+ * Minor fixes of decoding other capabilities.
+
+ * As usually, updated pci.ids to the current snapshot of the database.
+
+2017-02-25 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.5.4.
+
+ * Previous version broke compilation on systems, for which lib/types.h
+ did not provide a 64-bit integer type. It is provided everywhere now.
+
+2017-02-15 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.5.3.
+
+ * When lspci looks for Linux kernel modules, it uses the default
+ path to module directory provided by libkmod. Previously,
+ it tried to construct the path explicitly, which need not
+ work on all systems.
+
+ * Improved formatting of memory and I/O ranges behind a bridge.
+
+ * PCIe link capabilities now display GEN4 speed (16GT/s).
+
+ * PCIe device capabilities now show bits related to atomic operations.
+ Thanks to Satanand Burla for a patch.
+
+ * As usually, updated pci.ids to the current snapshot of the database.
+
+2016-10-03 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.5.2.
+
+ * The L1 power management capability is now decoded more
+ thoroughly. Thanks to Rajat Jain for the patch.
+
+ * The table of configuration registers used by setpci
+ had a bug in the definition of SUBSYSTEM_VENDOR_ID.
+
+2016-05-22 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.5.1.
+
+ * Fixed symbol versioning of pci_init().
+
+2016-05-19 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.5.0.
+
+ * New capabilities decoded: Downstream Port Containment,
+ Precision Time Measurement. Thanks to Keith Busch and
+ Jonathan Yong.
+
+ * Domain numbers extended to 31 bits. This will be used by the
+ Linux kernel on some machines in near future.
+
+ * Enhanced allocation regions are now decoded on Linux.
+
+ * The NetBSD back-end supports PCIe extended configuration space.
+
+ * Updated pci.ids to the current snapshot of the database.
+
+2016-01-03 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.4.1.
+
+ * New capabilities decoded: Process Address Space, Page Request
+ Interface, Enhanced Allocation. Thanks to David Daney and
+ David Woodhouse.
+
+ * DevCap SlotPowerLimit is now decoded for all components with
+ upstream ports.
+
+ * Database mirror at SourceForge moved to GitHub.
+
+ * Updated pci.ids to the current snapshot of the database.
+
+2015-09-14 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.4.0.
+
+ * On Linux, we report NUMA nodes of devices.
+
+ * The sysfs back-end does not die on read errors
+ of optional attributes. Instead, a warning is produced.
+
+ * Fixed several minor bugs.
+
+ * Updated pci.ids to the current snapshot of the database.
+
+2015-04-09 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.3.1.
+
+ * Removed hacks for backward compatibility with Linux libc5,
+ which were breaking newer non-glibc Linux systems. Thanks
+ to Felix Janda.
+
+ * Display VirtIO vendor-specific capability. Patch by Gerd
+ Hoffmann.
+
+ * Fixed memory leak in name cache.
+
+ * Updated pci.ids to the current snapshot of the database.
+
+2014-11-10 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.3.0.
+ (celebrating a one-year anniversary of the previous version :))
+
+ * Device names exported by BIOS are displayed on Linux.
+
+ * On Linux systems, HWDB is used to look up device names
+ when our ID database gives no match. (More precisely,
+ HWDB is consulted after local pci.ids, but before using
+ network to query online pci.ids.) Thanks to Tom Gundersen
+ for the initial patch.
+
+ * Added experimental back-end for OS X / Darwin. Thanks to
+ Richard Yao for providing it.
+
+ * Filters now support matching by device class. Original
+ patch by Matthew Wilcox, wrappers for ABI compatibility
+ by me.
+
+ * Interrupt Pin and Interrupt Line registers are displayed
+ for bridge devices, too.
+
+ * Several portability bugs have been fixed.
+
+ * Several typos have been fixed. Also, use of questionable
+ constructs in man pages has been reduced.
+
+ * PCIe link capabilities now include the ASPMOptComp bit.
+
+ * The "CRS Software Visibility" bit is now decoded properly.
+
+ * Updated pci.ids to the current snapshot of the database.
+
+2013-11-10 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.2.1.
+
+ * CardBus bridge capabilities are displayed.
+
+ * PCIe L1 PM substates are decoded.
+
+ * Various bugs were fixed in decoding of PCIe capabilities.
+
+ * The sysfs back-end does not spit out unnecessary warnings when
+ empty slots report only a partial device address. This actually
+ happens on IBM pSeries.
+
+ * Updated pci.ids to the today's snapshot of the database.
+
+2013-04-19 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.2.0.
+
+ * On newer Linux systems, we use libkmod to look up kernel modules
+ (modules.pcimap no longer exists). To facilitate this, libpci
+ is able to look up module aliases in sysfs.
+
+ * Various minor bug fixes.
+
+ * Updated pci.ids to the today's snapshot of the database.
+
+2012-06-25 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.10.
+
+ * Decoding of LTR/OBFF in PCIe capabilities.
+
+ * Various minor bug fixes.
+
+ * Updated pci.ids to the today's snapshot of the database.
+
+2012-01-14 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.9.
+
+ * Updated README.
+
+ * Wherever we mention the PCI ID database, we now refer to
+ http://pci-ids.ucw.cz/ and the sf.net site is mentioned only
+ as a mirror. This includes update-pciids.
+
+ * Decode PCIe Gen 3 speeds and link status fields.
+
+ * Various minor bug fixes.
+
+ * Updated pci.ids to the today's snapshot of the database.
+
+2011-10-02 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.8.
+
+ * More capabilities: Transaction Processing Hints, Latency Tolerance
+ Reporting. Thanks to Jesse Barnes.
+
+ * Added BeOS and Haiku ports. Contributed by Francois Revol.
+
+ * pciutils.pc now uses Libs.private properly.
+
+ * When we format a name and it does not fit in the buffer, we truncate
+ it instead of returning "buffer too small" instead. This works on all
+ platforms with sane (i.e., C99-compatible) snprintf().
+
+ * Various minor bug fixes.
+
+ * Updated pci.ids to the today's snapshot of the database.
+
+2010-01-31 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.7.
+
+ * Minor improvements and bug fixes in decoding of the Virtual Channel
+ capability.
+
+2010-01-24 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.6.
+
+ * More capabilities decoded: Virtual Channel (except arbitration
+ tables), Root Complex Link, Vendor-Specific (header only), SATA HBA.
+
+ * All extended capabilities have their version displayed (-vv or more).
+
+2010-01-19 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.5.
+
+ * Updated pci.ids to the today's snapshot of the database.
+
+ * When scanning extended capabilities, properly mask the lowest 2 bits,
+ which are currently reserved. This avoids unaligned access errors on
+ broken hardware (see tests/broken-ecaps).
+
+ * Large bar sizes are displayed in human-readable format (with units).
+ Thanks to Matthew Wilcox.
+
+ * Physical slot information is displayed correctly for multi-function cards.
+ Fixed by Matthew Wilcox.
+
+ * Fixed a couple of typos everywhere.
+
+ * Library: Fixed bugs in freeing of capabilities.
+
+ * Windows back-end compiles again.
+
+2009-08-14 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.4.
+ (sorry that I have missed 22/7 and released it later :-))
+
+ * Updated pci.ids to the today's snapshot of the database.
+
+ * Fixed memory and file descriptor leak in the dump back-end.
+
+ * The SR-IOV capability decoder now prints the VF BAR's.
+ Patch by Chris Wright, cleaned up by me.
+
+ * On request of certain company's lawyers, we now include a copy
+ of the GPL with our package. It seems that the pciutils are getting
+ mature if the most important bug of the month was this one ;-)
+
+2009-07-04 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.3.
+
+ * Updated pci.ids to the current snapshot of the database.
+
+ * The VPD parser now reports unknown and vendor-defined items
+ properly. It also stops on any item in unknown format, avoiding long
+ output on bogus VPD data. Thanks to Ben Hutchings and Matthew Wilcox.
+
+ * The MSI-X table size now matches the spec. Thanks to Michael S.
+ Tsirkin.
+
+ * The Power Management capability now includes the soft reset bit.
+ Thanks to Yu Zhao.
+
+ * Decoding of the Advanced Features capability has been added.
+ Thanks to Yu Zhao.
+
+ * The whole package compiles on GNU/kFreeBSD again.
+
+ The following patches have been contributed by Matthew Wilcox:
+
+ * The procfs back-end is able to cope with /proc/bus/pci
+ containing names with domains, which occur on sparc64 and
+ possibly other architectures due to a kernel bug.
+
+ * The sysfs back-end no longer complains when a slot address
+ is missing, which happens with old versions of Linux fakephp.
+
+ * The Device Serial Number capability is printed in the right
+ byte order.
+
+ * The MSI and MSI-X capabilities are printed in a prettier way.
+
+ * The tree output mode (`lspci -t') shows domain numbers only
+ at the root, which makes the output more compact.
+
+ * Updated documentation on the bus mapping mode (`lspci -M').
+
+2009-02-01 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.2.
+
+ * Fixed another silly bug in the command-line parser of setpci.
+
+2009-01-30 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.1.
+
+ * Updated pci.ids to the current snapshot of the database.
+
+ * The configure script now sets LC_ALL to avoid being fooled by
+ locale-dependent behavior of `tr'.
+
+ * The command-line parser of setpci did sometimes segfault on invalid
+ input. Thanks to Andreas Wiese for a fix.
+
+2009-01-18 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.1.0.
+
+ * Updated pci.ids to the current snapshot of the database.
+
+ * The Cygwin backend now works on Windows Vista. Thanks to Jonathan
+ Kotta.
+
+ * Fixed a bug in decoding of the SR-IOV capability. Patch by Yu Zhao.
+
+ * Details of some PCIe capabilities are displayed only with -vv.
+
+ * When a BAR is reported by the OS, but not by the device (i.e.,
+ it is marked as [virtual] in lspci), the [disabled] flag is
+ suppressed, because it does not make sense in such cases.
+ Patch by Yu Zhao.
+
+2008-12-13 Martin Mares <mj@ucw.cz>
+
+ * The source code of lspci has been split to multiple files, hopefully
+ making it easier to maintain.
+
+ * The library and lspci now know about physical slot names. So far,
+ they are provided by the sysfs back-end only. Thanks go to Alex Chiang.
+
+ * When a device has the VPD (Vital Product Data) capability and the
+ VPD data are supplied by the OS, they are decoded and printed in the
+ verbose mode. This currently works only on Linux with the sysfs
+ back-end. Thanks to Ben Hutchings of Solarflare for the patch.
+
+ * `setpci --version' now works properly.
+
+ * `setpci --dumpregs' prints a table of all known names of
+ registers and capabilities. This replaces the table of registers
+ in the setpci man page.
+
+ * The dry-run mode of setpci gives better feedback.
+
+ * The setpci utility is now able to address registers stored in PCI
+ capabilities (actually it allows a more general form of relative
+ addressing).
+
+ * The library has gained functions for working with PCI capabilities.
+
+ * Address Translation Services capability is now decoded. Patch by
+ Yu Zhao.
+
+2008-11-09 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.0.3.
+
+ * `lspci -k' now displays the subsystem ID, too. This makes `-k'
+ show everything needed to identify the device and the available
+ drivers, which was called for by many users.
+
+ * Fixed spelling of MSI. Patch by Matthew Wilcox.
+
+ * Better support for cross-compilation. Thanks to Alon Bar-Lev
+ for the patch.
+
+ * Fixed printing of the AER capability. Patch by Max Asbock.
+
+ * HT 1.02 capabilities are decoded as HT 1.03. Suggested by
+ Carl-Daniel Hailfinger.
+
+ * Fixed Cygwin build. Thanks to Steve Elliott for reporting the bug.
+
+ * Updated pci.ids to the current snapshot of the database.
+
+2008-09-19 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.0.2.
+
+ * Fixed a minor bug in the configure script, which caused warnings
+ about redefinition of symbols during compilation.
+
+2008-09-11 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.0.1.
+
+ * Updated pci.ids to the most recent snapshot.
+
+ * Added a Cygwin port. Patch by Christopher Voltz, ported to the
+ current tree by Martin Mares.
+
+ * Worked around compatibility problems with various default settings
+ of wget (we have to set --no-timestamping explicitly). Thanks to Ville
+ Skytta for pointing that out.
+
+ * Fixed printing of MSI capabilities. Thanks to Matthew Wilcox for
+ a patch.
+
+ * Added decoding of several PCI-X capabilities: device/link/slot 2,
+ Advanced Error Reporting, Access Control Services, Alternative Routing-ID,
+ Single Root I/O Virtualization. Thanks to Yu Zhao for patches.
+
+ * Fixed bug in filters which caused them to refuse vendor/device ID 0xffff.
+
+ * README: The linux-pci mailing list has moved to linux-pci@vger.kernel.org.
+
+ * Fixed several build problems: builds without PCI_USE_DNS on Linux
+ and with PCI_USE_DNS on Solaris and *BSD. Static library mode also
+ compiles again on old versions of GCC.
+
+2008-04-10 Martin Mares <mj@ucw.cz>
+
+ * Released as 3.0.0.
+
+ * Updated API and ABI version.
+
+2008-02-20 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.99.1-alpha2.
+
+ * Changed the default domain for the DNS resolver.
+
+2008-02-18 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.99.1-alpha1.
+
+ * The makefile system has been reworked. All configuration settings
+ are now passed to the configure script in environment variables,
+ allowing for easy tweaking in the top-level Makefile. All control
+ knobs are now described in the README.
+
+ * The libpci can be built as a shared library with properly restricted
+ symbol exports. Use `make SHARED=yes' to enable that or `make SHARED=local'
+ for a local testing build (with hardwired paths to the library, so that
+ it does not need installation).
+
+ * The example program has been moved from lib/example.c to the top-level
+ directory, because it should be built exactly as the other utilities
+ are. It has been also improved slightly to educate better.
+
+ * The i386-ports method is enabled on Linux/x86_64 as well.
+
+2008-02-13 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.10-net2.
+
+ * Support for resolving of PCI ID's using DNS together with a local
+ cache of resolved entries has been added. See the `-q' and `-Q' options
+ of lspci.
+
+ * The library now has a generic system of settable parameters, which
+ also include settings of the DNS resolver and cache. An `-O' option
+ has been added to lspci and setpci to allow setting of these options.
+
+ * Configuration of the access methods are now specified by the new
+ parameter system, replacing the pci_access->method_params array.
+
+ * Access methods now also have sensible names and help texts and it
+ is possible to look up method ID by a name.
+
+ * An `-A' switch has been added to both lspci and setpci, allowing to
+ select an arbitrary access method. The `-P' switch (configure proc
+ backend) has been removed as it is no longer needed and I do not know
+ any its user.
+
+ * Several source files have been split for better maintainability
+ (most notably the resolving of ID's).
+
+ * Man pages and help texts have been updated. A new man page `pcilib(7)'
+ has been added and description of library options has been moved there.
+
+ * When an unknown device ID is encountered, we print `Device <id>'
+ instead of `Unknown device <id>'. It uses less space and it also
+ should reduce the number of inexperienced users complaining that
+ the device is not supported by the OS. To lookup up OS drivers,
+ use the `-k' option.
+
+ * PCI_LIB_VERSION has been bumped to 0x020299.
+
+ * Makefile: stripping of the binaries during installation can be
+ overridden by the STRIP variable.
+
+ * lib/types.h: We use the integer types from <stdint.h> if the
+ compiler claims C99 support.
+
+2008-02-11 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.10.
+
+ * lspci.c, setpci.c: Cleaned up the list of options.
+
+ * lib/names.c: Fix displaying of errors reported by zlib.
+ Previously, the buffer containing the error message had
+ been deallocated by gzclose() before the message was printed.
+
+2008-01-07 Martin Mares <mj@ucw.cz>
+
+ * update-pciids.sh: Added quiet mode (-q). Clean up uncompressed
+ files left by previous versions of the pciutils. Patch by Mike
+ Frysinger.
+
+ * update-pciids.man: Mention the -q switch.
+
+2007-11-29 Martin Mares <mj@ucw.cz>
+
+ * lib/dump.c: Squashed compiler warnings about code with
+ no effect (there really were surplus *'s).
+
+2007-11-06 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.9.
+
+ * lspci.c: Added a new switch `-k' which requests printing
+ of information on kernel drivers attached to each device
+ and on kernel modules reporting the ability to handle the
+ device. So far, this is supported only on Linux with the
+ sysfs back-end, so it is implemented internally in the lspci
+ instead of the libpci. Thanks to Anicka <anicka@anicka.net>
+ for help.
+
+2007-10-19 Martin Mares <mj@ucw.cz>
+
+ * Makefile, lib/Makefile: Moved -lz from LDFLAGS to LDLIBS.
+ Also added an explicit pattern rule for linking to make sure
+ that LDLIBS is used on all platforms. Thanks to Dan Nicholson
+ for the suggestion.
+
+2007-10-19 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.8.
+
+ * pci.ids: Revised class codes to match Conventional PCI 3.0 specs.
+ Added a couple of new ones, renumbered `ADMA continuous operation'
+ prog-if to 0x30 (even the old SATA Class Code ECN doesn't mention
+ 0x40) and renumbered the satellite communication controllers.
+
+ * lib/header.h: Include `PCI hot-plug' and `Secure device'
+ capabilities from PCI 3.0 specs. Also added `SATA HBA' and
+ `Advanced features' caps from various ECN's.
+
+ * lspci.c: All known capabilities have at least their name displayed
+ now. When we are unable to decode them completely, we signalize it
+ with a `<?>' mark.
+
+ * lspci.man: Document `<?>' and also mention that extended config
+ space is currently available only with the linux_sysfs back-end.
+
+ * lspci.c: Decode the Debug port capability (per EHCI 0.96 spec).
+
+ * lspci.c: Big code cleanup: re-arranged functions in the code,
+ renamed everything related to capabilities to cap_* and
+ all options except verbose to opt_*.
+
+2007-10-14 Martin Mares <mj@ucw.cz>
+
+ * lib/[fno]bsd-*: Removed extraneous braces.
+
+2007-10-12 Martin Mares <mj@ucw.cz>
+
+ * Capability loop detection introduced 2.2.7 did not work
+ properly with extended capabilities. Fixed.
+
+2007-10-05 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.7.
+
+ * lspci.c (show_caps, show_ext_caps): Detect and report loops in
+ capability lists.
+
+ * lspci.c, lib/header.h: Finished decoding of the PCI Express
+ capability. The extended capabilities remain undecoded for now,
+ but at least the list of them has been updated to reflect the
+ current PCI Express 2.0 spec.
+
+ * lspci.c, lib/header.h: Decode new bits of traditional registers
+ as defined by PCIE / PCI-X. This includes discard timers in the bridge
+ control register and INTx enable/status in device control/status
+ registers.
+
+ * lib/fbsd-device.c: Support domains on new FreeBSD's. Contributed
+ by Marius Strobl.
+
+2007-09-12 Hasso Tepper <hasso@estpak.ee>
+
+ * Extended the fbsd-device backend to run on Dragonfly BSD.
+
+ * lspci.c: alloca() is declared in <stdlib.h> on BSD's, not <alloca.h>.
+
+2007-09-03 Martin Mares <mj@ucw.cz>
+
+ * Resurrected the Windows port, including cross-compilation by MinGW.
+ Patch by Samuel Bronson <naesten@gmail.com>.
+
+2007-08-31 Martin Mares <mj@ucw.cz>
+
+ * Makefile, lib/Makefile: `ar' and `ranlib' can be overridden to allow
+ cross-compilation.
+
+2007-08-27 Martin Mares <mj@ucw.cz>
+
+ * lib/names.c (pci_open): When calling gzopen(), use "rb" file mode
+ instead of "r". This is needed on DOS systems, where this function
+ somewhat illogically uses the binary flag for the compressed file
+ instead of the decompressed stream inside, where binariness really
+ matters.
+
+2007-08-14 Martin Mares <mj@ucw.cz>
+
+ * lspci.c (show_ht): Added decoding of Hypertransport MSI mapping capability,
+ based on a patch by Jason Gunthorpe.
+
+ * tests/cap-MSI-mapping: Added a test case. I plan to add test cases
+ (which are dumps of config space) for all new features.
+
+2007-06-20 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.6.
+
+ * Makefile: Added an "install-lib" target. Thanks to Dan Nicholson
+ for a patch.
+
+ * Makefile, lib/Makefile: Generate and install pkg-config file
+ for libpci. Again by Dan Nicholson.
+
+2007-06-20 Thomas Schwinge <tschwinge@gnu.org>
+
+ * lib/i386-io-hurd.h: Rewritten for new Hurd kernels.
+
+2007-05-04 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.5.
+
+ * pci.ids: Updated to the current snapshot of the database.
+
+2007-02-14 Martin Mares <mj@ucw.cz>
+
+ * lspci.c (show_express): Added PCI/PCI-X to PCI-Express Bridge type.
+ Patch by Mark Glines.
+
+2007-02-09 Martin Mares <mj@ucw.cz>
+
+ * pci.ids: Updated to the current snapshot of the database.
+
+2007-02-06 Martin Mares <mj@ucw.cz>
+
+ * Replaced bzero() by memset() everywhere, it's better to lose a tiny
+ bit of readability than maintain hacks to make it work on various systems.
+
+ * lib/configure: tr on Solaris is a bit weird and it requires `[A-Z]'
+ instead of `A-Z'. Fortunately, the extra brackets don't hurt otherwise.
+
+ * lib/types.h, lib/configure: Solaris should use <stdint.h> to get precise
+ integer types.
+
+2007-02-04 Martin Mares <mj@ucw.cz>
+
+ * lspci.c: alloca() needs <alloca.h>.
+
+2006-09-17 Martin Mares <mj@ucw.cz>
+
+ * lib/dump.c: Fixed a couple of bugs in the dump backend which caused
+ devices with domains or with extended config space to be read incorrectly.
+ Also, dumps with partial lines are allowed now.
+
+2006-09-10 Martin Mares <mj@ucw.cz>
+
+ * pci.ids: Updated to the current database snapshot.
+
+ * lspci.c (scan_device): If an error occurs when reading the standard config
+ header of a device, report it and ignore the device and return with exit
+ code 2 at the end.
+
+2006-09-09 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.4.
+
+ * maint/release: Remind the maintainer about the current PCI_LIB_VERSION.
+
+ * lib/pci.h: Updated PCI_LIB_VERSION.
+
+ * lspci.c (show_machine): In the `-m' mode, do proper shell escaping
+ if the fields printed contain quotes or backslashes.
+
+ * lspci.c (show_machine): Added a `-vmm' mode, which removes the misuse
+ of the `Device' tag for two different purposes.
+
+ * Makefile: Moved all system-dependent decisions to lib/configure,
+ since config.mk is included in the top-level Makefile anyway.
+
+ * lib/configure: When configuring for Linux, ignore the kernel version
+ and always build all access methods.
+
+ * Makefile (CFLAGS): Removed -Winline, it's not needed and triggers
+ at many non-interesting places with gcc 3.4.
+
+ * Whitespace cleanup of all source files.
+
+ * The pci.ids file can be stored compressed if zlib is available.
+ Added transparent decompression to lib/names.c, modified update-pciids.sh
+ to keep the file compressed if applicable, updated Makefiles.
+ Based on a patch by Matthew Wilcox, but all the bugs are mine.
+
+ * Makefile, README: Allow installation to be done to a different directory
+ from the one we will eventually be placed in by setting DESTDIR.
+ Patch by Matthew Wilcox.
+
+ * Added .gitignore files.
+
+ * Makefile (clean): Clean *.orig, too.
+
+ * Cleaned up usage of `char' and `byte'.
+
+ * lib/header.h: PCI_CLASS_DOCKING_OTHER should be 0x0a80, not 0x0a01.
+
+ * lib/header.h: Added PCI_CLASS_xxx constants for all classes currently
+ defined in the pci.ids file.
+
+2006-08-01 Martin Mares <mj@ucw.cz>
+
+ * lib/i386-io-hurd.h: Fixed a silly typo.
+
+2006-07-30 Martin Mares <mj@ucw.cz>
+
+ * lib/Makefile: Added a missing rule for obsd-device.o.
+
+ * lspci.c, lib/header.h: Added support for MSI per-vector masking.
+ Contributed by Petr Vandrovec.
+
+ * lspci.c, lib/header.h: Added support for the `bridge subsystem ID'
+ capability. Contributed by Petr Vandrovec.
+
+ * lspci.c (show_htype1): Fixed bug in printing of 64-bit prefetchable
+ memory windows. Fix by Petr Vandrovec.
+
+ * maint/release.pm: Exclude .git directory from releases.
+
+ * lib/i386-ports.c, lib/i386-io-*: Report failures during port access
+ initialization by a->warn() instead of calling printf() or perror()
+ directly.
+
+ * lib/i386-ports.c, lib/i386-io-*: Moved the logic which keeps track of
+ the port access state to generic code.
+
+ * lib/i386-io-hurd.h: Ask the kernel for I/O port access appropriately.
+ Contributed by Thomas Schwinge and Samuel Thibault.
+
+ * lib/i386-ports.c: Define _GNU_SOURCE, it's needed by the Hurd module.
+
+ * lib/header.h: Whitespace cleanups.
+
+ * lib/i386-io-windows.h: Fixed indentation and spelling.
+
+ * README.Windows: Mention that WinIO.dll is needed and where to get it.
+
+2006-06-27 Martin Mares <mj@ucw.cz>
+
+ * lib/Makefile, Makefile: Added rules for building and cleaning the example.
+
+ * lib/example.c: Show how to use PCI_FILL_CLASS and used another register
+ for demonstrating the config space accesses. The example is still very
+ incomplete, though.
+
+ * lib/pci.h, lib/generic.c: Added PCI_FILL_CLASS.
+
+ * lib/sysfs.c: Fetch device ID and class from sysfs, because they can
+ be modified by kernel fixups.
+
+ * lspci.c: Use PCI_FILL_CLASS. Thanks to Matthew Willcox for the patch.
+
+2006-06-03 Martin Mares <mj@ucw.cz>
+
+ * lspci.c: Introduced the -nn switch. Thanks to David N. Welton
+ for the idea.
+
+ * lib/names.c (pci_lookup_name): Introduced PCI_LOOKUP_MIXED and
+ cleaned up the name selection code.
+
+ * lib/names.c (pci_lookup_name): If PCI_LOOKUP_NO_NUMBERS is given,
+ pci_access->numeric_ids is ignored.
+
+ * lib/names.c (pci_lookup_name): When class is identified and
+ subclass isn't, display class name and subclass number.
+
+2006-05-31 Martin Mares <mj@ucw.cz>
+
+ * setpci.c (main): Don't crash when an invalid width is specified.
+
+ * lspci.c, setpci.c, common.c: die() now uses the right program
+ name when printing an error message.
+
+2006-05-19 Martin Mares <mj@ucw.cz>
+
+ * README: Mention the public GIT tree.
+
+ * README: Updated the list of ports.
+
+2006-05-10 Martin Mares <mj@ucw.cz>
+
+ * lib/configure, lib/fbsd-device.c: Added support for GNU/kFreeBSD.
+ Thanks to Petr Salinger for the patch.
+
+2006-05-05 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.3.
+
+ * lspci.c (scan_device): The -D switch didn't work as intended
+ in -m mode.
+
+2006-05-05 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.2.
+
+ * pci.ids: Updated to the current database snapshot.
+
+ * lib/obsd-device.c (and other files in lib/*): Added OpenBSD
+ interface by Matthieu Herrb <matthieu.herrb@laas.fr>, based on
+ the existing FreeBSD interface.
+
+ * Moved pciutils to a GIT repository, which now contains merged
+ history from both CVS and Arch. Good bye, TLA!
+
+2006-03-21 Martin Mares <mj@ucw.cz>
+
+ * lspci.c (show_slot_name): Avoid the previous changes in default
+ display of domain name when in machine-readable mode. However, `-D'
+ forces domain display even there.
+
+ * lspci.man: Added a warning on -m being the only format, which is
+ guaranteed to be stable between lspci versions.
+
+2006-03-13 Martin Mares <mj@ucw.cz>
+
+ * lspci.man: Documented the -D switch.
+
+ * lspci.c (show_slot_name, scan_device): If there are multiple PCI
+ domains or if the `-D' switch is used, show the domain number for
+ all devices.
+
+ * lspci.c (show_verbose): Report cache line size in bytes.
+
+ * update-pciids.sh: Use curl if available. Patch by Matthew Wilcox.
+
+2006-01-04 Martin Mares <mj@ucw.cz>
+
+ * lspci.c (show_express_link): Fixed decoding of link status register.
+ Thanks to Roland Dreier for his report.
+
+2005-12-12 Martin Mares <mj@ucw.cz>
+
+ * lib/nbsd-libpci.c: If libpci is used by an ordinary user, allow
+ at least read-only access. Patch by Pavel Celeda <celeda@liberouter.org>.
+
+ * Makefile: Corrected library path on NetBSD.
+
+2005-11-26 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.1.
+
+ * lspci.c (show_machine): Subsystem vendors were printed incorrectly
+ in machine-readable modes. Thanks to Pavel Celeda for a bug report.
+
+2005-10-11 Martin Mares <mj@ucw.cz>
+
+ * lspci.c (new_bus): Fixed a memory leak. Thanks to Paul Sangree for reporting it.
+
+2005-09-21 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.2.0.
+
+ * pci.ids: Updated copyright header.
+
+ * lib/sysfs.c (sysfs_get_resources): Removed warning about unsupported
+ 64-bit addresses, they are now always supported.
+
+ * lspci.c (show_bases): Corrected printing of 64-bit addresses
+ in bus-centric mode.
+
+ * lib/configure: Enable 64-bit addresses on all Linux systems.
+
+ * lib/types.h: Don't pad 64-bit addresses to 16 xigits, only to 8 if they
+ are shorter.
+
+2005-09-11 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.1.99-test11.
+
+ * lspci.c: Removed unnecessary trailing zeros in pci_lookup_name() calls.
+
+2005-09-10 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.1.99-test10.
+
+ * pci.ids: Synchronized with the current database snapshot.
+
+ * update-pciids.sh: Changed the default URL for downloading pci.ids
+ to http://pciids.sourceforge.net/v2.2/pci.ids.
+
+ * lib/pci.h (PCI_LIB_VERSION): Added version identifier.
+
+ * lib/names.c: Rewritten the name database module almost from scratch.
+ Everything is much cleaner and there are hopefully no more memory leaks;
+ pci_lookup_name() now uses varargs in a backward compatible fashion.
+ Introduced PCI_LOOKUP_NO_NUMBERS.
+
+ The new code supports subsystem entries not tied to a specific device.
+ I had to extend the format of pci.ids in order to support it, so I have
+ extended the idea of the "C" (class) blocks and introduced "S" blocks
+ for subsystems. To avoid doing more incompatible changes in the future,
+ the parser skips unknown single-letter blocks without reporting errors.
+
+2005-08-23 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.1.99-test9.
+
+ * Makefile (OPT): Removed -fomit-frame-pointer, it's default in recent
+ gcc versions when optimizing anyway.
+
+ * Made the path to pci.ids customizable and independent on the SHAREDIR.
+ Just override IDSDIR in the top-level Makefile.
+
+ * pci.ids: Synchronized with the current database snapshot.
+
+ * lspci.man, setpci.man: Improved the man pages. Documented all access
+ methods and which operations are privileged.
+
+ * lspci.c: Another rewrite of config register caching, now also including
+ lots of internal checks. It should be now perfectly able to cope with
+ portions of the configuration space being inaccessible, for example
+ due to insufficient access rights.
+
+ * lspci.c (show_pcix_nobridge, show_pcix_bridge): Cleaned up dumping of PCI-X
+ capabilities. Includes partial support for PCI-X 2.0 (probably incomplete
+ as I haven't seen the spec).
+
+ * lspci.c: Quell warnings about unused parameters.
+
+ * lspci.c: Removed C++ comments.
+
+ * lib/header.h: Merged definitions of extended capabilities and some new
+ PCI-X capability bits from linux-2.6.11/include/pci.h.
+
+ * lspci.c (show_caps): Try to scan extended capabilities only if the
+ device has either PCI-X or PCI Express capability in its normal capability
+ list.
+
+ * lib/dump.c (dump_init): Reading of dumps works again. The dump reader
+ now also remembers how much data it has read and refuses attempts to
+ access more than that.
+
+ * setpci.man, lspci.c, README: prefer spelling "buses" over "busses".
+
+ * lspci.c: If alloca() is not available, use xmalloc(), not malloc().
+
+ * lib/configure: Added x86_64 on Linux.
+
+2005-08-22 Martin Mares <mj@ucw.cz>
+
+ * lib/filter.c (pci_filter_parse_slot): Fixed parsing of domain filters.
+ Thanks to Matthew Wilcox for bug report.
+
+ * lspci.c: Corrected spelling of "Hz" (it was "hz" at several places).
+ Thanks to Krzysztof Oledzki for pointing that out.
+
+2004-08-19 Martin Mares <mj@ucw.cz>
+
+ * pciutils.spec (%attr): Setting attributes for /sbin/* is an anachronism,
+ because pciutils by default install to /usr/sbin. Update.
+
+2004-08-13 Martin Mares <mj@ucw.cz>
+
+ * Released as 2.1.99-test8.
+
+ * lib/types.h, lib/pci.h, lib/sysdep.h: `byte' and `word' are
+ no longer exported to the outside world.
+
+ * README.Windows: Updated.
+
+ * maint/release: Substitute version number in win32/config.h.
+
+ * win32/config.h: Added.
+
+ * lib/sysdep.h (bzero, strcasecmp): are macros on Windows.
+
+ * compat/getopt.[ch]: Added copies of getopt.[ch] from the GNU
+ C library for use on systems where the default libraries lack
+ proper getopt. Currently used by the Windows port.
+
+ * lib/sysdep.h, lib/types.h: Updates of the Windows port from
+ Alexander Stock.
+
+ * lib/types.h: If PCI_HAVE_Uxx_TYPES is defined, libpci doesn't define
+ its own u8, u16, u32, byte and word types and uses what the user
+ has supplied instead. With this change, all namespace clashes should
+ be gone.
+
+ * Makefile, lib/Makefile: Updated dependencies.
+
+ * pciutils.h: Include lib/sysdep.h and move NONRET et al. there.
+
+ * lib/sysdep.h: No need to include <sys/types.h> on FreeBSD, because
+ it is included unconditionally in lib/types.h.
+
+ * Moved system-dependent stuff from lib/internal.h to lib/sysdep.h,
+ which is also used by the utilities. Also moved type declarations
+ from lib/pci.h to lib/types.h.
+
+ * All files: Prepend a "PCI_" prefix in front of all configuration
+ defines to avoid namespace clashes. CAVEAT: If there is any libpci
+ application using the (undocumented) defines, it has to be updated.
+
+ * lib/Makefile: Killed HAVE_OWN_HEADER_H, since we stopped sharing
+ headers with the kernel a couple of years ago.
+
+ * lib/sysfs.c (sysfs_get_resources): We have 7 resources, not 6.
+
+ * lspci.c (show_rom): Use the same logic for printing disabled
+ or unassigned ROM's as we do in show_bases() for the other BAR's.
+
+ * lib/generic.c (pci_generic_fill_info): Better reaction to
+ invalid 64-bit addresses. Also d->hdrtype should not include
+ bit 7, which caused mysterious errors on multi-function devices.
+
+ * lib/generic.c (pci_generic_fill_info): Fill in base addresses
+ even if the regions are known to be disabled. It is the client's
+ job to interpret them. (And it is not a trivial job if you want
+ to do it correctly, since you need to check I/O and memory enables
+ on all upstream bridges, too.) However, it could be interesting to
+ introduce functions for interpreting the addresses and even for
+ mapping the regions and doing I/O on them.
+
+2004-07-30 Martin Mares <mj@ucw.cz>
+
+ * lspci.c: HyperTransport improvements from Maciej.
+
+2004-07-11 Martin Mares <mj@ucw.cz>
+
+ * lib/sysfs.c (sysfs_get_resources): Arguments now correspond
+ to the format string. [Patch by Bill Nottingham]
+
projects / pciutils.git / blobdiff