Hasso Tepper [Mon, 17 Sep 2007 10:31:07 +0000 (10:31 +0000)]
Add many devices to the uftdi(4). Sources of the info are mainly Linux and
Windows drivers.
Hasso Tepper [Mon, 17 Sep 2007 10:24:37 +0000 (10:24 +0000)]
uftdi(4) related usbdevs work:
- Shorten BBELECTRONICS -> BBELECTR.
- SIIG2 -> OCT. According to various sources (including Linux driver) SIIG
rebrands some Omnidirectional Control Technology products.
- Cleanup/sort the list of products using FTDI vendor ID.
- Add many new uftdi(4) related product ID's. Info mainly from Linux
driver and various Windows drivers.
Sascha Wildner [Sun, 16 Sep 2007 19:06:26 +0000 (19:06 +0000)]
Add a reference to wpa_supplicant(8) to the manual pages of all drivers
supporting it.
Sascha Wildner [Sun, 16 Sep 2007 18:55:16 +0000 (18:55 +0000)]
Add reference to SEE ALSO as well.
Sascha Wildner [Sun, 16 Sep 2007 18:52:17 +0000 (18:52 +0000)]
Add reference to bwi(4).
Sascha Wildner [Sun, 16 Sep 2007 18:26:18 +0000 (18:26 +0000)]
Add a manual page for the new bwi(4) driver (Broadcom BCM430x/4318
wireless cards).
In-discussion-with: sephe
Peter Avalos [Sun, 16 Sep 2007 17:02:49 +0000 (17:02 +0000)]
Make functions static and use M_ZERO instead of a separate bzero.
Sepherosa Ziehau [Sun, 16 Sep 2007 11:53:36 +0000 (11:53 +0000)]
Create per-interface sysctl tree.
Sepherosa Ziehau [Sun, 16 Sep 2007 11:31:20 +0000 (11:31 +0000)]
- Pass mgt packet's node into bwi_encap(), so fallback duration can be
calculate.
- Don't involve multicast packets into TX rate control.
Sepherosa Ziehau [Sun, 16 Sep 2007 10:20:15 +0000 (10:20 +0000)]
- Use RF/PHY information printing format to print BBP information.
- Print MAC information, instead of mixing it in the regwin information.
Sepherosa Ziehau [Sun, 16 Sep 2007 09:58:27 +0000 (09:58 +0000)]
For bwi(4) parts that use TX status registers:
- Update TXSTATUS0 register's layout according to v4 specification.
- Extract frame's TX count and whether the frame has been acknowledged.
- If the TX status indicates that the frame is pending or is AMPDU(?),
don't do any further processing (pending probably will not happen in
the current TX model).
Sepherosa Ziehau [Sun, 16 Sep 2007 08:25:41 +0000 (08:25 +0000)]
For bwi(4) parts that use TX status ring:
- Extract frame's TX count and whether the frame has been acknowledged.
- If the TX status indicates that the frame is pending, don't do any further
processing (this probably will not happen in the current TX model).
Sepherosa Ziehau [Sun, 16 Sep 2007 04:24:30 +0000 (04:24 +0000)]
- Add radio tap support
- Fix macro name; OFDM PLCP header uses 'rate' instead of 'signal'
- Fix comment
- Add error message, if interrupt setup fails
Peter Avalos [Sun, 16 Sep 2007 00:59:59 +0000 (00:59 +0000)]
Allow selection of the affected terminal using redirection
of standard input or standard output.
Obtained-from: FreeBSD
Peter Avalos [Sun, 16 Sep 2007 00:59:13 +0000 (00:59 +0000)]
Markup fixes.
Obtained-from: FreeBSD
Sascha Wildner [Sat, 15 Sep 2007 23:52:49 +0000 (23:52 +0000)]
if_ioctl takes a pointer to struct ucred.
While I'm here change the argument 'command' to 'cmd' for consistency.
Sascha Wildner [Sat, 15 Sep 2007 21:28:15 +0000 (21:28 +0000)]
Fix 'type qualifiers ignored on function return type' warnings.
Sascha Wildner [Sat, 15 Sep 2007 21:24:59 +0000 (21:24 +0000)]
Fix 'value computed is not used' warnings.
Sascha Wildner [Sat, 15 Sep 2007 20:06:39 +0000 (20:06 +0000)]
Fix some kprintf format warnings.
Sascha Wildner [Sat, 15 Sep 2007 18:21:12 +0000 (18:21 +0000)]
Add missing space.
Sascha Wildner [Sat, 15 Sep 2007 13:18:40 +0000 (13:18 +0000)]
Silence gcc 4.1's "packed attribute ignored..." warnings.
No differences in object files.
Taken-from: FreeBSD
Sepherosa Ziehau [Sat, 15 Sep 2007 10:53:31 +0000 (10:53 +0000)]
Make sure that the received frame contains at least an extra header,
a smallest frame (ACK or CTS) and FCS.
Sepherosa Ziehau [Sat, 15 Sep 2007 09:59:29 +0000 (09:59 +0000)]
Convert RSSI into signal strength (relative to noise floor)
Sascha Wildner [Sat, 15 Sep 2007 09:58:06 +0000 (09:58 +0000)]
Bring in some miscellaneous changes from OpenBSD and point to an
archive provided by sephe for the firmware files.
Sepherosa Ziehau [Sat, 15 Sep 2007 07:19:23 +0000 (07:19 +0000)]
Shut up GCC41 warning: different pointer signedness
Sepherosa Ziehau [Sat, 15 Sep 2007 07:12:03 +0000 (07:12 +0000)]
Bug found by GCC41 warning:
Wrong __packed position, we want to pack the anonymous structure not the field.
Sascha Wildner [Fri, 14 Sep 2007 23:47:53 +0000 (23:47 +0000)]
Fix typos, mostly effect -> affect.
Sascha Wildner [Fri, 14 Sep 2007 23:22:38 +0000 (23:22 +0000)]
Fix typo.
Sascha Wildner [Fri, 14 Sep 2007 22:53:56 +0000 (22:53 +0000)]
Make the firmware_image_load() and firmware_image_unload() functions
known to the world by adding a minimal firmware(9) manual page.
Sascha Wildner [Fri, 14 Sep 2007 21:04:56 +0000 (21:04 +0000)]
Fix typo.
Sascha Wildner [Fri, 14 Sep 2007 20:02:03 +0000 (20:02 +0000)]
Remove two unused entries.
Sascha Wildner [Fri, 14 Sep 2007 19:47:03 +0000 (19:47 +0000)]
Add missing LIBRARY sections and extend mdoc.local accordingly.
Sascha Wildner [Fri, 14 Sep 2007 19:08:59 +0000 (19:08 +0000)]
Remove several mann/catn directories in /usr/share.
It seems they were added for Tcl originally in FreeBSD but nothing gets
installed there anymore for a long time.
In-discussion-with: hasso
Sascha Wildner [Fri, 14 Sep 2007 06:41:07 +0000 (06:41 +0000)]
Fix path.
Spotted-by: Noah Yan <noah.yan@gmail.com>
Chris Turner [Thu, 13 Sep 2007 23:40:24 +0000 (23:40 +0000)]
testing, 123
check, check, check.
sibilance, sibilance.
Sascha Wildner [Thu, 13 Sep 2007 20:02:52 +0000 (20:02 +0000)]
Add a skeletal kernconf(5) manual page for documenting the structure
of a kernel configuration file.
It only documents the keywords for now but can (and will eventually)
be extended to document our numerous kernel options.
It's kinda based on NetBSD's options(4).
Suggested-by: dillon (sort of)
[http://leaf.dragonflybsd.org/mailarchive/bugs/2007-08/msg00064.html]
Sascha Wildner [Thu, 13 Sep 2007 13:55:35 +0000 (13:55 +0000)]
Remove NO_WERROR.
Sascha Wildner [Thu, 13 Sep 2007 10:55:55 +0000 (10:55 +0000)]
Fix typo.
Sascha Wildner [Thu, 13 Sep 2007 10:34:04 +0000 (10:34 +0000)]
Escape the ' chars in the example.
Sascha Wildner [Thu, 13 Sep 2007 10:33:10 +0000 (10:33 +0000)]
Add missing .El
Yonghong Yan [Thu, 13 Sep 2007 04:23:59 +0000 (04:23 +0000)]
a helloWorld ci test by yanyh
Sepherosa Ziehau [Wed, 12 Sep 2007 12:02:09 +0000 (12:02 +0000)]
Prepare to support polling(4) on multiple cpus:
- Add per-cpu polling(4) context, move global variables into it.
- Make global sysctl nodes per-cpu. Fake original global sysctl
nodes by creating alias to the lowest cpu's sysctl nodes, mainly
for compatibility.
Hasso Tepper [Wed, 12 Sep 2007 09:26:32 +0000 (09:26 +0000)]
Oops, remove comma.
Reminded-by: swildner@
Hasso Tepper [Wed, 12 Sep 2007 08:42:59 +0000 (08:42 +0000)]
Update the agp(4) manpage. General description and example based on NetBSD
agp(4) manpage, ioctl(2) interface description from FreeBSD.
Obtained from: NetBSD and FreeBSD with modifications
Hasso Tepper [Wed, 12 Sep 2007 08:31:44 +0000 (08:31 +0000)]
Update the agp(4) code to the latest one from FreeBSD HEAD. This brings in
a lot of newer hardware support.
Obtained-from: FreeBSD
Hasso Tepper [Wed, 12 Sep 2007 07:59:31 +0000 (07:59 +0000)]
Add bus_alloc_resources() and bus_release_resources() functions to allow to
simplify the code and to make it easier to port drivers (initially the
agp(4)) from FreeBSD.
Obtained-from: FreeBSD
Matthew Dillon [Mon, 10 Sep 2007 15:13:55 +0000 (15:13 +0000)]
Fix bugs in the handling of CIDR specifications such as 'route add 128/8
<gateway>'. Also rectify the way netmasks are deduced from the address class.
Submitted-by: "Nuno Antunes" <nuno.antunes@gmail.com>
Matthew Dillon [Mon, 10 Sep 2007 15:08:43 +0000 (15:08 +0000)]
kern_access() had the same bug kern_stat() had with regards to a
vnode/namecache deadlock when dealing with stale NFS mounts.
Reported-by: elekktretterr@exemail.com.au
Sepherosa Ziehau [Mon, 10 Sep 2007 14:08:28 +0000 (14:08 +0000)]
- Add interrupt moderation support by only using TIMER intr; IM is not
possible when RX or TX intr is enabled. With this change, my nfe(4)
can TX at line speed finally (mtu 1500).
- Add per iface sysctl node hw.nfeX.imtime to enable, disable and set
interrupt moderation time. Set it to -1 will disable IM; 64-128
seems to be a reasonable value range. This sysctl variable is _not_
on by default, and can be tuned when the iface is running.
- Nuke global sysctl nodes, make them per iface.
# We may consider using TIMER intr under all situations, but with larger
# imtime when IM is not enabled, since for certain type of nfe(4), TX
# interrupts may never come.
Simon Schubert [Sun, 9 Sep 2007 23:32:43 +0000 (23:32 +0000)]
Resync patch with contrib.
Sascha Wildner [Sun, 9 Sep 2007 15:43:38 +0000 (15:43 +0000)]
Switch even more to gcc41.
Sepherosa Ziehau [Sun, 9 Sep 2007 09:14:38 +0000 (09:14 +0000)]
- Make fwe(4) aware of IFF_POLLING setting in its if_init().
- Allow IFF_POLLING flag to be turned on even when the interface is not
up yet, since after above fix, all drivers that support polling(4) will
turn on/off interrupt in their if_init() based on IFF_POLLING.
This fixes the bug reported by many people that enabling "polling" in
rc.conf's ifconfig_ifaceX does not work.
Sepherosa Ziehau [Sun, 9 Sep 2007 06:21:23 +0000 (06:21 +0000)]
Correct DEVICE_POLLING support in vr_init()
Sepherosa Ziehau [Sun, 9 Sep 2007 05:11:28 +0000 (05:11 +0000)]
NULL should never be passed to ether_poll_deregister(). Try catching any
brain dead caller by assertion.
Sepherosa Ziehau [Sun, 9 Sep 2007 04:30:25 +0000 (04:30 +0000)]
- Fix comment.
- Use ifp as local variable, indirect accessing ifp through pollrec is not
meaningful here.
- Break long lines.
Sepherosa Ziehau [Sun, 9 Sep 2007 03:51:25 +0000 (03:51 +0000)]
Nuke the hack in vr(4) that polling(4) will be turned when watchdog times out;
if interrupt does not work at all, polling(4) can be turned manually.
Approved-by: dillon@
Sascha Wildner [Sat, 8 Sep 2007 20:50:50 +0000 (20:50 +0000)]
Use .Dv and start sentences on a new line.
Sepherosa Ziehau [Sat, 8 Sep 2007 12:50:34 +0000 (12:50 +0000)]
It does not make sense to call interface's initialization routine when
the interface is up and only layer 3 address is to be changed.
Submitted-by: "Devon H. O'Dell" <devon.odell@coyotepoint.com> (w/ mod)
Sepherosa Ziehau [Sat, 8 Sep 2007 12:35:04 +0000 (12:35 +0000)]
Kill token ring remainder.
Sepherosa Ziehau [Sat, 8 Sep 2007 11:20:22 +0000 (11:20 +0000)]
- Add missing header protector
- Strip blank line at the end of file
Sascha Wildner [Sat, 8 Sep 2007 10:49:00 +0000 (10:49 +0000)]
Remove previous hack to cope with CVS IDs.
It wasn't properly documented and also prevented quiz questions starting
with # from being asked. Not very likely but we shouldn't break program
semantics because of administrative reasons.
The datfiles have now been added to the exclusion list.
Sascha Wildner [Sat, 8 Sep 2007 09:48:59 +0000 (09:48 +0000)]
Synchronize manual page installation of gcc34 and gcc41 and switch the
various MLINKS to gcc41 which is default now.
Sepherosa Ziehau [Sat, 8 Sep 2007 06:16:27 +0000 (06:16 +0000)]
Hook bwi(4)
Sepherosa Ziehau [Sat, 8 Sep 2007 06:15:54 +0000 (06:15 +0000)]
Add WIP support for Broadcom wireless chips:
- STA and MONITOR operation mode are supported
- WPA/WPA2 is supported (host based)
- Use V3 firmware
- 11b/g mode is supported
TODOs (AKA defects ;)
- Add the missing part of PCIe initialization (need testers)
- Hook TX rate control support in our 802.11 layer
- HOSTAP and IBSS operation mode support
- Add radio tap support
- Use V4 firmware for newer chips
- Use firmware based crypto for V4 firmware
- V3 and V4 firmware coexist
- Make LED blink
- Make my 11b only card work :]
- Correctly configure CW and MAC level acknowledge rates based on the
characteristic of the BSS/IBSS that we are going to join
And blah, blah, blah ...
Known to work devices, i.e. tested devices:
o BCM4318 (Linksys WPC54G ver.3, cardbus)
BBP: id 0x4318, rev 0x2, pkg 2
MAC: rev 9
PHY: type 2, rev 7, ver 3
RF: BCM2050, rev 8
o BCM4306 (Buffalo WLI2-PCI-G54S, PCI)
BBP: id 0x4306, rev 0x3, pkg 0
MAC: rev 5
PHY: type 2, rev 2, ver 2
RF: BCM2050, rev 2
o BCM4319 (unknown vendor, mini-PCI)
BBP: id 0x4318, rev 0x2, pkg 2
MAC: rev 9
PHY: type 2, rev 7, ver 3
RF: BCM2050, rev 8
o BCM4309 (unknown vendor, mini-PCI)
BBP: id 0x4306, rev 0x3, pkg 1
MAC: rev 5
PHY: type 2, rev 2, ver 2
RF: BCM2050, rev 2
TX performance of above devices at different OFDM rates (unit Mbits/s):
OFDM rate 54 48 36 24 18 12 9 6
TX perf 23.x 21.x 18.x 14.x 11.x 8.x 4.x(1) 6.x(2)
(1) BCM4319/BCM4306 listed above can only do 2.xMbits/s at 9Mbits/s
(2) BCM4319 listed above has unstable TX performance, 4.x-6.xMbits/s,
at 6Mbits/s
(3) There is a bug(to be found) in driver which reverses 9Mbits/s and 6Mbits/s
Thank folks at bcm-specs.sipsolutions.net for their reverse engineering work.
Thank johill and mb on #bcm-specs IRC channel for answering my questions.
#
# Following devices are also known to work, but have poor TX performance.
# They have same characteristic that MAC rev is less than 5.
# o BCM4306 (Dell TrueMobile 1300 rev A01, cardbus)
# BBP: id 0x4306, rev 0x2, pkg 0
# MAC: rev 4
# PHY: type 2, rev 1, ver 1
# RF: BCM2050, rev 2
# o BCM4306 (unknown vendor, mini-PCI)
# BBP: id 0x4306, rev 0x2, pkg 0
# MAC: rev 4
# PHY: type 2, rev 1, ver 1
# RF: BCM2050, rev 2
#
# Certain 11b only cards may work depending on the PHY rev
#
Peter Avalos [Sat, 8 Sep 2007 05:01:18 +0000 (05:01 +0000)]
Upgrade to OpenSSH 4.7p1.
This brings some performance enhancements and bug fixes.
Peter Avalos [Sat, 8 Sep 2007 04:52:53 +0000 (04:52 +0000)]
Merge from vendor branch OPENSSH:
Import OpenSSH 4.7p1.
Peter Avalos [Sat, 8 Sep 2007 04:52:53 +0000 (04:52 +0000)]
Import OpenSSH 4.7p1.
Simon Schubert [Sat, 8 Sep 2007 01:29:40 +0000 (01:29 +0000)]
Yugoslavia split up into Serbia and Montenegro some time ago.
Simon Schubert [Sat, 8 Sep 2007 01:29:14 +0000 (01:29 +0000)]
Allow comment lines in input files.
Sascha Wildner [Fri, 7 Sep 2007 23:24:20 +0000 (23:24 +0000)]
The weekly script runs Saturday morning.
Found-in: FreeBSD
Matthew Dillon [Fri, 7 Sep 2007 21:42:59 +0000 (21:42 +0000)]
General userfs fleshing out work. Abstract out construction and
deconstruction functions and take advantage of the new getpages/putpages
support.
Matthew Dillon [Fri, 7 Sep 2007 19:31:30 +0000 (19:31 +0000)]
Remove a bogus assertion. in_vm86call may have been set by some unrelated
thread. We have to check that we are in an emulated 8086 mode before we
can assert that the MP lock is held.
Reported-by: Rumko <rumcic@gmail.com>
Sascha Wildner [Fri, 7 Sep 2007 08:14:57 +0000 (08:14 +0000)]
Oops, forgot one .Dv
Sascha Wildner [Fri, 7 Sep 2007 08:01:29 +0000 (08:01 +0000)]
* Start sentences on a new line.
* Fix typo: depreciated -> deprecated
* Use .Dv for defined values.
Peter Avalos [Fri, 7 Sep 2007 02:43:49 +0000 (02:43 +0000)]
Fix typo.
Reported-by: "Trevor Kendall" <trevorjkendall@gmail.com>
Peter Avalos [Fri, 7 Sep 2007 00:09:02 +0000 (00:09 +0000)]
Use >= instead of > to explicitly see which WARNS level we're dealing with.
Sascha Wildner [Thu, 6 Sep 2007 21:03:55 +0000 (21:03 +0000)]
Use bcd2bin().
Sascha Wildner [Thu, 6 Sep 2007 18:17:24 +0000 (18:17 +0000)]
Remove redundant initialization.
Sascha Wildner [Thu, 6 Sep 2007 11:29:30 +0000 (11:29 +0000)]
Use .Dv
Sascha Wildner [Thu, 6 Sep 2007 11:11:19 +0000 (11:11 +0000)]
Document some ATM media settings.
Peter Avalos [Wed, 5 Sep 2007 22:11:01 +0000 (22:11 +0000)]
Describe CCVER's default as gcc41.
Peter Avalos [Wed, 5 Sep 2007 21:40:10 +0000 (21:40 +0000)]
Add a build option (NO_GCC34) to prevent building gcc version 3.4.
This does not attempt to remove gcc-3.4, only to prevent it from building.
Hasso Tepper [Wed, 5 Sep 2007 08:23:43 +0000 (08:23 +0000)]
Better setpgid(2) documentation.
Obtained-from: NetBSD with modifications
Matthew Dillon [Wed, 5 Sep 2007 05:28:32 +0000 (05:28 +0000)]
Set si_iosize_max to silence run-time warnings.
Reported-by: Vincent Stemen <vs1@crel.us>
Matthew Dillon [Tue, 4 Sep 2007 22:07:54 +0000 (22:07 +0000)]
Change the virtual kernel's default hz to 20, because the kqueue timers we
use can't really go much higher. This fixes sleep intervals.
Reported-by: Multiple People
Simon Schubert [Tue, 4 Sep 2007 17:08:38 +0000 (17:08 +0000)]
Use gnuregex.h instead of a private header whose declarations are wrong.
This private header declares the return type of re_search to be off_t,
i.e. 64 bit. Our (old) gnu regex library however only returns ints. As a
result, the generated opcode erroneously only looked at the sign of %edx,
which happened to contain a random value.
DragonFly-bug: <http://bugs.dragonflybsd.org/issue801>
Reported-by: Vincent Stemen <vs1@crel.us>
Simon Schubert [Tue, 4 Sep 2007 16:53:57 +0000 (16:53 +0000)]
Consolidate all local CVS patches into the contrib sources.
Include file prefix ordering is a horror when patching header files,
so get over this once and for all and put the patches where they belong.
Matthew Dillon [Tue, 4 Sep 2007 00:48:18 +0000 (00:48 +0000)]
Force an over-the-wire transaction when resolving the root of an NFS mount
point. The namecache will cache the mount point for us so this should not
reduce performance. We need to know that the mount point is good or
things like the namecache resolver could end up looping forever trying to
resolve a stale NFS mount.
Reported-by: elekktretterr@exemail.com.au (Petr)
Peter Avalos [Mon, 3 Sep 2007 19:24:53 +0000 (19:24 +0000)]
Remove options that don't make sense for a vkernel, and make SMP default.
Peter Avalos [Mon, 3 Sep 2007 18:48:12 +0000 (18:48 +0000)]
Update for libarchive 2.2.7.
Peter Avalos [Mon, 3 Sep 2007 18:47:14 +0000 (18:47 +0000)]
Merge from vendor branch LIBARCHIVE:
Import libarchive-2.2.7.
This brings support for mtree files, gtar 1.17 sparse files, and
refines suid/sgid restore handling.
Peter Avalos [Mon, 3 Sep 2007 18:47:14 +0000 (18:47 +0000)]
Import libarchive-2.2.7.
This brings support for mtree files, gtar 1.17 sparse files, and
refines suid/sgid restore handling.
Matthew Dillon [Mon, 3 Sep 2007 17:32:32 +0000 (17:32 +0000)]
Bring in FreeBSD/1.177 - fix a bug in a call made to useracc(). This
permissions check is redundant since copyin/copyout also perform a
validation so there is no known security issue.
Reported-by: "Constantine A. Murenin" <cnst@FreeBSD.org>
Obtained-from: FreeBSD
Matthew Dillon [Mon, 3 Sep 2007 17:06:24 +0000 (17:06 +0000)]
Add a MNTK_ flag to the mount structure allowing a VFS to specify that
no submounts under the VFS are to be allowed. Adjust procfs and linprocfs
to use the feature.
Submitted-by: "Nicolas Thery" <nthery@gmail.com>
Sascha Wildner [Sun, 2 Sep 2007 19:30:48 +0000 (19:30 +0000)]
Use .Pa
Sepherosa Ziehau [Sun, 2 Sep 2007 13:27:23 +0000 (13:27 +0000)]
Switch ipfw from ipfw1 to ipfw2.
Approved-by: dillon@
Submitted-by: Gary Allan <dragonfly@gallan.plus.com> (w/ modification)
Peter Avalos [Sat, 1 Sep 2007 20:43:24 +0000 (20:43 +0000)]
TAILQ_FOREACH to enhance readability.
Peter Avalos [Sat, 1 Sep 2007 18:42:08 +0000 (18:42 +0000)]
Sync with OpenBSD.
- Save and restore not only scale, but ibase as well.
- Interchange roles: bc becomes parent, dc child. This leads to saner
tty handling.
- Avoid segv when a syntax error occurs while reading from command line.
- Various whitespace, warns cleanup, etc.
Hasso Tepper [Fri, 31 Aug 2007 13:39:35 +0000 (13:39 +0000)]
Handle baudrate requests algorithmically with newer chips (not old SIO),
allowing all sorts of strange rates one might need.
Obtained-from: OpenBSD
Hasso Tepper [Fri, 31 Aug 2007 13:25:32 +0000 (13:25 +0000)]
Simplify the way how chip type is determined. Instead of managing insane
case constructs with info what vendor/product ids combinations have which
chip, use the info device descriptors provide.
- FT2232 dual UART chips have bNumInterfaces set to 2 (obviously).
- Original SIO have bcdDevice set to 0x0100.