summaryrefslogtreecommitdiff
path: root/sys/dev/usb
AgeCommit message (Collapse)Author
2009-07-10timeout_add -> timeout_add_msecBret Lambert
ok damien@
2009-07-10timeout_add -> timeout_add_msecBret Lambert
ok damien@
2009-07-02Revert the 'optimization' I attempted in r1.21 since it turns outKenneth R Westerback
that returning COMPLETE for non-SCSI_POLL commands is significant. Fixes double free of xs and subsequent crashes caused by pulling out USB media while i/o is active. Problem analyzed & fix discovered by hshoexer@. A more general fix may follow post-4.6. ok miod@ deraadt@.
2009-06-28Better wording for the not-supported-video-control message.Marcus Glocker
Requested by and OK deraadt@
2009-06-24print RFv1/RFv2 for 8187L dmesgs; since code paths are very differentMartynas Venckus
also print 8187B/8187L rev. ids for unknown revisions and remove dots as suggested by jsg@. ok jsg@
2009-06-21simplify write_s16 routine and just use memory stack instead of twoMartynas Venckus
buffers. ok jsg@. tested by kevlo@ and myself
2009-06-17fix endianness issue on BE.Damien Bergamini
2009-06-06regenYojiro Uo
2009-06-06add another UDL deviceYojiro Uo
- LCD-4300U (century corp.) ok mglocker@
2009-06-06- implement support for rtl8187bMartynas Venckus
- change state properly in urtw_stop. disable interrupts, rx, and tx. fixes scanning, ifconfig down, and some other issues - fix rxgain; for ver. 2 - remove delays; that have been removed 8185 -> 8187 new one tested by jsg@, a bunch of people on tech@, and myself old one tested by nick@
2009-06-05catch up NetBSD kern/41048.Yojiro Uo
check for a non-NULL configuration descriptor before dereferencing. ok deraadt@
2009-06-05this one tooMartynas Venckus
2009-06-05reduce diff i will send in a minuteMartynas Venckus
2009-06-04- factor out rf into its own structureMartynas Venckus
- clean up the softc a bit - remove unused SET_CHANNEL - init set sens
2009-06-04- 8187 uses different anaparam valuesMartynas Venckus
- use function pointer we've saved; that'll make things easier for 8187b.
2009-06-04detect & print the revision (rev. d, b) of rtl8187l chip using theMartynas Venckus
tx_conf register. if you get "not supported" in your dmesg; it's actually an (early) rtl8187b that's using rtl8187l id; so you should test the diff on tech@. otherwise; rtl8187b devices not matched yet.
2009-06-04- prepare for multiple revisions; urtw_lookup stuffMartynas Venckus
- fix some comments - remove var from a declaration; for consistency
2009-06-04- make comments understandable by lintMartynas Venckus
- reduce diff a bit
2009-06-04difs&eifs are 8187 regsMartynas Venckus
2009-06-04- KNFMartynas Venckus
- nitemize
2009-06-04- update io functions to support indexMartynas Venckus
- add index macros
2009-06-04- fix some commentsMartynas Venckus
- remove some unused fields from softc - spacing
2009-06-04spacingMartynas Venckus
2009-06-04instead of having some 'generic' data str. with "NB: don't use itMartynas Venckus
for tx" for some fields, and functions trying to do everything; split it into rx_data & tx_data. this way it's less error-prone. i've actually fixed a bug doing this. "please start comitting this stuff" jsg@
2009-06-04trying to split some parts;Martynas Venckus
- tx_start never actually free'd mbufs; oops report/test/ok jsg@
2009-06-04regenDamien Bergamini
2009-06-04the Sitecom WL-608 is a run(4) (although it is not 80.11n capable).Damien Bergamini
it has an RT2070 MAC/BB which is a (cheap) single-chip 802.11g solution. USB ID provided by Michael Kiesel.
2009-06-03syncMartynas Venckus
2009-06-03realtek rtl8187b devices that will run urtw. ok jsg@Martynas Venckus
2009-06-02ehci_alloc_sqtd_chain() was doing signed math on len, thus the great codeTheo de Raadt
curlen -= curlen % mps; can for very large transfers result in an intermediate variable growing larger than it should. Probably can't really happen in the real world. Do the same u_int repairs to the other matching drivers ok blambert kjell miod kettenis
2009-06-02Avoid a buf oflow in uhidev_set_report_async() if we have to prepend theMiod Vallat
report id and the data to report is too long (this should probably use dynamic allocation for large reports).
2009-06-02Fix some minor format string problems found in a maze of false positivesTheo de Raadt
provided by Parfait ok oga
2009-06-01Use the background color attribute in udl_erasecols and udl_eraserowsMarcus Glocker
instead hardcoding ``0x0000''.
2009-06-01Sort argument list of udl_draw_char similar as udl_fb_* friends.Marcus Glocker
2009-06-01Allow a font width other than 8 pixels.Marcus Glocker
2009-06-01Better naming for some framebuffer functions.Marcus Glocker
2009-06-01Reduce data payload for font drawing to a fixed size per character.Marcus Glocker
2009-06-01I forgot to apply the End Of Command token for async xfers. Maybe thatMarcus Glocker
caused the stalls seen in some rare conditions? Who knows ...
2009-06-01oops, bad merge.Damien Bergamini
2009-06-01- cleanup tx pathDamien Bergamini
- nitemize
2009-06-01fix endianness issue.Damien Bergamini
tested by jsg@ on tadpole sparc64 laptop.
2009-05-31We are able to address 256 pixels (16bit) per command not just 128 pixels.Marcus Glocker
This safes us some protocol overhead, especially when bit blittering large regions.
2009-05-31Add AL2230S support, extracted from part of a massiveJonathan Gray
'sync to p4' style commit in FreeBSD.
2009-05-31PV2000 -> AL2230SJonathan Gray
2009-05-31Match on siemens zd1211b, support for the al2230s radioJonathan Gray
is not yet in, but this makes things easier.
2009-05-31regenJonathan Gray
2009-05-31Siemens ZD1211B.Jonathan Gray
2009-05-24Mention broken cursor drawing.Marcus Glocker
2009-05-24regenMarcus Glocker
2009-05-24List product name instead chip type.Marcus Glocker
generated by cgit v1.2.3 (git 2.25.1) at 2025-01-12 06:54:45 +0000