summaryrefslogtreecommitdiff
path: root/xhfc
AgeCommit message (Collapse)Author
2010-12-03remove redundant activate_requestNoe Rubinstein
2010-12-03Fix spanno/port/port+1 confusion in tracesNoe Rubinstein
2010-12-02if it works, do break itNoe Rubinstein
2010-11-26remove nonsenseNoe Rubinstein
2010-11-25commentsNoe Rubinstein
2010-11-25traces and todoNoe Rubinstein
2010-11-25remove redundant codeNoe Rubinstein
2010-11-25State machine changesNoe Rubinstein
Forgot to commit last friday
2010-11-19update notesNoe Rubinstein
2010-11-18Is the state machine guaranteed to work?Noe Rubinstein
2010-11-12compiles; panics.Noe Rubinstein
2010-11-12debug traces, cleanup/fixNoe Rubinstein
2010-11-05moreNoe Rubinstein
2010-11-05state handling code nearly completeNoe Rubinstein
2010-11-04expiry_T1, expiry_T3, set_st_stateNoe Rubinstein
2010-11-04start rewriting software part of ST state machineNoe Rubinstein
2010-10-29change values for timersNoe Rubinstein
note: TIMER_1_MS does not mean "1 millisecond timer", it means "timer 1, in milliseconds".
2010-10-29todo/commentsNoe Rubinstein
2010-10-22remove codeNoe Rubinstein
2010-10-22remove xhfc_span_startup and shutdownNoe Rubinstein
2010-10-22fix some errors in the previous commitsNoe Rubinstein
2010-10-22 #ifdef AUDIONoe Rubinstein
#endif
2010-10-22s/pci_register_driver/dahdi_pci_init/Noe Rubinstein
Also cosmetic changes
2010-10-21xhfc_span.shutdown unusedNoe Rubinstein
2010-10-21xhfc_span->ts not relevant anymoreNoe Rubinstein
2010-10-21spantype description in spanconfigNoe Rubinstein
I forgot to change xhfc_span->span.spantype in xhfc_spanconfig
2010-10-21change V_REV to 1 for B fifos. del comments.Noe Rubinstein
2010-10-19added a new (very very low quality way) to propagate audioGuillaume Knispel
2010-10-17add some audio for XHFCGuillaume Knispel
2010-10-14use __func__ instead of __FUNCTION__Guillaume Knispel
2010-10-14don't share the interrupt of the xhfcGuillaume Knispel
2010-10-14properly free_irq where needed in error handling code pathGuillaume Knispel
2010-10-14rework of some return error codesGuillaume Knispel
2010-10-11commentsGuillaume Knispel
2010-10-11relax the cpu :)Guillaume Knispel
2010-10-11commentsGuillaume Knispel
2010-10-11add parentheses around macro arguments in definitionsGuillaume Knispel
2010-10-11build with GPIOGuillaume Knispel
2010-10-11don't configure half of the PCM in enable_interruptsGuillaume Knispel
2010-10-11remove unused r_su_sel1 varGuillaume Knispel
2010-10-11disable the F1_1 signal for nowGuillaume Knispel
2010-10-11switch XHFC PCM back to 2 MBit/sGuillaume Knispel
2010-10-11update GPIO lines used for XHFC /INT and /RESGuillaume Knispel
2010-10-08remove sync_source module param, which was not producing any effect on the chipGuillaume Knispel
2010-10-08Don't exec isr fifo and state payloads during initGuillaume Knispel
In case of HW glitch, shared interrupt, or spurious interrupt there was no self-waranty that the complicated routines talking to dahdi and to the chip (hfc_handle_state, hfc_update_st_timers) would not be called, which could maybe have resulted in strange results. This change extend the use of xhfc->running so that it can be used in the ISR xhfc_interrupt() as a flag to protect those calls. It is now carefully set in enable_interrupts() using set_mb() (and also in disable_interrupts(), which is only used in init when it does not change the initial value 0 of xhfc->running, and in cleanup just before free_irq() -- note that the disabling does not have synchronous caracteristics). Also, xhfc->span is now also set with set_mb() for good measure. Unrelated changes (and I slap myself for doing that!): * noop xhfc_ioctl() has been simplified * comments added / changed
2010-10-07indentationGuillaume Knispel
2010-10-07Merge branch 'master' of hard:git/driversNoe Rubinstein
2010-10-07commit ba506d3 improperly removed a '+': fix thatGuillaume Knispel
2010-10-07add comment about xhfc timerGuillaume Knispel
2010-10-07remove LSD-induced control structuresNoe Rubinstein
generated by cgit v1.2.3 (git 2.39.1) at 2024-06-02 11:45:40 +0000