X-Git-Url: http://git.tremily.us/?a=blobdiff_plain;f=TODO;h=7ced4364f7e6f14c80497c0d3b04e471b95fec59;hb=a6c04a29951077b153ee79075ddc7d7d423bd23e;hp=b1b5f87f11c489b827f542dad2cde05f641e379b;hpb=99cb4c902d851dd202bac35a3ee1e526dd2a16fc;p=comedi.git

diff --git a/TODO b/TODO
index b1b5f87f..7ced4364 100644
--- a/TODO
+++ b/TODO
@@ -1,42 +1,52 @@
 
-Doesn't get the right RTL flags
+make menuconfig
 
-das6402 needs cleanup
+bug: buffer resize to 0 convinces comedi that buffer doesn't exist
 
-check each driver init for stuff
+doc: how to use kcomedilib
 
-grep XXX
+comedi_rt_timer segfaults when the lower-level driver disappears
+and reappears.  This is a problem with the Comedi core, not the
+driver.  I don't really know how to fix this.  Actually, I do.
+kcomedilib needs an implementation of comedi_open() and comedi_close().
 
-register_symtab
+das6402 needs cleanup
 
 fix PCI code, use pci_find_device()
 
-standardize error codes
-
-verify PCI-6503 code
+pci_enable_device() needs to be added to many drivers
 
 get rid of spurious "char-major-98-0 not found"
 
-verify 2.0.x operation
+spinlocks
 
-develop 8253 and Am9513 code
+autoloading causes BUG on failure.
 
-increase mod use count for RT
+das1800: kmalloc could fail on large buffer sizes (see dt282x).  Needs
+to be fixed to use a bounce buffer.
 
-different error code for board error vs buffer overflow
+move /proc/comedi to /proc/driver/comedi
 
-change _recognize to const char *
+Drivers that do DMA need to be fixed for big-endian machines.
 
-implement poll/select
+ni_pcimio: doesn't do DMA with AO
 
-fix mmap
+ni_pcimio: ai COMEDI_WAKE_EOS doesn't work with DMA
 
-ni_mio_common: gpct code broken
+create separate fops structure for unconfigured devices (probably not)
 
-ni_mio_common: external trigger doesn't work
+ni_mio_common: resetting AO sets ao chan 1 to -10V
 
-spinlocks
+Dan Christian 07/01: rtd520.o: problems with buffer
+
+ni_pcimio: polling
+
+migrate COMEDI_CB interface to COMEDI_EV flags, move #defines to comedidev.h
+
+rename prealloc_buf and prealloc_bufsz
+
+fix RT interrupts and comedi_irq[] dumbness
 
-ni_mio_common: subd 5 fails test_insn with -EINVAL -- should support
-  readback, and/or should return -EIO (?)
+ni_mio_common: analog output doesn't work correctly with insns after
+  running a command