- 08 Dec, 2006 40 commits
-
-
Jiri Slaby authored
Release everything what was allocated and check return value of isa probing. Release only ISA boards in module exit, since pci have their own pci-probing-remove. Signed-off-by:
Jiri Slaby <jirislaby@gmail.com> Cc: Alan Cox <alan@lxorguk.ukuu.org.uk> Signed-off-by:
Andrew Morton <akpm@osdl.org> Signed-off-by:
Linus Torvalds <torvalds@osdl.org>
-
Jiri Slaby authored
Delete macros and functions, that are implemented in kernel yet (strtoul, min, tolower). Expand one function body in place, where it is called from. Signed-off-by:
-
Jiri Slaby authored
Some functions are now called from pci probing functiuon which is __devinit, not __init, correct this to not free functions after init if hotplug enabled. Signed-off-by:
-
Jiri Slaby authored
This driver expect everything to work. Implement fail paths logic to release regions, irq hangler, memory... if something is in bad state. Signed-off-by:
-
Jiri Slaby authored
Too many information is printed out (they may be easily obtained through sysfs), wipe them out in probe function. Convert rest of them to dev_ variants. Signed-off-by:
-
Jiri Slaby authored
Convert stallion driver to pci probing instead of pci_dev_get iteration. Signed-off-by:
-
Jiri Slaby authored
Check regions if they are free before we touch them. Release them in failpaths. Signed-off-by:
-
Jiri Slaby authored
Use pci_ friends for memory remapping of pci devices. Signed-off-by:
-
Jiri Slaby authored
If pci_register_driver fails, but eisa_driver_register doesn't, we don't call misc_deregister, but returns error. Return OK in such cases. Signed-off-by:
-
Jiri Slaby authored
Use kernel helpers for changing timer internals. Signed-off-by:
-
Jiri Slaby authored
Use Lindent to cleanup whitespace. Wrap long lines by hand. Signed-off-by:
-
Jiri Slaby authored
sx_remove_code contents were used twice. Call this function instead. Signed-off-by:
-
Jiri Slaby authored
Fix race condition which may occurs when multiple cards are probed at the same time. Add mutex to critical sections to avoid this situation. Signed-off-by:
-
Jiri Slaby authored
Most users won't need old probe code -- make it depndent on CONFIG_(E)ISA. Cc: <R.E.Wolff@BitWizard.nl> Signed-off-by:
-
Jiri Slaby authored
Instead of finding eisa directly, use eisa probing. Cc: <R.E.Wolff@BitWizard.nl> Signed-off-by:
-
Jiri Slaby authored
Mark as much as possible functions as __devinit to free them after driver initialization (if no hotplug). Cc: <R.E.Wolff@BitWizard.nl> Signed-off-by:
-
Jiri Slaby authored
Convert self-implemented kzalloc to kernel kcalloc. Cc: <R.E.Wolff@BitWizard.nl> Signed-off-by:
-
Jiri Slaby authored
Convert old pci code to pci probing. Cc: <R.E.Wolff@BitWizard.nl> Signed-off-by:
-
Jiri Slaby authored
Since tty->index is signed and may be < 0, we should assign this to int not uint. There is already a check to ensure if it is not negative, but gcc complains with -W flag enabled and it is perfectly correct: drivers/char/isicom.c:953: warning: comparison of unsigned expression < 0 is always false Fix this issue by converting `line' variable from uint to int. Signed-off-by:
-
Jiri Slaby authored
We don't need RCS_ID & co. Extract them from the code and define only one macro -- SX_VERSION. Signed-off-by:
-
Jiri Slaby authored
Don't init timer in such complicated way. Use DEFINE_TIMER and then only mod_timer to reset the expiration. Signed-off-by:
-
Jiri Slaby authored
Use wait_for_completion+complete instead of variables+msleep hack. Signed-off-by:
-
Jiri Slaby authored
Value returned from kamlloc may be NULL, we should check if ENOMEM occured. Signed-off-by:
Alan Cox <alan@redhat.com> Signed-off-by:
-
Jiri Slaby authored
Use pci_request_region in pci probing function instead of request_region. Signed-off-by:
-
Jiri Slaby authored
Instead of registering all devices in register_tty_driver, register devices in probe function and register only port_count devices. Signed-off-by:
-
Jiri Slaby authored
Memsetting of global static variables is not needed in the init code. Signed-off-by:
-
Jiri Slaby authored
ISA is not supported by this driver, remove parts, that take care of this. Signed-off-by:
-
Jiri Slaby authored
Rename init function from setup to init and mark it as __init, not __devinit. Signed-off-by:
-
Jiri Slaby authored
Simple functions (those, that calls reuest_irq, request_region et al) may be expanded directly in code. It makes probe function more readable and transparent. Signed-off-by:
-
Alan Cox authored
This turns on the split input/output speed features and arbitary baud rate handling for the x86-64 platform. Nothing should break if you use existing standard speeds. If you use the new speed stuff then you may see some drivers failing to report the speed changes properly in error cases. This will be worked on further. For the working cases this all seems happy. I'll post a test suite used to test the basic stuff as well. Patches for i386 will follow when I get a moment but are basically the same. If people could patch/test-suite other architectures and submit them that would be great. Signed-off-by:
-
Alan Cox authored
Having split out the user and kernel structures it turns out that some non glibc C libraries pull their termios struct from the kernel headers directly or indirectly. This means we must keep "struct termios" as the library sees it correct for the old ioctls. Not a big problem just shuffle the names and ifdef around a bit [akpm@osdl.org: build fix] Signed-off-by:
-
Alan Cox authored
This is the grungy swap all the occurrences in the right places patch that goes with the updates. At this point we have the same functionality as before (except that sgttyb() returns speeds not zero) and are ready to begin turning new stuff on providing nobody reports lots of bugs If you are a tty driver author converting an out of tree driver the only impact should be termios->ktermios name changes for the speed/property setting functions from your upper layers. If you are implementing your own TCGETS function before then your driver was broken already and its about to get a whole lot more painful for you so please fix it 8) Also fill in c_ispeed/ospeed on init for most devices, although the current code will do this for you anyway but I'd like eventually to lose that extra paranoia [akpm@osdl.org: bluetooth fix] [mp3@de.ibm.com: sclp fix] [mp3@de.ibm.com: warning fix for tty3270] [hugh@veritas.com: fix tty_ioctl powerpc build] [jdike@addtoit.com: uml: fix ->set_termios declaration] Signed-off-by:
Martin Peschke <mp3@de.ibm.com> Acked-by:
Peter Oberparleiter <oberpar@de.ibm.com> Cc: Cornelia Huck <cornelia.huck@de.ibm.com> Signed-off-by:
Hugh Dickins <hugh@veritas.com> Signed-off-by:
Jeff Dike <jdike@addtoit.com> Cc: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it> Signed-off-by:
-
Alan Cox authored
This is the core of the switch to the new framework. I've split it from the driver patches which are mostly search/replace and would encourage people to give this one a good hard stare. The references to BOTHER and ISHIFT are the termios values that must be defined by a platform once it wants to turn on "new style" ioctl support. The code patches here ensure that providing 1. The termios overlays the ktermios in memory 2. The only new kernel only fields are c_ispeed/c_ospeed (or none) the existing behaviour is retained. This is true for the patches at this point in time. Future patches will define BOTHER, ISHIFT and enable newer termios structures for each architecture, and once they are all done some of the ifdefs also vanish. [akpm@osdl.org: warning fix] [akpm@osdl.org: IRDA fix] Signed-off-by:
-
Alan Cox authored
In order to sort out our struct termios and add proper speed control we need to separate the kernel and user termios structures. Glibc is fine but the other libraries rely on the kernel exported struct termios and we need to extend this without breaking the ABI/API To do so we add a struct ktermios which is the kernel view of a termios structure and overlaps the struct termios with extra fields on the end for now. (That limitation will go away in later patches). Some platforms (eg alpha) planned ahead and thus use the same struct for both, others did not. This just adds the structures but does not use them, it seems a sensible splitting point for bisect if there are compile failures (not that I expect them) Signed-off-by:
-
Jiri Slaby authored
Many prototypes are useless, since functions are declared before they are called. Also some code was easy to move to resolve this dependency and delete prototypes. Signed-off-by:
-
Jiri Slaby authored
Use __init macro for functions, that may be freed after initialization. Signed-off-by:
-
Jiri Slaby authored
- Do not inline such long functions, it won't speed up anything. - Remove prototypes of these functions. Signed-off-by:
-
Jiri Slaby authored
- Move code from stl_init into module init function, because calling it was the only one thing, that it did. - Move this code to the end of the driver (usual place for this) to resolve dependencies simply -- without prototypes. Signed-off-by:
-
Jiri Slaby authored
Typedefs are considered ugly in the kernel. Eliminate them. Signed-off-by:
-
Jiri Slaby authored
casts of NULL are unnecessary. And so casts to (void *) are. Signed-off-by:
-
