- 30 Dec, 2009 2 commits
-
-
Kevin Wolf authored
* Up to now, the cdi-device struct was created during the init_driver call of the driver and init_device only completed the initialisation. The new model is that init_device gets bus information (e.g. PCI address, device/vendor ID) and decides if it can handle the device. If so, it creates a cdi_device object and returns it. This way device creation is decoupled from driver initialisation and hot-plugging becomes possible. Signed-off-by:
Kevin Wolf <kevin@tyndur.org> Acked-by:
Max Reitz <max@tyndur.org> Acked-by:
Matthew Iselin <matthew@theiselins.net>
-
Kevin Wolf authored
+ Each device gets a structure that links it to its bus. For now this means moving the cdi_pci_device out of the specific device struct of each driver into the generic cdi_device. Signed-off-by:
-
- 28 Dec, 2009 1 commit
-
-
Kevin Wolf authored
! iso9660: If we start reading in the middle of a cache block, we must not try to copy a whole block, but stop at the end of the cache block. Signed-off-by:
-
- 22 Dec, 2009 4 commits
-
-
Kevin Wolf authored
- cdi_device.type is redundant, we already have a pointer to the driver which stores the type. Remove it. Signed-off-by: -
+ Added mempool header written by Kevin as a temporary solution for the USB host controller drivers coming soon Signed-off-by:
-
Kevin Wolf authored
! cdi-osdep.h: Matt fixed the English documentation for CDI_DRIVER and I forgot to synchronise the German one in the same commit. This patch fixes it. Signed-off-by: -
Kevin Wolf authored
! iso9660: If an error occurs during a read operation, the CDI cache returns a NULL pointer. We better check for that. Signed-off-by:
-
- 21 Dec, 2009 2 commits
-
-
Kevin Wolf authored
! ata: The error register is only valud if ERR is set in the status register Signed-off-by: -
+ Some drivers need direct access to the PCI configuration space Signed-off-by:
-
- 20 Dec, 2009 1 commit
-
-
! e1000: This patch merely fixes a signed/unsigned comparison warning in the e1000 driver that was missed in the previous warning fix commit. Signed-off-by:
-
- 18 Dec, 2009 1 commit
-
-
* Use C99 (better: gnu99) instead of implicit C89 * Ignore errors upon make clean Signed-off-by:
-
- 11 Dec, 2009 2 commits
-
-
Signed-off-by:
-
* Update cdi-osdep.h to contain OS-specific structs for PCI and DMA Signed-off-by:
-
- 06 Dec, 2009 4 commits
-
-
Kevin Wolf authored
* cdi: The destroy function doesn't need to get the driver struct passed as an argument. The driver struct is static anyway. * cdi: Drivers don't need to wrap cdi_{fs,net,storage}_driver if they don't extend the struct (and there is no reason to do so). Signed-off-by: -
Kevin Wolf authored
* cdi: Drivers register their initialization function in the generic driver structur now. They no longer have a main function (if possible), instead the entry point is moved to somewhere in the OS specific CDI library. ! cdi: Drivers need to specify their type Signed-off-by: -
Kevin Wolf authored
* cdi: Initialize driver structures statically instead of using init functions that set the values + cdi: Add a CDI_DRIVER definition to each driver Signed-off-by: -
Kevin Wolf authored
* cdi: OS dependent parts of cdi.h are now in cdi-osdep.h + cdi: CDI_DRIVER macro which is used to declare drivers (in preparation of the removal of main() from the drivers) Signed-off-by:
-
- 05 Dec, 2009 1 commit
-
-
+ Numbers for CDI_VIDEO, CDI_AUDIO, CDI_AUDIO_MIXER, CDI_USB_HCD and CDI_USB; these drivers are expected to be the next CDI drivers, hence they must be ordered. Signed-off-by:
-
- 12 Nov, 2009 1 commit
-
-
This patch merely fixes warnings in the e1000, pcnet, and sis900 drivers. Signed-off-by:
-
- 11 Nov, 2009 1 commit
-
-
This patch changes the pcnet driver to use the cdi_wait_irq function instead of implementing its own solution. Signed-off-by:
-
- 17 Oct, 2009 1 commit
-
-
Kevin Wolf authored
+ Makefile: Add doc target to generate doxygen documentation in both German and English (once comments are bilingual) Signed-off-by:
-
- 04 Sep, 2009 1 commit
-
-
* pcnet: Verwendet jetzt die 2-Klausel-BSD-Lizenz Signed-off-by:
Jörg Pfähler <bluecoder@gmx.de> Signed-off-by:
-
- 21 Aug, 2009 2 commits
-
-
Kevin Wolf authored
Most CDI drivers assume a strange libc like the tyndur one. This patch cleans them up and allows to compile the drivers on Linux. Signed-off-by: -
Kevin Wolf authored
Add a Makefile to the CDI repository. Drivers are compiled using the system compiler, but they are not linked. This allows at least some basic compile testing without moving patches into the source tree of an OS. Signed-off-by:
-
- 20 Aug, 2009 1 commit
-
-
Signed-off-by:
-
- 15 Aug, 2009 1 commit
-
-
+ cdi: This patch adds a CDI driver for the NE2K (PCI) NIC. Signed-off-by:
-
- 11 Aug, 2009 4 commits
-
-
Max Reitz authored
! string.h sollte für memset() & Co. inkludiert werden Signed-off-by:
Max Reitz <xanclic@googlemail.com> Signed-off-by:
Kevin Wolf <mail@kevin-wolf.de>
-
Max Reitz authored
+ SSIZE32-Bit setzen wäre sinnvoll Signed-off-by:
-
Kevin Wolf authored
* e1000: Eigentlich brauchen wir die Debugmeldungen im Moment nicht mehr Signed-off-by: -
Kevin Wolf authored
- cdi: ip-Kommandozeilenparameter in Netzwerktreibern entfernt Signed-off-by:
-
- 07 Aug, 2009 2 commits
-
-
svn-taljeth authored
* cdi: Auch modulare Monolithen koennen mit cdi_run_drivers etwas anfangen, also nicht von CDI_STANDALONE abhaengig machen. Signed-off-by:
-
svn-taljeth authored
* ata: Verwendet jetzt die 2-Klausel-BSD-Lizenz Signed-off-by:
Antoine Kaufmann <toni@tyndur.org> Signed-off-by:
Janosch Gräf <janosch.graef@gmx.net> Signed-off-by:
-
- 06 Aug, 2009 5 commits
-
-
svn-taljeth authored
! rtl8139: tyndur-spezifisches in #ifdef * rtl8139: Unbenoetigtes #include <string.h> entfernt Signed-off-by:
-
svn-taljeth authored
! e1000: Es gibt keinen Grund mehr, Ausgaben auf VT2 zu machen, und portabel ist es sowieso nicht Signed-off-by:
-
svn-taljeth authored
! pcnet: tyndur-spezifischen Code in #ifdef ! pcnet: Unnoetiger tyndur-spezifischer Code entfernt Signed-off-by:
-
svn-taljeth authored
! sis900: tyndur-spezifische Sachen muessen in #ifdef stehen, damit andere Systeme den Treiber 1:1 benutzern koennen. Signed-off-by:
-
svn-taljeth authored
+ cdi: In cdi.h wird TYNDUR definiert, damit Treiber etwas haben, womit sie ihre #ifdefs machen koennen Signed-off-by:
-
- 05 Aug, 2009 3 commits
-
-
svn-meuchler authored
* ramdisk: Verwendet jetzt die 2-Klausel-BSD-Lizenz Signed-off-by:
Alexander Siol <alex@tyndur.org> Signed-off-by:
-
svn-meuchler authored
* serial: Lizenzaenderung zu 2-Klausel-BSD Signed-off-by:
-
svn-taljeth authored
* floppy: Lizenzaenderung zu 2-Klausel-BSD Signed-off-by:
-
