summaryrefslogtreecommitdiffstats
path: root/gfxnvidia.c
Commit message (Collapse)AuthorAgeFilesLines
* Add additional error handling to pcidev_readbar() callersNiklas Söderlund2013-08-231-0/+3
| | | | | | | | | | | This is mostly a leftover of Niklas' "remove exit call from pcidev_init" patch. While not explicitly necessary detecting errors early is usually a good idea. Corresponding to flashrom svn r1718. Signed-off-by: Niklas Söderlund <niso@kth.se> Signed-off-by: Stefan Tauner <stefan.tauner@alumni.tuwien.ac.at> Acked-by: Stefan Tauner <stefan.tauner@alumni.tuwien.ac.at>
* Automatically unmap physmap()sStefan Tauner2013-08-141-9/+2
| | | | | | | | | | | | Similarly to the previous PCI self-clean up patch this one allows to get rid of a huge number of programmer shutdown functions and makes introducing bugs harder. It adds a new function rphysmap() that takes care of unmapping at shutdown. Callers are changed where it makes sense. Corresponding to flashrom svn r1714. Signed-off-by: Stefan Tauner <stefan.tauner@alumni.tuwien.ac.at> Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
* Decouple BAR reading from pci device init, handle errors gracefullyCarl-Daniel Hailfinger2013-01-051-4/+7
| | | | | | | | | | | | | | | Pcidev_init() now returns struct pci_device * instead of a BAR stored in PCI config space. This allows for real error checking instead of having exit(1) everywhere in pcidev.c. Thanks to Niklas Söderlund for coming up with the original error handling patch which was slightly modified and folded into this patch. Move the declaration of struct pci_device in programmer.h before the first user. Corresponding to flashrom svn r1644. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Stefan Tauner <stefan.tauner@alumni.tuwien.ac.at>
* Unify PCI init and let pcidev clean itself upStefan Tauner2013-01-041-5/+1
| | | | | | | | | | | | | | | | | | | Previously the internal programmer used its own code to initialize pcilib. This patch extracts the common code from the internal programmer and pcidev_init() into pcidev_init_common(). This fixes the non-existent PCI cleanup of the internal programmer and adds an additional safety by checking for an already existing PCI context. We got a nice shutdown function registration infrastructure, but did not use it very wisely. Instead we added shutdown functions to a myriad of programmers unnecessarily. In this patch we get rid of those that do only call pci_cleanup(pacc) by adding a shutdown function the pcidev.c itself that gets registered by pcidev_init(). Corresponding to flashrom svn r1642. Signed-off-by: Stefan Tauner <stefan.tauner@alumni.tuwien.ac.at> Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
* Unify usbdev_status and pcidev_status into dev_entryStefan Tauner2012-12-271-1/+1
| | | | | | | | | | | | | | | | | | | Once upon a time usbdev_status was created for the ft2232 programmer. Its IDs are semantically different to pcidev_status because they indicate USB instead of PCI IDs, but apart from that both data structures are equal. This change makes life easier for everything involved in handling and printing the status of devices that is noted in those structures by combining them into dev_entry. It is still possible to distinguish between PCI and USB devices indirectly by using the struct programmer's type field. Also, add a programmer column to the PCI and USB devices lists. Corresponding to flashrom svn r1632. Signed-off-by: Stefan Tauner <stefan.tauner@alumni.tuwien.ac.at> Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
* Some ISO C fixesCarl-Daniel Hailfinger2012-08-271-1/+1
| | | | | | | | | | | | | | | | | This patch just fixes a limited number of bits not conforming to c99 by using - __asm__ instead of just asm - {0} instead of {} for struct initialization - h_addr_list[0] instead of h_addr to access the host address in struct hostent - #include <strings.h> where needed (for ffs and strcasecmp) Based on a previous patch by Carl-Daniel. Corresponding to flashrom svn r1585. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Signed-off-by: Stefan Tauner <stefan.tauner@alumni.tuwien.ac.at> Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
* Automatically release I/O permissions on shutdownCarl-Daniel Hailfinger2012-07-211-2/+2
| | | | | | | | | | | | | | | | | | | | | Get_io_perms() is renamed to rget_io_perms() and automatically registers a function to release I/O permissions on shutdown. Actually release I/O permissions on Solaris and iopl()-supporting operating systems like Linux. This patch fixes quite a few programmers which forgot to release I/O permissions on shutdown, and it simplifies the shutdown and error handling code for all others. Do not call exit(1) if I/O permissions are denied and return an error instead. This part of the patch was written by Niklas Söderlund. Corresponding to flashrom svn r1551. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Signed-off-by: Niklas Söderlund <niso@kth.se> Acked-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
* Hide hwaccess.h from public APIPatrick Georgi2012-07-201-0/+1
| | | | | | | | | | | | | Move hwaccess.h #include from flash.h to individual drivers. libflashrom users need flash.h, but they do not care about hwaccess.h and should not see its definitions because they may conflict with other hardware access functions and #defines used by the libflashrom user. Corresponding to flashrom svn r1549. Signed-off-by: Patrick Georgi <patrick.georgi@secunet.com> Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
* Add struct flashctx * parameter to all functions accessing flash chipsCarl-Daniel Hailfinger2011-12-181-2/+8
| | | | | | | | | | | | | | | | All programmer access function prototypes except init have been made static and moved to the respective file. A few internal functions in flash chip drivers had chipaddr parameters which are no longer needed. The lines touched by flashctx changes have been adjusted to 80 columns except in header files. Corresponding to flashrom svn r1474. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
* Register Parallel/LPC/FWH programmers the same way SPI programmers are ↵Carl-Daniel Hailfinger2011-11-091-2/+12
| | | | | | | | | | | | | | | | | | | | registered All programmers are now calling programmer registration functions and direct manipulations of buses_supported are not needed/possible anymore. Note: Programmers without parallel/LPC/FWH chip support should not call register_par_programmer(). Additional fixes: Set max_rom_decode.parallel for drkaiser. Remove abuse of programmer_map_flash_region in it85spi. Annotate several FIXMEs in it85spi. Corresponding to flashrom svn r1463. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
* Random whitespace and coding-style fixesUwe Hermann2011-07-281-1/+1
| | | | | | | | | | | | Also, indentation fixes, e.g. due to conversion to msg_*, use ARRAY_SIZE where possible, wrap overly long line, etc. Compile-tested. There should be no functional changes. Corresponding to flashrom svn r1397. Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de> Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
* Rename CHIP_BUSTYPE_FOO to BUS_FOOCarl-Daniel Hailfinger2011-07-271-1/+1
| | | | | | | | | It's shorter to type, and we have less problems with the 80 column limit. Corresponding to flashrom svn r1396. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
* Use shutdown callback mechanism to shutdown programmersDavid Hendricks2011-06-141-12/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | This patch attempts to resolve some programmer shutdown ordering issues by having the programmer init functions register shutdown callbacks explicitly wherever it makes most sense. Before, assumptions were made that could lead to the internal programmer's state changing before the external programmer could be shut down properly. Now, each programmer cleans up after itself and (hopefully) performs each operation in the correct order. As a side-effect, this patch gives us a better usage model for reverse operations such as rpci_* and rmmio_*. In the long-run, this should make reversing the initialization process easier to understand, less tedious, and less error-prone. In short, this patch does the following: - Registers a shutdown callback during initialization for each programmer. - Kills the .shutdown function pointer from programmer_entry struct. Also, make most shutdown functions static. - Adds a few minor clean-ups and corrections (e.g. missing physunmap() calls). TODO: Remove forward declaration of serprog_shutdown() (added to simplify diff) Corresponding to flashrom svn r1338. Signed-off-by: David Hendricks <dhendrix@google.com> Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
* Remove vendorid parameter from pcidev_init()Carl-Daniel Hailfinger2011-03-071-2/+1
| | | | | | | | | | | | | | | Simplify pcidev_init by killing the vendorid parameter which was pretty useless anyway since it was present in the pcidevs parameter as well. This also allows us to handle multiple programmers with different vendor IDs in the same driver. Fix compilation of flashrom with only the nicrealtek driver. Corresponding to flashrom svn r1274. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
* Revert PCI config space writes on shutdownCarl-Daniel Hailfinger2010-11-101-8/+4
| | | | | | | | | | | | | | | This means all chipset enables etc. will be undone on shutdown. Reversible PCI config space writes now use rpci_write_*(). PCI config space writes which are one-shot (e.g. communication via config space) should continue to use the permanent pci_write_* variants. Extend the number of available register_shutdown slots to 32. Corresponding to flashrom svn r1232. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
* Remove duplicate includes from the codeStefan Reinauer2010-10-061-1/+0
| | | | | | | Corresponding to flashrom svn r1196. Signed-off-by: Stefan Reinauer <stepan@coresystems.de> Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
* Compile gfxnvidia by default, but disallow write/eraseCarl-Daniel Hailfinger2010-07-291-0/+3
| | | | | | | Corresponding to flashrom svn r1117. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
* Split off programmer.h from flash.hCarl-Daniel Hailfinger2010-07-271-0/+1
| | | | | | | | | | | | | | | | | | | | | Programmer specific functions are of absolutely no interest to any file except those dealing with programmer specific actions (special SPI commands and the generic core). The new header structure is as follows (and yes, improvements are possible): flashchips.h flash chip IDs chipdrivers.h chip-specific read/write/... functions flash.h common header for all stuff that doesn't fit elsewhere hwaccess.h hardware access functions programmer.h programmer specific functions coreboot_tables.h header from coreboot, internal programmer only spi.h SPI command definitions Corresponding to flashrom svn r1112. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
* Convert MMIO accesses of non-internal PCI-based programmers to be ↵Carl-Daniel Hailfinger2010-07-271-2/+2
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | endian-agnostic Convert all PCI-based external programmers to use special little-endian accessors for all MMIO regions of PCI devices. This patch does _not_ touch the internal programmer (which is PCI-based as well). Huge thanks go to Misha Manulis who worked with me to create a first version of this patch for the satasii programmer based on modification of generic code. Huge thanks also go to Segher Boessenkool for suggesting the pci_mmio_ prefix for the abstraction layer. NOTE to package maintainers: With this patch, compilation and usage of flashrom should be safe on x86, x86_64, MIPS (little and big endian) and PowerPC (big endian). The internal programmer is disabled on non-x86/x86_64 (but it compiles). The atahpt, nic3com, nicnatsemi, nicrealtek and rayer_spi can not be compiled on non-x86/x86_64 because port space I/O is not (yet) supported. Please compile with default settings on x86/x86_64 and with the following settings on all other architectures: make CONFIG_NIC3COM=no CONFIG_NICREALTEK=no CONFIG_NICNATSEMI=no CONFIG_RAYER_SPI=no Corresponding to flashrom svn r1111. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Misha Manulis <misha@manulis.com>
* Refine PCI BAR masks handling for drkaiser and gfxnvidiaCarl-Daniel Hailfinger2010-07-171-2/+7
| | | | | | | | | | | | | | Use the BAR value returned by pcidev_init which automatically applies the correct BAR mask for the drkaiser driver. Truncate flash chip addresses to fit into the 128 kB memory window for drkaiser and pick the same window size for gfxnvidia. Uwe tested all operations successfully on a Dr. Kaiser card. Corresponding to flashrom svn r1089. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
* Various places in the flashrom source feature custom parameter extraction ↵Carl-Daniel Hailfinger2010-07-061-2/+2
| | | | | | | | | | | | | | | | | | | | from programmer_param This led to wildly differing syntax for programmer parameters, and it also voids pretty much every assumption you could make about programmer_param. The latter is a problem for libflashrom. Use extract_param everywhere, clean up related code and make it more foolproof. Add two instances of exit(1) where we have no option to return an error. Remove six instances of exit(1) where returning an error was possible. WARNING: This changes programmer parameter syntax for a few programmers! Corresponding to flashrom svn r1070. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
* Kill global variables, constants and functions if local scope sufficesCarl-Daniel Hailfinger2010-07-031-1/+1
| | | | | | | | | | | | | | | | | | | | | | | Constify variables where possible. Initialize programmer-related variables explicitly in programmer_init to allow running programmer_init from a clean state after programmer_shutdown. Prohibit registering programmer shutdown functions before init or after shutdown. Kill some dead code. Rename global variables with namespace-polluting names. Use a previously unused locking helper function in sst49lfxxxc.c. This is needed for libflashrom. Effects on the binary size of flashrom are minimal (300 bytes shrinkage), but the data section shrinks by 4384 bytes, and that's a good thing if flashrom is operating in constrained envionments. Corresponding to flashrom svn r1068. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net> Acked-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de>
* Replace PCI_OK/PCI_NT by OK/NTMichael Karcher2010-02-241-23/+23
| | | | | | | | | | We don't need to duplicate OK and NT as PCI_OK and PCI_NT if the symbols are already there (defined for the chipset enable table). Corresponding to flashrom svn r911. Signed-off-by: Michael Karcher <flashrom@mkarcher.dialup.fu-berlin.de> Acked-by: Uwe Hermann <uwe@hermann-uwe.de>
* Convert all messages in gfxnvidia.c to the new message infrastructureSean Nelson2010-01-091-1/+1
| | | | | | | Corresponding to flashrom svn r850. Signed-off-by: Sean Nelson <audiohacked@gmail.com> Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
* Add initial support for flashing some NVIDIA graphics cardsUwe Hermann2009-09-301-0/+104
The new option is '-p gfxnvidia', rest of the interface is as usual. I tested a successful identify and read on a "RIVA TNT2 Model 64/Model 64 Pro" card for now, erase and write did NOT work properly so far! Please do not attempt to write/erase cards yet, unless you can recover! In addition to the NVIDIA handling code it was required to call programmer_shutdown() in a lot more places, otherwise the graphics card will be disabled in the init function, but never enabled again as the shutdown function is not called. The shutdown handling may be changed to use atexit() later. Corresponding to flashrom svn r737. Signed-off-by: Uwe Hermann <uwe@hermann-uwe.de> Acked-by: Luc Verhaegen <libv@skynet.be>
OpenPOWER on IntegriCloud