summaryrefslogtreecommitdiffstats
path: root/pci.c
Commit message (Collapse)AuthorAgeFilesLines
* (get_pci_description) introduce new API for drakx's stage1 in order to find ↵Thierry Vignaud2011-10-181-0/+20
| | | | unknown PCI devices
* (pci_probe) don't exit(1) on systems without pciThierry Vignaud2011-10-181-0/+5
| | | | | | | | | arm systems don't always have pci busses so killing the app for this reasons is just bad idea. The workaround is to check for pci procfs files and if not found return an empty set. v1: by rtp v2: simplify (tv)
* (pci_probe) do not crash if there're more than 100 PCI devicesThierry Vignaud2009-12-011-1/+1
| | | | (usb_probe() was already safe)
* (pci_probe,usb_probe) commentThierry Vignaud2009-12-011-0/+2
|
* (pci_probe) no need to lookup extended PCI capabilitiesThierry Vignaud2009-12-011-1/+1
|
* (pci_probe) simplify reusing pci_find_cap()Thierry Vignaud2009-12-011-7/+1
|
* class_prog is now unusedThierry Vignaud2009-09-301-2/+0
|
* (pci_probe) explainThierry Vignaud2009-09-301-0/+1
|
* (pci_probe) kill useless empty lineThierry Vignaud2009-09-301-1/+0
|
* restore a lost commentThierry Vignaud2009-09-301-0/+1
|
* typo fixThierry Vignaud2009-09-301-5/+5
|
* (pci_probe) no need to allocate that much space on the stackThierry Vignaud2009-09-301-2/+2
|
* (pci_probe) use libpci in order to retrieve PCI capabilities in order to haveThierry Vignaud2009-09-301-1/+11
| | | | a better test for identifying PCI Express devices
* (pci_probe) reuse more stuff from libpci in order to retrieve PCI subThierry Vignaud2009-09-301-6/+5
| | | | ids, class and revision
* (pci_probe) simplify using libpci in order to retrieve PCI config spaceThierry Vignaud2009-09-301-14/+3
|
* (pci_probe) change type of buffer (no real change)Thierry Vignaud2009-09-301-1/+1
|
* (pci_probe) make sure not to retrieve values from previous devicesThierry Vignaud2009-09-301-0/+1
|
* split out CONFIG_SPACE_ZIZE (needed for next commits)Thierry Vignaud2009-09-301-1/+3
|
* (pci_probe) fix inverted test for choosing between '8139cp' &Thierry Vignaud2009-09-281-1/+1
| | | | | '8139too' drivers (#53349)
* (pci_probe) try harder to fetch the right driver between '8139cp' &Thierry Vignaud2009-09-231-0/+7
| | | | | '8139too' according to PCI revision (#53349)
* (pci_probe) retrieve PCI revision (#42576)Thierry Vignaud2009-09-231-0/+1
|
* kill doble includeThierry Vignaud2009-09-231-1/+0
|
* Do not lose pointer when realloc decides to move itPascal Terjan2009-06-251-1/+2
|
* (pci_probe) adapt to pciutils-3.x APIThierry Vignaud2008-05-141-2/+1
|
* try harder to reduce number of relocations (most remaining ones are due toThierry Vignaud2007-08-271-2/+2
| | | | | arrays of pointers)
* remove old 8139too/gdth hardcoded rules (already in modules.alias)Olivier Blin2007-08-141-10/+0
|
* (pci_probe) simplifyThierry Vignaud2007-08-041-2/+1
|
* (pci_probe) kill some quirks that are either obsolete now that weThierry Vignaud2007-08-041-30/+1
| | | | | | resolve modalias (which bring wildcard support) or were removed from the kernel (eg: sata_nv do not more probe on class)
* reuse modprobe code in order to resolve modalias (need to be optimized)Thierry Vignaud2007-08-041-1/+1
|
* (pci_probe) typo fix :-(Thierry Vignaud2007-02-261-1/+1
|
* export PCI domainThierry Vignaud2007-02-261-0/+1
|
* (pci_probe) get PCI class as reported by pciutils (might be used by harddrakeThierry Vignaud2007-02-261-1/+3
| | | | | in order to display something nicer than ->{media_type})
* (struct pciusb_entry) rename class_ as class_idThierry Vignaud2007-02-261-6/+6
|
* (pci_probe) remove an extra space wronly added in r125837Thierry Vignaud2007-02-261-1/+1
|
* (pci_probe) replace a couple magic valyes by constants from pciutilsThierry Vignaud2007-02-261-3/+4
|
* stop performing descriptions lookup from pcitable for PCIThierry Vignaud2007-02-261-1/+1
|
* (pci_probe) use pciutils in order to get device description from ↵Thierry Vignaud2007-02-261-0/+5
| | | | /usr/share/pci.ids
* switch to pciutils as PCI listing backendThierry Vignaud2007-02-261-62/+38
|
* always initialize n_pci_domains, stick to 0 on opendir() error and no matchGwenolé Beauchesne2006-01-051-1/+1
|
* add support for pci domainsGwenolé Beauchesne2006-01-051-2/+50
|
* oops: sata_via doesn't support for CLASS probing whereas sata_nv doesThierry Vignaud2005-03-301-2/+2
|
* libify and simplifyPascal Rigaux2005-03-141-3/+1
|
* setting .nb to 0 seems a better idea than setting .entriesPascal Rigaux2005-03-111-1/+2
|
* handle a few more special cases (gdth, snd-vx222, 8139too, and agp bridges)Thierry Vignaud2005-02-171-0/+20
|
* detect VIA SATA controllers since new sata_via.c driver probes themThierry Vignaud2005-02-171-0/+2
| | | | | | this way... (we should really add a class/... matching table like kernel's pcimap is)
* all PCI_CLASS_BRIDGE_CARDBUS cards are yenta_socket (says kudzu)Pascal Rigaux2004-12-071-2/+8
|
* do full-probe by default; remove support for no full-probeThierry Vignaud2003-08-191-8/+2
|
* - detect ohci1394 & ehci-hcd based on the pci classPascal Rigaux2003-07-311-2/+13
| | | | | (as done in RedHat's kudzu)
* Use read() instead of fread() to read from "/proc/bus/pci/%02x/%02x.%d".Pascal Rigaux2003-04-221-4/+8
| | | | | | | | | | | | Thanks a lot to Tom Cox for finding this bug: The proc.c module in the kernel source clearly states that reading more than 64 bytes can cause problems. The pci.c module in the ldetect library uses the buffered fread() function. This function always reads in blocks, so when run as root, the read always tried to read more than the user requested amount.
* no error message when -p is not used and there is neither pci nor usb busPascal Rigaux2002-11-051-8/+11
|