aboutsummaryrefslogtreecommitdiffstats
path: root/com32/lib
Commit message (Collapse)AuthorAgeFilesLines
* pci: Rename get_module_name_from_pci_idssyslinux-3.74-pre13syslinux-3.74-pre12Erwan Velu2009-03-291-1/+1
| | | | | | | | | Impact: none Rename get_module_name_from_pci_ids to get_module_name_from_pcimap since the module names are read from the modules.pcimap and not pci.ids file. Signed-off-by: Sebastian Herbszt (herbszt@gmx.de)
* pci: Check if configuration type is PCI_CFG_NONEErwan Velu2009-03-261-0/+7
| | | | | | Impact: prevent from grabbing configuration when PCI is PCI_CFG_NONE Signed-off-by: Sebastian Herbszt (herbszt@gmx.de)
* pci: Remove "(void)cfgtype"Erwan Velu2009-03-261-2/+0
| | | | | | | Impact: none Remove useless "(void)cfgtype" Signed-off-by: Sebastian Herbszt (herbszt@gmx.de)
* vesacon: remove unused com32sys_t variablessyslinux-3.74-pre11H. Peter Anvin2009-03-201-3/+0
| | | | | No more direct __intcall()'s in this code, so remove unused register set variables.
* vesacon: migrate disable fixes from ansiconH. Peter Anvin2009-03-201-7/+8
| | | | | Copy the disabled fixes already done in ansicon from checkin 22cd4e378e3aafd402f43030941d6f6b300d0cea.
* ansicon: handle the disabled case properlyH. Peter Anvin2009-03-201-6/+6
| | | | | | | | | Even when disabled, we need to set the screen size, because the downstream code depends on it. Set it to 25x80 (maybe we should swtich this is 24x80) in this case. Furthermore, when disabled, we should return success == bytes "written", not zero.
* vesa: fixing wrong total_memory sizeErwan Velu2009-03-141-1/+1
| | | | | | Regarding the vesa spec, total_memory is reported to be 16bits That 32bits field had shifted the whole structure and prevented reading other fields properly.
* hdt: Adding PCI latencyErwan Velu2009-03-141-0/+1
| | | | Adding the pci latency information
* hdt: irq have to be in dev_infoErwan Velu2009-03-141-1/+7
| | | | Moving irq stuff to the dev_info structure to keep coherent
* hdt: Adding legacy IRQ stuffErwan Velu2009-03-131-1/+28
| | | | | | | | In the cli mode, the pci context feature the following : pci> show irq Under cli, "pci> show device <x>" display the IRQ when device features one. The menu mode behave the same way.
* Merge branch 'hdt'H. Peter Anvin2009-03-125-52/+263
|\
| * libcom32: add strpcpy()H. Peter Anvin2009-03-122-2/+22
| | | | | | | | | | Add strpcpy() -- by analogy with mempcpy(), a strcpy() which returns a pointer to the end of the string (specifically, to the final null.)
| * Moving dmi & cpuid stuff to gpl subdirsErwan Velu2009-03-128-1502/+0
| | | | | | | | | | | | | | | | | | Fixing copyrights Moving dmi includes to gplinclude/dmi Moving dmi code to gpllib/dmi/ Moving cpuid includes to gplinclude/ Moving cpuid code to gpllib/ Fixing Makefiles accordingly
| * com32/gpllib: clean up Makefile structureH. Peter Anvin2009-03-122-1/+1
| |
| * Add gpllib, gplincludeH. Peter Anvin2009-03-121-0/+5
| | | | | | | | | | Add separate gpllib and gplinclude directories, and have a NOGPL variable in MCONFIG that can be set to exclude these directories.
| * PCI: flexible modules.pcimap & pci.ids pathsErwan Velu2009-03-111-6/+6
| | | | | | | | | | | | While detecting the pci names, class name & kernel modules, it's better to let the user choosing the path instead of the harcoded value "/" is not always the wanted path
| * hdt: Cleaning DMI implementationerwan2009-03-031-0/+3
| | | | | | | | | | | | Previous commit was incomplete. Removing useless warnings by defining structures & functions properly Using existing types instead of u8,u16,u32
| * hdt:Cleaning DMI implementationerwan2009-03-027-0/+1499
| | | | | | | | | | Removing useless warnings by defining structures & functions properly Using existing types instead of u8,u16,u32
| * hdt: PCI: Adding categoryErwan Velu2009-02-261-1/+4
| |
| * hdt: Adding pxe_get_nic_type(t_PXENV_UNDI_GET_NIC_TYPE *gnt)Erwan Velu2009-02-252-1/+62
| |
| * hdt: We have to iterate & copy the vendor name to all matching pci devices ↵Erwan Velu2009-02-251-4/+13
| | | | | | | | as all pci devices might not have a matching vendor/product id. This will help us having the proper vendor device with an unknown product.
| * hdt: Adding more defined values for pci components Enabling mulitple kernel ↵Erwan Velu2009-02-251-21/+26
| | | | | | | | modules per pci device Updating pcitest
| * hdt: Returning ENOPCIIDS and ENOMODULESPCIMAP when pci.ids and ↵Erwan Velu2009-02-251-3/+3
| | | | | | | | modules.pcidmap files are missing
| * hdt: Improving performances by not computing product ids when the vendor ↵Erwan Velu2009-02-251-30/+25
| | | | | | | | | | | | doesn't exist on the local system. Removing useless break() that prevent other com32 module to load
| * hdt: Improving detection speed by skipping uncessary linesErwan Velu2009-02-251-0/+13
| |
| * hdt: No need to only copy vendor nameErwan Velu2009-02-251-8/+6
| |
| * hdt: Performance optimisations : avoiding hex_to_int in pci_domain loopsErwan Velu2009-02-251-14/+29
| |
| * hdt: Adding get_class_name_from_pci_ids to grab class name from the pciidsErwan Velu2009-02-251-0/+87
| |
| * hdt: Improving clean targetErwan Velu2009-02-251-1/+1
| |
* | linux.c32: fix initrd alignmentH. Peter Anvin2009-03-121-4/+4
| | | | | | | | | | | | | | | | | | | | | | | | Impact: fix boot failure on Dell E6500 and possibly other platforms Linux may not be able to access a part of an initrd which resides in a page that is part RAM, part non-RAM. Make sure we round the end of the memory range we want to use down to a page boundary (as well as round the start up.) Reported-by: Pierre-Alexandre Meyer <pierre@mouraf.org> Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* | shuffler: more corrections to shuffle buffer assignmentH. Peter Anvin2009-03-101-12/+19
| | | | | | | | We really need to be able to allocate the shuffle buffer itself...
* | shuffler: fix the handling of more than one block of descriptorsH. Peter Anvin2009-03-101-7/+15
|/ | | | | | When allocating more than one block of descriptors, make sure we account for all the memory used correctly. Otherwise, very bad things happen.
* syslinux_derivative_info(): use the buffer we already haveH. Peter Anvin2009-02-181-4/+3
| | | | Use the data buffer already set aside for syslinux_derivative_info().
* Constructed data objects can't be common; hack derivative_infoH. Peter Anvin2009-02-181-7/+5
| | | | | | | | | - Constructed data objects can't be common. Earlier version of gcc didn't if the objects were explicitly listed extern, but newer ones need an explicit __attribute__((nocommon)) or -fno-common. - Make syslinux_derivative_info() save the entire reply.
* getcwd.c: need <string.h>H. Peter Anvin2009-02-181-0/+1
| | | | getcwd.c needs <string.h>
* COM32 API: Add functions for directory useGene Cumm2009-02-107-0/+181
| | | | | | | | | COM32: Add directory functions getcwd(), opendir(), readdir() and closedir(). This depends on the patch that I just submitted creating the COMBOOT API calls. Signed-off-by: Gene Cumm <gene.cumm@gmail.com> Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* Add comboot interface to access the keyboard remapping tableH. Peter Anvin2009-02-032-0/+45
| | | | | | Add a comboot interface to access the keyboard remapping table, which can be used for either querying or changing the keyboard map.
* com32: make _Exit just an ordinary symbolH. Peter Anvin2008-10-141-0/+7
| | | | | | | | Make _Exit just an ordinary symbol which aliases _exit. This avoids some versions of gcc complaining about the inline, which is rather unnecessary anyway. Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* Older binutils wants parens around macrosDag Wieers2008-10-141-1/+1
| | | | | | | Older versions of gcc/binutils apparently really want parentheses around macro arguments. Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* e820: pre-initialize bufferH. Peter Anvin2008-10-131-0/+2
| | | | | | | | Pre-initialize the e820 buffer to all zero, except for the active bit; we have found that BIOSes will make assumptions about the pre-existing value of the buffer and this seems the most conservative value. Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* com32/lib: fix reversed test when looking for extattr in e820H. Peter Anvin2008-10-131-1/+1
| | | | | | | Fix reversed test when looking for extattr in e820: force the extattr values only when they are *not* present... Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* Support "extended attributes" for INT 15h, AX=E820hH. Peter Anvin2008-10-131-1/+8
| | | | | | | | | | | | | | | Some blithering idiot thought it was a good idea to introduce "extended attributes" for INT 15h, AX=E820h, and in doing so, breaking compatibility with ALL E820 users out there. F*cking morons. Implement handling of extended attributes in: - e820 parsing in the core - e820 parsing in libcom32 - e820 parsing *and proxying* in memdisk The latter is the really painful one. Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* Remove line-leading gas-style comments in files compiled with -std=gnu99H. Peter Anvin2008-09-263-45/+47
| | | | | | | | Some versions of gcc barf if we have hash-style comments in files when -std=gnu99 is used. Remove from assembly files that are subject to being assembled that way. Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* setjmp.S: use C-style commentsH. Peter Anvin2008-09-261-15/+15
| | | | | | | Earlier versions of cpp would get confused by the gas-style comments at line starts; replace them with C-style comments. Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* Fix "make clean" confusionsyslinux-3.72H. Peter Anvin2008-09-251-4/+2
| | | | | | | | "make clean" must NOT remove stuff that "make installer" should not rebuild. This is a somewhat unfortunate choice of naming, but that's the way it is. Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* com32/lib: fix "make clean" by proper parens for "find"H. Peter Anvin2008-09-251-1/+1
| | | | | | "-o" clauses bind looser than -print0, so we need parens. Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* ansicon, vesacon: beep on ^GH. Peter Anvin2008-09-094-0/+16
| | | | | | If we get a ^G, issue a beep. Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* Major Makefile cleanups; gcc 4.3.0 compatiblityH. Peter Anvin2008-08-201-2/+2
| | | | | Cleanup and centralize the Makefile system even more. Fix a gcc 4.3 incompatibility in memdisk (definition of strlen).
* pci: fix pci_scan()Sebastian Herbszt2008-08-121-2/+4
| | | | Signed-off-by: H. Peter Anvin <hpa@zytor.com>
* pci: cosmetic fixesSebastian Herbszt2008-08-121-34/+34
| | | | Signed-off-by: H. Peter Anvin <hpa@zytor.com>