Commit Graph

97326 Commits

Author SHA1 Message Date
joerg
8800d320f1 Define wapbl_flush_fn_t only for the kernel. 2008-11-10 20:30:31 +00:00
cliff
b5cb47739c - nullcom.c provides nulldev-like console device for silent gzboot
- GEMINI_dram_0x01600000/Makefile uses nullcom, shhh
2008-11-10 20:30:12 +00:00
joerg
3fbdfc8af9 Reduce internals of WAPBL exposed to the rest of the system. 2008-11-10 20:12:13 +00:00
joerg
5787d39971 Fix typo in fw name. 2008-11-10 18:05:12 +00:00
cegger
cf508a059b mainbus_attach: reduce diff a little between i386 and amd64 2008-11-10 14:36:59 +00:00
cliff
ac51de51b3 - slave CPU using LPC UART or console
- don't use VERBOSE_INIT_ARM
- dont configure obio UART
- using intr -1 for now, causes callout interrupt polling
  can use that until serirq from 8712 works
2008-11-10 04:39:43 +00:00
cliff
129e7690e8 master cpu does not need LPC stuff, comment out for now 2008-11-10 04:37:14 +00:00
cliff
3b02e84035 - GEMINI_CONSOLE_VBASE renamed to GEMINI_UART_VBASE 2008-11-10 04:35:25 +00:00
cliff
5f656fb56e - use actual device address defines instead of abstract console defines
in mmu_init_table\gemini_start.S and devmap[]\gemini_machdep.c;
  this allows some flexibility e.g. to map 'em all, which is useful
  e.g. when using obio uart to debug lpc uart.

- add mmu_init_table entries for Gemini global regs, lpchc, and lpcio
  these are needed to consinit() when console is lpc uart
2008-11-10 04:33:42 +00:00
cliff
23f46ff40f - use actual device address defines instead of abstract console defines for in mmu_init_table\gemini_start.S and devmap[]\gemini_machdep.c;
this allows some flexibility e.g. to map 'em all, which is useful
  e.g. when using obio uart to debug lpc uart.

- add mmu_init_table entries for Gemini global regs, lpchc, and lpcio
  these are needed to consinit() when console is lpc uart

- now have 2 versions of consinit, one for obio uart,
  the other for lpc uart; we use CONSADDR to determine
  which gets compiled.

- remove a useless bus_space map/unmap from obio uart version of consinit
  comcnattach() will do his own...
2008-11-10 04:30:46 +00:00
cliff
a8ab989c8b in critical_devs[], master cpu gets com0;
slave cpu gets geminilpchc0 (and children)
2008-11-10 04:11:17 +00:00
cliff
332c188109 reduce callout polling period to (hz/16)
once we get serial irq from 8712 uart working
this should become a non-issue
2008-11-10 04:07:30 +00:00
cliff
f5b85be633 gemini_lpc_bus_ops is needed for consinit(), so remove 'static' 2008-11-10 04:05:35 +00:00
uebayasi
ed729f9508 Whitespace. 2008-11-10 01:06:43 +00:00
bouyer
71d4948e50 Fix kern/39769: race condition in TCP timers
When a TCP timer is disarmed (with callout_stop()) in the general case
callout_invoking() isn't checked, so the timer handler could be called run
when the current interrupt handler exits, athough the timer is disarmed.
This case cause bad things like TCPT_REXMT and TCPT_PERSIST being both pending,
causing a panic (see the PR for details).
Close the issue by aborting the handler if the timer is not callout_expired().
(the EXPIRED flag being cleared by callout_stop()).
2008-11-09 17:32:38 +00:00
martin
3ab32a0a5e KNF is nice, but not if it causes a warning that breaks compilation :-) 2008-11-09 15:48:13 +00:00
cegger
76d489193e struct device * -> device_t 2008-11-09 15:34:14 +00:00
bjs
8b2e3ec035 Use agp_flush_cache() instead of wbinvd(). 2008-11-09 14:26:14 +00:00
cegger
32c1f8c732 Nuke last parameter from mpaci_scan_apics() and mpbios_scan().
It is unused.
2008-11-09 14:24:14 +00:00
pgoyette
7355d18b97 Leave the task of displaying sensor states to envsys_sysmon() unless we
are in debug mode.
2008-11-09 13:54:06 +00:00
bouyer
5da863e7e9 Assert that bp->b_error == 0 if all data have been transfered in
physio_biodone(). We have more chance to have the fauly code in the
stack trace when doing the check here than in physio_done().
2008-11-09 12:18:07 +00:00
cliff
fe95ceb60e these kernel configs are lmeant to be used to create flashable memory disk rooted
embedded kernels using non-SMP master/slave dual core division of resources.
still very much in progress at this time.
2008-11-09 09:39:18 +00:00
cliff
443abea81b - use option GEMINI_MASTER to pick up clock, icu
- ATADEBUG now here but commented out
- deconfigure timer#1 now that system clock and stat clock are combined
- add umass and wd for use w/ ehci
- uncomment wdc0, wd1; add atabus and wd* for use w/ wdc
- warning the gzipped binary GEMINI kernel image will not fit  in a 1MB flash area, even gzipped
2008-11-09 09:36:08 +00:00
cliff
bd4b8f9765 - add autoconfig support for wdc at Gemini obio
- add autoconfig support for com type UART on IT8712 superio chip
  attached at lpc bus at lpc host controller at Gemini obio.
2008-11-09 09:19:44 +00:00
cliff
2842dccfaa preliminary support for com type UART in IT8712 superio chip
attached at LPC bus at Gemini LPC Host Controller at obio on Gemini.
2008-11-09 09:15:42 +00:00
cliff
3100fdac39 - we are combining stat and system clocks to use the same timer
the master and slave CPU each get one timer for those functions;
  both CPUs can use the free-running ref timer;
  it is initialized by CPU#0 and after that all access is read-only
- use correct OBIO_*_DEFAULT locator indicies
2008-11-09 09:11:09 +00:00
cliff
694f117a78 - use #if NPCI > 0 arond pci stuff so we can compile when pci is not configured
- in critical_devs[], divide up timers attachment between master and slave CPUs
  and provide icu#1 attachment for slave cpu
2008-11-09 09:04:33 +00:00
cliff
6a1577ee20 autoconfig glue for wdc at obio on Gemini 2008-11-09 08:59:05 +00:00
cliff
fcb80bc9f2 - we are combining stat and system clocks, so STATHZ is now HZ 2008-11-09 08:56:47 +00:00
cliff
7a2b04a933 - GEMINI_LPCP_BASE is renamed GEMINI_LPCIO_BASE
- add some minimal Gemini GLOBAL reg defines
- add Gemini LPC Host Controller reg defines
- add Gemini MIDE Controller reg defines
2008-11-09 08:54:46 +00:00
cliff
ed27a42072 - initarm progress putchars and printfs are now #ifdef VERBOSE_INIT_ARM
- use GEMINI_SLAVE, GEMINI_MASTER defines to distinguish between
  primary and secondary CPUs, who owns what devices, etc.
- add short delay()s to allow console printf output to drain before reset
  unless if ref_sc is not yet initialized, in which case tough.
- add devmap entriy for LPC
- add devmap entry for RAMDISK if MEMORY_DISK_DYNAMIC is defined
  i.e. if the ramdisk is external to the kernel, in which case
  we "know" where is lives by agreement with firmware.
- call md_root_setconf() if MEMORY_DISK_DYNAMIC is defined
2008-11-09 08:49:40 +00:00
cliff
a80e63a8a0 add defines for LPC and RAMDISK VADDRs for use in devmap 2008-11-09 08:40:59 +00:00
cliff
8e3f3d7b7b those progress putchars are now #ifdef VERBOSE_INIT_ARM 2008-11-09 08:39:31 +00:00
abs
66b0aa634e Regenerate from GENERIC.in 1.75
Enable 68060 support in FALCON & BOOTX kernels for CT60/63 upgrades
Drop KERNFS from INSTALL kernels
Drop 68040 support from ATARITT kernels
2008-11-09 07:50:39 +00:00
abs
0282f5b26e Enable 68060 support in FALCON & BOOTX kernels for CT60/63 upgrades
Drop KERNFS from INSTALL kernels
Drop 68040 support from ATARITT kernels
2008-11-09 07:49:59 +00:00
tsutsui
9f2a500c35 Make compile this with options DIAGNOSTIC. 2008-11-09 07:01:42 +00:00
pgoyette
18913ef363 Temp sensor should report a critical-overtemp condition (ENVSYS_SCRITOVER)
not a simple critical (ENVSYS_SCRITICAL) condition.
2008-11-09 03:36:04 +00:00
christos
4c2b1bf33a PR/39885: Sébastien Bocahu: add R500 chips support 2008-11-09 02:35:42 +00:00
christos
2688948f51 regen. 2008-11-08 17:27:15 +00:00
christos
71ede8ab96 Support for Intel G35 as found on Asus P5E-VM HDMI motherboard from
Milos Negovanovic
2008-11-08 17:26:28 +00:00
he
ddda8f3c69 Bump install ramdisk image size from 1440k to 1500k, so that it fits again.
The install kernel is subsequently compressed, so there should be nothing
magic about the 1440k constant.
2008-11-08 10:24:23 +00:00
he
22d492ac84 Adapt to changes in arm/arm32/mem.c -- use a mutex to control
access to DEV_MEM.  OK'ed by matt.
2008-11-08 10:18:10 +00:00
cliff
583e22f8aa remove (unused) KERNEL_BASE_EXT option 2008-11-08 09:42:46 +00:00
cliff
15bc29b700 on-chip gemini uart is a COM_TYPE_16550_NOERS 2008-11-08 09:39:27 +00:00
cliff
56126c94ae use correct OBIOCF_*_DEFAULT locator values 2008-11-08 09:37:13 +00:00
dyoung
8e791deb98 To finish the device_t/softc split, convert some casts to struct
auich_softc * to device_private() calls.
2008-11-08 00:26:35 +00:00
cegger
83fde857cc buildfix: use PRIxPTR rather %lx. 2008-11-07 19:50:00 +00:00
he
761f38db5c In wsdisplay_kbdholdscreen(), just return if sc->sc_focus is NULL.
There is apparently no guarantee sc->sc_focus will be non-NULL,
as proved by a crash I experienced earlier today (admittedly on 4.0.1).
2008-11-07 19:33:13 +00:00
rafal
dd332648f3 Let hpcboot load kernels linked with PA != VA; use the PA addresses specified
instead of the VA addresses.  Should make no difference for kernels where the
physical and virtual addresses are the same.
2008-11-07 16:28:26 +00:00
rafal
dba38eab03 Explicitly set LOADADDR of bss/sbss so naiive bootloaders (cough, hpcboot)
can deal with the BSS/SBSS of kernels linked with PA != VA.
2008-11-07 16:23:58 +00:00