Commit Graph

823 Commits

Author SHA1 Message Date
Peter Grehan
1a6f777efe Always isync after a mtmsr. While perhaps not strictly necessary for PSL_EE
bit banging according to the OEA, it's better to be conservative than
having to continually audit uses of this inline.
2004-08-07 00:20:00 +00:00
Peter Grehan
2184ddd1f7 In pmap_page_protect, clear the vm page's PG_WRITEABLE flag if
downgrading to read-only. Found by triggering the KASSERT in
vm_pageout_flush().
2004-08-05 12:44:12 +00:00
Alan Cox
684a62b7bf - Push down the acquisition and release of Giant into pmap_enter_quick()
on those architectures without pmap locking.
 - Eliminate the acquisition and release of Giant in vm_map_pmap_enter().
2004-08-04 22:03:16 +00:00
Mark Murray
d23a262fc5 Making a loadable null.ko for /dev/(null|zero) proved rather
unpopular, so remove this (mis)feature.

Encouragement provided by:	jhb (and others)
2004-08-03 19:24:54 +00:00
Maxime Henrion
9f1b87f106 Instead of calling ia32_pause() conditionally on __i386__ or __amd64__
being defined, define and use a new MD macro, cpu_spinwait().  It only
expands to something on i386 and amd64, so the compiled code should be
identical.

Name of the macro found by:	jhb
Reviewed by:	jhb
2004-08-03 18:44:27 +00:00
Peter Grehan
e796443d47 Remove race condition between reading of MSR, setting md_savecrit,
and setting MSR. This was most evident with the idle proc running
with interrupts disabled and causing a lockup. Switch over to the
i386 style which does things in the right order.

debug assisted by:  gallatin, and the invaluable KTR option.
2004-08-03 04:14:55 +00:00
Suleiman Souhlal
09e7619917 Remove 'device mem' from GENERIC, which markm@ mistakingly added.
We don't have mem/kmem yet.

Approved by:	grehan (mentor)
2004-08-02 11:08:48 +00:00
Peter Grehan
016927054b Kernel traps were not being passed to trap_fatal in some
circumstances.

Spotted by:  gallatin
2004-08-02 02:37:29 +00:00
Mark Murray
8ab2f5ecc5 Break out the MI part of the /dev/[k]mem and /dev/io drivers into
their own directory and module, leaving the MD parts in the MD
area (the MD parts _are_ part of the modules). /dev/mem and /dev/io
are now loadable modules, thus taking us one step further towards
a kernel created entirely out of modules. Of course, there is nothing
preventing the kernel from having these statically compiled.
2004-08-01 11:40:54 +00:00
Alan Cox
9bb0e06861 - Push down the acquisition and release of Giant into pmap_protect() on
those architectures without pmap locking.
 - Eliminate the acquisition and release of Giant from vm_map_protect().

(Translation: mprotect(2) runs to completion without touching Giant on
alpha, amd64, i386 and ia64.)
2004-07-30 20:38:30 +00:00
Suleiman Souhlal
a6e340aabe Add comment explaining struct reg and struct fpreg must match the trapframe.
Approved by:	grehan (mentor)
2004-07-29 13:39:27 +00:00
Suleiman Souhlal
009a0e433b Implement MD parts of ptrace.
Approved by:	grehan (mentor)
2004-07-29 13:34:50 +00:00
Robert Watson
1a8cfbc450 Pass a thread argument into cpu_critical_{enter,exit}() rather than
dereference curthread.  It is called only from critical_{enter,exit}(),
which already dereferences curthread.  This doesn't seem to affect SMP
performance in my benchmarks, but improves MySQL transaction throughput
by about 1% on UP on my Xeon.

Head nodding:	jhb, bmilekic
2004-07-27 16:41:01 +00:00
Peter Grehan
e60a36f564 Properly implement kdb_cpu_{set|clear}_singlestep to allow DDB to
continue from breakpoints.
2004-07-27 07:06:20 +00:00
Peter Grehan
74fa1336d6 Make sure icache is sync'd whenever memory is touched. It may
be more optimal to override the BKPT_WRITE macro, but DDB performance
isn't really a goal at this stage...
2004-07-27 07:04:58 +00:00
Peter Grehan
1f1b01c0a5 Save DAR/DSISR in DDB regsave area when stack overflow detected. It's
hard to work out where the problem was without these.
2004-07-27 03:46:34 +00:00
Peter Grehan
5dd954f3de Improve boot-time debugging with DDB by extracting the ksym start/end
values from the loader.
2004-07-27 03:41:34 +00:00
Alan Cox
ab50a26230 Implement the protection check required by the pmap_extract_and_hold()
specification.

Reviewed and tested by:	grehan@
2004-07-26 18:10:10 +00:00
Andrew Gallatin
d593f6eda9 Let ddb know powerpc is big endian so as to make ddb output
human readable.

Obtained from: sparc64/include/db_machdep.h
2004-07-23 14:45:15 +00:00
Peter Grehan
2a1c4385c3 Detect kernel stack excursion into guard pages. Drop into KDB
with a wired stack if this is found.

Mostly obtained from:  NetBSD
2004-07-23 05:33:24 +00:00
Peter Grehan
a76b77653c Bring KDB stack size into line with thread stack size (4 pages). 2004-07-23 05:31:14 +00:00
Peter Grehan
c06a377abc Allow DSI exceptions to invoke DDB. 2004-07-23 05:27:17 +00:00
Peter Grehan
163fac2c58 The ADDR16 relocations were assuming that non-local symbols had an
addend of 0. This isn't correct, and was quite easy to break by
referring to the address of an element within a structure.

 However, fixing this exposed the fact that symbol lookups for
local variables were returning the base of the section they
were contained in. This case is detected by comparing the return
value from elf_lookup() to the relocbase+addend value: if it is
lesser, but greater than relocbase, then relocbase+addend is
taken to be the authoritative value.

bug reported by:  gallatin
2004-07-23 00:46:05 +00:00
Peter Grehan
bddfaa895c Update the callframe structure to leave space for the frame pointer
and saved link register as per the ABI call sequence. Update code
that uses this (fork_trampoline etc) to use the correct genassym'd
offsets.

 This fixes the 'invalid LR' message when backtracing kernel
threads in DDB.
2004-07-22 01:28:51 +00:00
Andrew Gallatin
82ab2f48a4 Make this compile: add sys/module.h and KDBify. 2004-07-22 00:54:01 +00:00
Marcel Moolenaar
fd32d93b97 Unify db_stack_trace_cmd(). All it did was look up the thread given
the thread ID and call db_trace_thread().
Since arm has all the logic in db_stack_trace_cmd(), rename the
new DB_COMMAND function to db_stack_trace to avoid conflicts on
arm.
While here, have db_stack_trace parse its own arguments so that
we can use a more natural radix for IDs. If the ID is not a thread
ID, or more precisely when no thread exists with the ID, try if
there's a process with that ID and return the first thread in it.
This makes it easier to print stack traces from the ps output.

requested by: rwatson@
tested on: amd64, i386, ia64
2004-07-21 05:07:09 +00:00
Peter Grehan
8c18dab95e elf_cpu_load_file no longer has an __unused variable. Also, don't
bother syncing the icache for the special case of the kernel (id == 1),
since the loader has already done this.

__unused use reported by:  gallatin
2004-07-20 02:40:57 +00:00
Peter Grehan
1c4ba0be13 Properly obey PPC context synchronization rules when modifying
the address translation bits of the MSR. This fixes the boot-time
panic reported by Drew Gallatin.
2004-07-20 02:22:36 +00:00
Andrew Gallatin
df11547f5d Fix printing of long doubles to match the size that
gcc is using.  This fixes devstat consumers (like vmstat, iostat,
systat) so they don't print crazy zillion digit numbers for
disk transfers and bandwidth.

According to gcc, long doubles are 64-bits, rather than 128 bits
like the SVR4 ABI spec wants them to be..  Note that MacOSX also treats
long doubles as 64-bits, and not 128 bits, so we are in good company.

Reviewed by: das
Approved by: grehan
2004-07-19 23:56:07 +00:00
David Schultz
479f8d2214 Make FLT_ROUNDS correctly reflect the dynamic rounding mode. 2004-07-19 08:17:25 +00:00
Peter Grehan
3327cde241 Use the version field to identify the partial context used by
KSE process-scope threads.
2004-07-19 07:21:46 +00:00
Peter Grehan
8653f72f6e Empty GENERIC.hints file needed by make release.
Noticed by:  Suleiman Souhlal <refugee@segfaulted.com>
2004-07-19 02:08:22 +00:00
Maxim Konovalov
aa355a2679 In -CURRENT pseudo devices are not statically assigned at compile time,
remove a stale comment.

PR:		kern/62285
2004-07-18 09:03:12 +00:00
Peter Grehan
b7ef1c19e4 Resurrect kld support. Support ADDR16_HA/LA relocations, and sync
the icache on module load. Requires "-mlongcall" support, in gcc >= 3.3
but needs a bugfix to support gcc arith builtins.
2004-07-17 07:26:32 +00:00
Alan Cox
3d2e54c317 Push down the acquisition and release of the page queues lock into
pmap_protect() and pmap_remove().  In general, they require the lock in
order to modify a page's pv list or flags.  In some cases, however,
pmap_protect() can avoid acquiring the lock.
2004-07-15 18:00:43 +00:00
David Xu
53dbf30349 Add ptrace_clear_single_step(), alpha already has it for years, the function
will be used by ptrace to clear a thread's single step state.
2004-07-13 07:22:56 +00:00
Peter Grehan
441e42eaf4 Rename low-level code ddb -> db. Use KDB instead of DDB.
Fix bug in setup of stack frame where 8 bytes wasn't being
saved for the callee's frame pointer and saved LR.
2004-07-12 22:32:08 +00:00
Peter Grehan
b188ee2269 Bring into KDB new order. 2004-07-12 22:26:20 +00:00
Peter Grehan
20d21fe9ef - DDB -> KDB, with kdb routines
- ddb -> db for low-level trapcode
- implement makectx. I think it only matters that the stack is setup
  correctly.
- bring over ddb_trap_glue and rename to db_trap_glue
2004-07-12 22:25:09 +00:00
Peter Grehan
def828503c No need for ddb option. Never a need for ipkdb option. 2004-07-12 22:22:53 +00:00
Peter Grehan
e3c25f5d33 Catch up with gratuitous ddb -> db renaming 2004-07-12 22:22:09 +00:00
Peter Grehan
d8570880e8 Bring into line with KDB. Bring in NetBSD updates for backtrace routine,
although it really needs a decent re-work.
2004-07-12 22:21:34 +00:00
Peter Grehan
a136236e5c Remove unused NetBSD code. Bring mem r/w routines into here in line
with sparc64, although keep the size deref checks: they are useful
when accessing PCI space where some devices may not implement byte
access.
2004-07-12 22:20:01 +00:00
Peter Grehan
4539337da0 Gratuitous namechange to avoid low-level association with ddb. 2004-07-12 22:18:02 +00:00
Peter Grehan
abdc8ff1bc Add prototype for KDB's makectx routine 2004-07-12 22:17:20 +00:00
Peter Grehan
5f40873a1f Remove old NetBSD-derived unused code and stuff that is now obsolete
due to KDB.
2004-07-12 22:16:50 +00:00
Peter Grehan
9917bd81ed DDB -> KDB, and rename low-level trap handler to avoid name conflict. 2004-07-12 22:16:04 +00:00
Peter Grehan
f4ec5aece7 kdb.h for PowerPC. Stubs for now. 2004-07-12 22:15:03 +00:00
Peter Grehan
24c1c57156 Add new KDB option, and also drop in long-held fxp/dc eth drivers. 2004-07-12 22:14:21 +00:00
Alan Cox
eba90ac75f pmap_remove_pages() must not remove wired mappings. Since
pmap_remove_pages() is an optimization, its implementation is optional.

Discussed with:	grehan
2004-07-12 04:40:26 +00:00