summaryrefslogtreecommitdiffstats
path: root/sys
Commit message (Collapse)AuthorAgeFilesLines
* Initialize 'blocked' to NULL. I think this was a real problem, but Ipeter2003-07-231-0/+1
| | | | | am not sure about that. The lack of -Werror and the inline noise hid this for a while.
* Stop GCC from whining when people use a 16 bit port number for inb() and outb()phk2003-07-231-2/+2
|
* Turn -Werror back on.peter2003-07-231-1/+1
|
* ARGH. I *knew* I'd eventually accidently commit this. Change 5.1-XP backpeter2003-07-231-1/+1
| | | | to 5.1-CURRENT.
* Provide a knob for raising the inline instruction estimate threshold andpeter2003-07-233-1/+7
| | | | | | | set an initial value. This is aimed at getting us closer to being able to turn -Werror back on and we can adjust the settings later on. Yes, we could turn off -Wno-inline instead, but that would hide the effect of gcc's bogo-estimator ignoring inline (either rightly or wrongly).
* Annotate pmap_changebit() as __always_inline. This function wasalc2003-07-232-2/+4
| | | | | | | written as a template that when inlined is specialized for the caller through constant value propagation and dead code elimination. Thus, the specialized code that is generated for pmap_clear_reference() et al. avoids several conditional branches inside of a loop.
* Use macros from apic.h to when writing to the ICR to send IPIs to startupjhb2003-07-236-30/+72
| | | | | | APs rather than magic numbers. Tested by: scottl
* Add a new macro APIC_ICRLO_RESV_MASK that contains all of the reservedjhb2003-07-234-2/+5
| | | | | | | | fields in the low 32 bits of the local APIC ICR register. Use this macro in place of APIC_RESV2_MASK when masking off existing bits from the ICR when writing to it to send an IPI. Tested by: scottl
* Undo single-intance inlining which is way above the comfort limit for GCC.phk2003-07-231-3/+2
|
* Undo agressive inlining which GCC previously wisely ignored.phk2003-07-232-152/+169
| | | | | The this eliminates 67% of the text segment (relative to respected inline requests).
* Handle the new MEDIA definitions.harti2003-07-231-0/+9
|
* Convert a lot of uma_zalloc() calls to be NOWAIT instead of WAITOK. Allharti2003-07-239-30/+48
| | | | | these may be called from contexts where we cannot sleep (callout handlers for example).
* Get rid of the zone for network interfaces. We have converted this toharti2003-07-231-7/+0
| | | | use malloc(9).
* MFi386: revision 1.1090.nyan2003-07-231-0/+10
|
* Remove _ARCH_INDIRECT, it was the glue to enable having both hardwarepeter2003-07-231-67/+0
| | | | | | and software versions of the floating point code in libm. The runtime selection was done by reading the hw.floatingpoint sysctl via __get_hw_float().
* We sloppily created an array for the high FP registers (f32-f127),marcel2003-07-231-2/+96
| | | | | | but this just created a weird inconsistency when porting gdb(1). Instead, we name each high FP register seperately, like we do for all the other registers.
* Attempt to preempt any new gcc-ism references topeter2003-07-221-0/+6
| | | | | | __attribute__((__always_inline__)) by adding an __always_inline macro (used like __dead2 etc). __inline_damnit has also been suggested but we have a precedent of keeping the names similar so they are easier to find.
* There is strong reason to believe that gcc 4 will also supportimp2003-07-221-1/+1
| | | | __attribute__((__nonnull__(x)), assume it so.
* Remove all but one of the inlines here, this reduces the code size byphk2003-07-221-10/+9
| | | | 2032 bytes and has no measurable impact on performance.
* add IPSEC_FILTERGIF suport for FAST_IPSECsam2003-07-221-0/+7
| | | | | | PR: kern/51922 Submitted by: Eric Masson <e-masson@kisoft-services.com> MFC after: 1 week
* De-inline some functions. It doesn't gain us anything and bloatsmux2003-07-221-13/+13
| | | | | | code size by 3616 bytes. Furthemore, it was previously ignored by GCC. While I'm at it, fix some bogus comments.
* Allocate network interfaces from malloc() instead of using a zone.harti2003-07-221-6/+6
| | | | | Usually one needs only a couple of them so using a zone is waste of memory (esp. on multi-cpu systems).
* Remove the zone limits for all the zones used in the ATM code.harti2003-07-2213-21/+14
| | | | | | | | | | | These were a left over from when the private memory pools were converted to use uma zones. The limit of UMA zones, however, works differently. When a zone is limited to only one or two pages than, on multi-cpu systems, processes can get stuck on the zonelimit, because all remaining free items are in caches of other CPUs. Also add rudimentary error handling in some places (panic) when a zone cannot be created.
* EISA_SLOTS is mandantory to get opt_eisa.hticso2003-07-222-2/+2
| | | | Put it into MI files.
* relocate eisa into MI files.ticso2003-07-223-2/+1
| | | | Suggested by: jhb
* Revert stuff which accidentally ended up in the previous commit.phk2003-07-229-46/+15
|
* Don't attempt to inline large functions mb_alloc() and mb_free(),phk2003-07-2210-17/+48
| | | | | | it more than doubles the text size of this file. GCC has wisely ignored us on this previously
* Move inlined function ie_ack() up to before first use.phk2003-07-221-8/+8
|
* Don't inline very large functions.phk2003-07-221-1/+1
| | | | Gcc has silently not been doing this for a long time.
* Don't inline ridiculously very large functions.phk2003-07-221-4/+4
| | | | | Compared to the contents of these functions, an extra function call is nano-peanuts.
* Don't complain about inlines for genassymphk2003-07-221-1/+1
|
* When padding an mbuf chain to have a length that is a multipleharti2003-07-221-0/+1
| | | | | | of 48 bytes for AAL0, we also need to update the packet header. Spotted by: Anil Madhavapeddy <anil@recoil.org>
* Initiate de-orbit burn for fpu-less operation. 386+387 is stillpeter2003-07-2258-14046/+6
| | | | | theoretically supportable, but you'd really be happier with FreeBSD 2.1.8 on it.
* Fix a 64 bit bug lost in the inline warning noise. sizeof is not alwayspeter2003-07-221-1/+1
| | | | an int. Do not printf() it as though it were.
* swp_pager_hash() was called before it was instantiated inline. This madepeter2003-07-221-29/+29
| | | | gcc (quite rightly) unhappy. Move it earlier.
* Go back to 64 bit precision for fadd/fsub/fsqrt etc. This is because onpeter2003-07-222-16/+18
| | | | | | | AMD64, gcc (and the ABI) expects the x87 unit to be running in 80/64 mode (not 64/53) so that it can use it for 'long double' operations. It takes the expected precision differences into account when generating code.
* Extend the machine/ieeefp.h that was inherited from i386 to supportpeter2003-07-221-55/+165
| | | | | | | | | the SSE mxcsr register as well. Since gcc will intermix SSE2 and x87 FP code, the fpsetround() etc mode had better be the same. There are hooks to enable these inlines to be instantiated inside libc for non-gcc or C++ callers. (g++ doesn't like the inlines that tried to extract an integer and convert it to an enum).
* Add several vendor, API and media definitions. This has beenharti2003-07-221-5/+27
| | | | forgotten in the previous commit to harp and should unbreak world.
* Remove __nonnull() on the second argument of strto[u]l() which I usedphk2003-07-221-2/+2
| | | | | | to test that the warning actually was emitted. Spotted by: scottl
* Erm, my previous commit was wrong and sis_tick() was only called each timecognet2003-07-222-1/+6
| | | | | | | | | | sis_ioctl() was called, so one had to use ifconfig each time the cable got plugged in to be able to use the connection. Do it a better way now, add a "in_tick" field in the softc structure, call timeout() in sis_tick() and don't call it in sis_init() if in_tick is non-zero. Reported by: Landmark Networks Pointy hat to: cognet
* Fix apparent typo in previous commit.silby2003-07-221-1/+1
|
* o Add monitor mode support. This tested fine with prism cards but may requiresam2003-07-211-2/+29
| | | | | | | | newer lucent/hermes firmware than indicated (investigating). I'm committing this now since it shouldn't hurt anything. o Vaguely related, add bogus frame length check from netbsd. Obtained from: netbsd
* add safe driver until we can verify it as machine-independentsam2003-07-211-0/+10
|
* safenet driver config gluesam2003-07-212-0/+6
| | | | Sponsored by: Global Technology Associates, Inc.
* hook safe driver to the buildsam2003-07-211-0/+1
|
* SafeNet crypto driversam2003-07-211-0/+38
| | | | | Sponsored by: Global Technology Associates, Inc. MFC after: 1 day
* SafeNet crypto driver (supports only RNG and symmetric ops; no PK support yet)sam2003-07-213-0/+2876
| | | | | Sponsored by: Global Technology Associates, Inc. MFC after: 1 day
* Commit Ian Dowse's workaround for acpi resume resetting after thepeter2003-07-211-0/+8
| | | | | | | | | | LAZY_SWITCH changes. He pointed out the acpi code sets up an identity mapping in the current vmspace and that got messed up by the %cr3 being out of sync with the current page directory. As a workaround, restore %cr3 across the sleep/resume. A more complete fix would be to undo the lazy state and clear the pm_active bit from the borrowed pmap, but this works and people are currently hurting. I'll clean this up. This is mostly Ian's patch, plus a PAE tweak from me.
* Enable eisa support on alpha.ticso2003-07-213-0/+3
| | | | | We need eisa bridge support on some models, which now depends on having eisa in the kernel.
* Add a new macro __nonnull(x) to use the new GCC33 attribute which checksphk2003-07-212-16/+22
| | | | | | | | | that an argument is not a NULL pointer. Apply various obvious places. I belive __printf*() implies __nonnull() so it is not needed on functions already tagged that way.
OpenPOWER on IntegriCloud