ƒ«E192226 187 139 138 200 120 683 256 125 124 132 123 113 121 142 227 134 134 329 163 106 160 270 136 172 111 534 598 527 116 120 329 606 204 178 124 254 310 272 251 162 188 127 135 155 107 289 124 117 106 131 110 141 671 357 331 122 604 840 118 195 137 242 112 139 121 126 162 128 788 163 309 184 104 147 182 157 165 163 267 376 488 496 237 240 1388 213 247 135 180 233 123 196 132 164 109 253 100 169 1141 183 940 2358 1098 115 130 119 237 1572 555 361 572 151 408 107 127 140 225 186 467 424 746 140 419 167 131 185 369 315 138 341 146 149 130 124 1017 146 128 118 118 427 142 135 145 118 679 542 498 1605 371 135 154 137 168 198 500 1331 692 166 143 257 196 156 332 170 266 206 171 154 216 183 151 169 238 276 106 276 161 130 113 254 121 248 138 304 267 235 150 198 K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-16T22:07:41.471138Z K 7 svn:log V 45 correct range in comment pointed out by alc END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-16T22:08:00.773565Z K 7 svn:log V 44 correct range in comment pointed out by alc END K 10 svn:author V 8 brueffer K 8 svn:date V 27 2009-05-16T22:18:58.956548Z K 7 svn:log V 102 Drop clause 3 of the license as per rev. 1.35 from OpenBSD. Obtained from: OpenBSD MFC after: 3 days END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-16T22:20:42.639992Z K 7 svn:log V 26 initialize pcpu rm queues END K 10 svn:author V 5 dougb K 8 svn:date V 27 2009-05-16T22:22:31.373791Z K 7 svn:log V 588 Enhance the safety of the -U option: 1. In several places make sure that the mtree database is not empty using the -s argument to test instead of -f. (I thought I'd already changed this, but obviously not.) 2. When deriving the list of changed files use a colon-delimited list. 3. If the list of changed files is empty, unset the variable, and test for a non-empty CHANGED in diff_loop() before entering that routine. Enhance the speed of the -U option by using an internal case statement instead of echoing the variable to an external grep for every file. Fix indentation on one line. END K 10 svn:author V 8 rmacklem K 8 svn:date V 27 2009-05-16T22:31:38.015380Z K 7 svn:log V 158 Changed sys/fs/nfs_clbio.c in the same way Alan Cox changed sys/nfsclient/nfs_bio.c for r192134, so that the sources stay in sync. Approved by: kib (mentor) END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-16T22:54:16.432392Z K 7 svn:log V 31 restore osd support to prisons END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-16T23:33:07.321412Z K 7 svn:log V 30 deregister osd on thread exit END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-16T23:56:45.354650Z K 7 svn:log V 38 enable adaptive spinning on zfs locks END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-16T23:57:47.601011Z K 7 svn:log V 29 MFC adaptive spinning change END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T00:01:57.053013Z K 7 svn:log V 19 enable adaptive sx END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T01:31:28.084122Z K 7 svn:log V 27 SAVESTART implies SAVENAME END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T01:32:47.462873Z K 7 svn:log V 48 generalize assert as savestart implies savename END K 10 svn:author V 5 brian K 8 svn:date V 27 2009-05-17T04:00:43.896863Z K 7 svn:log V 132 Add a -d option to ps to display descendant info with the output. This is similar to linux's -H (or -f) switch. MFC after: 3 weeks END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T04:04:25.597636Z K 7 svn:log V 40 set createtxg prop name PR: bin/130105 END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T04:05:27.007714Z K 7 svn:log V 40 set createtxg prop name PR: bin/130105 END K 10 svn:author V 5 brian K 8 svn:date V 27 2009-05-17T04:34:14.107197Z K 7 svn:log V 234 When finding processes, ignore ourself and our ancestors. It is almost always surprising when you kill a 'sh -c ...' ancestor or when you kill yourself when using -f. Add a -a switch for backwards compatibility. MFC after: 3 weeks END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T05:54:25.445815Z K 7 svn:log V 69 add zfs oid to bsnmpd PR: bin/129360 Submitted by: Ulrich Spoerlein END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T05:55:13.933149Z K 7 svn:log V 12 add zfs oid END K 10 svn:author V 3 alc K 8 svn:date V 27 2009-05-17T06:45:30.589271Z K 7 svn:log V 68 Merge r191964: Eliminate a case of unnecessary page queues locking. END K 10 svn:author V 6 danger K 8 svn:date V 27 2009-05-17T08:25:02.442082Z K 7 svn:log V 174 - do not create and mount new file systems on top of the old ones on every invocation of this script once we already have one (in case tmpmfs="YES"). Reviewed by: dougb END K 10 svn:author V 8 brueffer K 8 svn:date V 27 2009-05-17T10:58:50.593444Z K 7 svn:log V 39 Improve the accf_dns_load description. END K 10 svn:author V 7 fabient K 8 svn:date V 27 2009-05-17T11:45:37.576186Z K 7 svn:log V 76 MFC 190395: Allow compile from c++ for libpmc Reviewed by: jkoshy (mentor) END K 10 svn:author V 5 trasz K 8 svn:date V 27 2009-05-17T12:21:11.693801Z K 7 svn:log V 17 Fix the example. END K 10 svn:author V 2 ed K 8 svn:date V 27 2009-05-17T12:30:25.839707Z K 7 svn:log V 442 Several cleanups to tty_info(), better known as Ctrl-T. - Only pick up PROC_LOCK once, which means we can drop the PGRP_LOCK right after picking up PROC_LOCK for the first time. - Print the process real time, making it consistent with tools like time(1). - Use `p' and `td' to reference the process/thread we are going to print. Only use pick-variables inside the loops. We already did this for the threads, but not the processes. END K 10 svn:author V 3 bms K 8 svn:date V 27 2009-05-17T15:42:41.277801Z K 7 svn:log V 505 MFC rev 191654: Fix an obvious bug in getsourcefilter()'s use of struct __msfilterreq; the kernel will return in msfr_nsrcs the number of source filters in-mode for a given multicast group. However, the filters themselves were never copied out, as the libc function clobbers this field with zero, causing the kernel to assume the provided vector of struct sockaddr_storage has zero length. This bug would only affect users of SSM multicast, which is shimmed in 7.x. Picked up during mtest(8) refactoring. END K 10 svn:author V 2 ed K 8 svn:date V 27 2009-05-17T16:17:48.380784Z K 7 svn:log V 435 Print an extra newline when not at the first column already. This makes siginfo output look a lot better when pressing it the first time when in sh(1), for example: $ load: 0.00 cmd: sh 1945 [ttyin] 3.94r 0.00u 0.00s 0% 1960k load: 0.00 cmd: sh 1945 [ttyin] 4.19r 0.00u 0.00s 0% 1960k will now become: $ load: 0.00 cmd: sh 1945 [ttyin] 3.94r 0.00u 0.00s 0% 1960k load: 0.00 cmd: sh 1945 [ttyin] 4.19r 0.00u 0.00s 0% 1960k END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T17:49:57.290643Z K 7 svn:log V 22 enable forced unmount END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T17:52:35.765893Z K 7 svn:log V 26 add ALT_BREAK_TO_DEBUGGER END K 10 svn:author V 8 rmacklem K 8 svn:date V 27 2009-05-17T17:54:01.772706Z K 7 svn:log V 231 Added a SYSCTL to sys/fs/nfsserver/nfs_nfsdport.c so that the value of nfsrv_dolocallocks can be changed via sysctl. I also added some non-empty descriptor strings and reformatted some overly long lines. Approved by: kib (mentor) END K 10 svn:author V 8 rmacklem K 8 svn:date V 27 2009-05-17T19:33:48.224428Z K 7 svn:log V 508 Fix the acquisition of local locks via VOP_ADVLOCK() by the experimental nfsv4 server. It was setting the a_id argument to a fixed value, but that wasn't sufficient for FreeBSD8. Instead, set l_pid and l_sysid to 0 plus set the F_REMOTE flag to indicate that these fields are used to check for same lock owner. Since, for NFSv4, a lockowner is a ClientID plus an up to 1024byte name, it can't be put in l_sysid easily. I also renamed the p variable to td, since it's a thread ptr. Approved by: kib (mentor) END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-17T19:46:50.436049Z K 7 svn:log V 111 fix 11a channel use; mark OFDM operation correctly Submitted by: Lucius Windschuh END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-17T19:51:08.502437Z K 7 svn:log V 86 add TRENDnet TEW-504UB/EU Submitted by: Lucius Windschuh END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T20:16:38.109906Z K 7 svn:log V 30 define compat shim for M_ZERO END K 10 svn:author V 3 alc K 8 svn:date V 27 2009-05-17T20:26:00.654015Z K 7 svn:log V 161 Introduce vfs_bio_set_valid() and use it from ffs_realloccg(). This eliminates the misuse of vfs_bio_clrbuf() by ffs_realloccg(). In collaboration with: tegge END K 10 svn:author V 3 alc K 8 svn:date V 27 2009-05-17T20:40:41.735827Z K 7 svn:log V 217 Eliminate a pointless call to pmap_clear_reference() from vm_pageout_scan(). If the page belongs to an object with a reference count of zero, then it can't have any managed mappings on which to clear a reference bit. END K 10 svn:author V 2 bz K 8 svn:date V 27 2009-05-17T20:53:10.044446Z K 7 svn:log V 180 Unbreak options VIMAGE builds, in a followup to r192011 which did not introduce INIT_VNET_NET() initializers necessary for accessing V_loif. Submitted by: zec Reviewed by: julian END K 10 svn:author V 2 bz K 8 svn:date V 27 2009-05-17T20:55:33.727869Z K 7 svn:log V 159 Put the structs under #ifndef VIMAGE_GLOBALS as some constants are not defined for VIMAGE_GLOBALS and thus broke the build for that option. Reported by: csjp END K 10 svn:author V 2 bz K 8 svn:date V 27 2009-05-17T20:59:09.956268Z K 7 svn:log V 71 Add a missing INIT_VNET_NET() to get VIMAGE closer to full LINT again. END K 10 svn:author V 8 brueffer K 8 svn:date V 27 2009-05-17T21:28:37.227064Z K 7 svn:log V 91 Document sbuf_new_auto(). While here, add a missing `-' in phk's name. MFC after: 3 days END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T22:17:47.603189Z K 7 svn:log V 33 include fcntl.h for namei change END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T22:46:23.600914Z K 7 svn:log V 41 create dedicated 7.2 branch with zfs v13 END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T23:14:26.774409Z K 7 svn:log V 61 nothing references ni_dirfd remove to avoid breaking the ABI END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T23:17:56.793239Z K 7 svn:log V 13 fix comments END K 10 svn:author V 3 alc K 8 svn:date V 27 2009-05-17T23:25:53.297835Z K 7 svn:log V 196 Several changes to vfs_bio_clrbuf(): Provide a more descriptive comment. Eliminate dead code. The page cannot possibly have PG_ZERO set. Eliminate unnecessary blank lines. Reviewed by: tegge END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T23:30:13.542237Z K 7 svn:log V 30 MFC sys portion of update v13 END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T23:31:57.892843Z K 7 svn:log V 23 MFC user bits from v13 END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T23:32:24.483927Z K 7 svn:log V 12 add zfs oid END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T23:40:59.969964Z K 7 svn:log V 37 add self to list to avoid size limit END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T23:42:38.753030Z K 7 svn:log V 16 MFC updated ZFS END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-17T23:44:10.841774Z K 7 svn:log V 47 rename branch to fcs for "fast content server" END K 10 svn:author V 3 imp K 8 svn:date V 27 2009-05-18T01:00:11.743330Z K 7 svn:log V 578 Turns out that my BCM4318 has a PCI ID of 0x4319, which lead me to believe it was a BCM4319. However, it is the a/b/g variation of the BCM4318. The chip itself is labelled BCM4318EKFBG, and the board is BCM94318MKABG. Paradox's patch includes the type of 802.11 wireless for each card, but changes all the names (I don't think the latter is quite right). Import that part of the patch, but keep the current set of BCM names (with a minor tweak for the 4306 ones). I'll need to verify them via some other means. Obtained from: http://paradox.lissyara.su/bwi.diff (partially) END K 10 svn:author V 2 bz K 8 svn:date V 27 2009-05-18T01:05:09.057442Z K 7 svn:log V 265 tbr_timeout() is a timer driven function[1]. While the previous commit made LINT happy this does the proper looping over all vnets as we are only called `globally' and not once per vnet instance. Reported by: zec [1] Missed by: bz [1] in r192264 Reviewed by: zec END K 10 svn:author V 3 imp K 8 svn:date V 27 2009-05-18T01:07:38.582981Z K 7 svn:log V 238 Correct types of PHY, per http://bcm-specs.sipsolutions.net/PHYVersioning # Note: The driver doesn't support either these PHY types, so this is # effectively a nop. Submitted by: "ddk" Obtained from: http://paradox.lissyara.su/bwi.diff END K 10 svn:author V 5 brian K 8 svn:date V 27 2009-05-18T01:45:52.628179Z K 7 svn:log V 28 Remove redundant whitespace END K 10 svn:author V 7 delphij K 8 svn:date V 27 2009-05-18T01:51:52.626704Z K 7 svn:log V 507 DMA synchronization fixes: - In bce_rx_intr(), use BUS_DMASYNC_POSTREAD instead of BUS_DMASYNC_POSTWRITE, as we want to "read" from the rx page chain pages. - Document why we need to do PREWRITE after we have updated the rx page chain pages. - In bce_intr(), use BUS_DMASYNC_POSTREAD and BUS_DMASYNC_PREREAD when before and after CPU "reading" the status block. - Adjust some nearby style mismatches/etc. Pointed out by: yongari Approved by: davidch (no objection) but bugs are mine :) END K 10 svn:author V 6 qingli K 8 svn:date V 27 2009-05-18T02:25:45.495879Z K 7 svn:log V 744 This patch resolves the following issues: -- A routing socket message is not generated when an IPv6 address is either inserted or deleted from an interface. The missing routing message problem was discovered by Randall Stewart and Michael Tuxen during SCTP testing. -- Previously when an IPv6 address is configured on an interface, if the prefix length is /128, then a host route is instaleld in the kernel for this address. But this host route is not deleted when that IPv6 address is removed from the interface. -- Routes to the link-local all-nodes multicast address and the interface-local all-nodes multicast address are not removed when the last IPv6 address is removed from an interface. Reviewed by: bz, gnn END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T02:40:11.006234Z K 7 svn:log V 24 merge fast_net backport END K 10 svn:author V 7 dchagin K 8 svn:date V 27 2009-05-18T04:07:46.242827Z K 7 svn:log V 99 Implement MSG_CMSG_CLOEXEC flag for linux_recvmsg(). Approved by: kib (mentor) MFC after: 1 month END K 10 svn:author V 6 adrian K 8 svn:date V 27 2009-05-18T04:50:31.755822Z K 7 svn:log V 42 Disable some un-needed console debugging. END K 10 svn:author V 6 adrian K 8 svn:date V 27 2009-05-18T04:56:37.014273Z K 7 svn:log V 146 The merge in r189699 reverted part of the work done in a previous commit (r188036.) Re-revert that change so the Xen networking functions again. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T05:47:58.623644Z K 7 svn:log V 18 add local changes END K 10 svn:author V 7 yongari K 8 svn:date V 27 2009-05-18T06:02:54.213846Z K 7 svn:log V 43 Use ANSI C declarations for all functions. END K 10 svn:author V 7 yongari K 8 svn:date V 27 2009-05-18T06:05:50.223236Z K 7 svn:log V 25 Remove register keyword. END K 10 svn:author V 7 yongari K 8 svn:date V 27 2009-05-18T06:09:10.390016Z K 7 svn:log V 30 Remove trailling whitespaces. END K 10 svn:author V 7 yongari K 8 svn:date V 27 2009-05-18T06:13:56.721167Z K 7 svn:log V 66 Remove return statement at the end of functions that return void. END K 10 svn:author V 7 yongari K 8 svn:date V 27 2009-05-18T06:19:17.079849Z K 7 svn:log V 32 style(9) - space after keyword. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T06:20:37.602537Z K 7 svn:log V 693 merge HEAD updates 191028 - use a shared lock for reads - remove stale comment 191033 - update bufring for ALTQ 191080 Extend route command: - add show as alias for get - add weights to allow mpath to do more than equal cost - add sticky / nostick to disable / re-enable per-connection load balancing This adds a field to rt_metrics_lite so network bits of world will need to be re-built. 191117 add an llentry to struct route{_in6} to allow it to be passed around with the rtentry 191120 make LLTABLE visible to netinet 191126 191148 Change if_output to take a struct route as its fourth argument in order to allow passing a cached struct llentry * down to L2 END K 10 svn:author V 7 yongari K 8 svn:date V 27 2009-05-18T06:32:38.483394Z K 7 svn:log V 67 s/u_int8_t/uint8_t/g s/u_int16_t/uint16_t/g s/u_int32_t/uint32_t/g END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T06:46:34.714332Z K 7 svn:log V 214 Import changes from HEAD 191038 191154 add utility routine for updating an struct llentry * 191158 191159 191160 191161 191162 191221 191255 191257 191258 191259 191324 191440 191441 191442 191603 191611 191612 END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T06:54:53.063302Z K 7 svn:log V 90 merge fixes 191605 191606 191607 191608 191609 191610 191899 192009 192011 192085 192282 END K 10 svn:author V 7 yongari K 8 svn:date V 27 2009-05-18T07:04:03.617112Z K 7 svn:log V 9 style(9) END K 10 svn:author V 7 yongari K 8 svn:date V 27 2009-05-18T07:10:48.259350Z K 7 svn:log V 51 Consistently use tab characters instead of spaces. END K 10 svn:author V 7 yongari K 8 svn:date V 27 2009-05-18T07:13:42.410912Z K 7 svn:log V 86 Remove extra parenthesis used in macros. These macros are not used in driver, though. END K 10 svn:author V 3 alc K 8 svn:date V 27 2009-05-18T07:24:06.598956Z K 7 svn:log V 65 MFC r180308 Enable the creation of a kmem map larger than 4GB. END K 10 svn:author V 3 kib K 8 svn:date V 27 2009-05-18T08:41:59.916188Z K 7 svn:log V 73 MFC r192151: Initialize f_vnode before calling d_fdopen() cdevsw method. END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-18T10:33:23.170351Z K 7 svn:log V 67 Garbage collect unused NETISR_{ATM,NETGRAPH,PPP} netisr constants. END K 10 svn:author V 8 brueffer K 8 svn:date V 27 2009-05-18T11:23:15.918665Z K 7 svn:log V 169 Remove license clauses 3 and 4 as per rev. 1.2 from NetBSD. Also, FreeBSD 8.0 will be the first release including this driver. (1) PR: 134093 (1) Submitted by: gavin END K 10 svn:author V 2 ed K 8 svn:date V 27 2009-05-18T12:03:43.469259Z K 7 svn:log V 284 Mark the clock sysctls as MPSAFE. These sysctls don't need any form of locking. At least cp_times is used by powerd very often, which means I get 50% less calls to non-MPSAFE sysctls on my system. The other 50% is consumed by dev.cpu.0.freq, but this seems to need Giant for Newbus. END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-18T14:02:55.706050Z K 7 svn:log V 391 Binding interrupts to a CPU consists of two parts: setting up CPU affinity for the interrupt thread, and requesting that underlying hardware direct interrupts to the CPU. For software interrupt threads, implement a no-op interrupt event binder that returns success, so that the interrupt management code will just set the ithread's affinity and succeed. Reviewed by: jhb MFC after: 1 week END K 10 svn:author V 3 imp K 8 svn:date V 27 2009-05-18T15:31:26.543975Z K 7 svn:log V 403 Fix a typo from the original driver. We need to write ctrl2 into RF register 0x52, not ctrl1. This appears to be a mistake in the bcm reverse engineering page, and has been corrected there. Tracing through the code, this is more in keeping with the "documented" register. Sephe thinks it looks interesting and may be worth fixing. :) Submitted by: ddkprog at yahoo com Reviewed by: Sepherosa Ziehau END K 10 svn:author V 3 imp K 8 svn:date V 27 2009-05-18T15:46:34.647630Z K 7 svn:log V 144 Remove unnecessary comments. 11A read/write is the same: its just the classic indirect register dance. Submitted by: ddkprog at yahoo not com END K 10 svn:author V 4 emax K 8 svn:date V 27 2009-05-18T16:00:18.463379Z K 7 svn:log V 146 Teach btpand(8) to recognized Bluetooth device node names. Tested by: Daniel O'Connor MFC after: 3 days END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-18T17:08:57.265019Z K 7 svn:log V 1290 Further refinement of netisr2: - Add netisr2_getqlimit() and netisr2_setqlimit() interfaces to allow protocols to query and manipulate per-workstream queue depth limits. This is required for routing socket and IPv4 netisrs which currently offer this functionality. - Add netisr2_getqdrops() and netisr2_clearqdrops() interfaces to allow protocols to query drops across CPUs, as well as clear drop statistics. This is required for IPv4. - Use u_int64_t rather than u_int for stats. - Rather than passing in each parameter individually for netisr2_register(), netisr2_unregister(), define a public struct netisr_handler, with padding, to describe protocols. - Explicitly enumerate policies supported by netisr2, rather than deriving them from implemented function pointers; this allows multiple policies to depend on the same function pointers if desired. We implement three policies now: NETISR_POLICY_SOURCE, NETISR_POLICY_FLOW, NETISR_POLICY_CPU. - Now that we use swi's, we can acquire the netisr lock around processing runs, since the wakeup can be waited for without holding the workstream lock. - Garbage collect NWS_SWI_BOUND and manual binding with sched_bind(), use intr_event_bind() now that this is supported for software interrupt threads. END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-18T17:10:26.818126Z K 7 svn:log V 116 Update netisr2 consumers for new netisr2 interfaces -- queue limits, stats, and struct netisr_handler registration. END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-18T17:14:23.379132Z K 7 svn:log V 154 MFC: Simplify the interrupt code a bit by always including the ie_disable and ie_eoi methods and collapsing down to a single intr_event_create() routine. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T17:14:48.972965Z K 7 svn:log V 41 make adaptive spinning on SX the default END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-18T17:18:40.564935Z K 7 svn:log V 88 add bpf_track eventhandler for monitoring bpf taps attached/detached Reviewed by: csjp END K 10 svn:author V 3 kan K 8 svn:date V 27 2009-05-18T17:20:24.243906Z K 7 svn:log V 140 Remove empty files and do nto try to build them. Apparently, they are problematic for CTF users. PR: 119298 Submitted by: Julian H. Stacey END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T17:37:27.369999Z K 7 svn:log V 29 merge rt+lle caching support END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-18T17:39:12.438303Z K 7 svn:log V 103 MFC: Use __null for NULL for GNU C++ 4.x and use the same value for NULL in the kernel as in userland. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T17:42:25.997373Z K 7 svn:log V 38 make lock shared on read for sendfile END K 10 svn:author V 2 bz K 8 svn:date V 27 2009-05-18T17:48:46.540587Z K 7 svn:log V 73 Add two missing INIT_VNET_INET6(curvnet) to make VIMAGE kernels happier. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T17:57:33.991423Z K 7 svn:log V 15 glue in flowid END K 10 svn:author V 7 attilio K 8 svn:date V 27 2009-05-18T18:04:42.879770Z K 7 svn:log V 156 Fix the world breakage, on amd64, introduced by the last libthread_db commit. Reported by: many Sponsored by: Sandvine Incorporated Pointy hat to: attilio END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-18T18:12:45.837941Z K 7 svn:log V 9 fix typo END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T18:34:03.091244Z K 7 svn:log V 75 Merge dependent callout changes 171053 172025 172184 173760 173842 176013 END K 10 svn:author V 6 marcel K 8 svn:date V 27 2009-05-18T18:37:18.736938Z K 7 svn:log V 1044 Add cpu_flush_dcache() for use after non-DMA based I/O so that a possible future I-cache coherency operation can succeed. On ARM for example the L1 cache can be (is) virtually mapped, which means that any I/O that uses temporary mappings will not see the I-cache made coherent. On ia64 a similar behaviour has been observed. By flushing the D-cache, execution of binaries backed by md(4) and/or NFS work reliably. For Book-E (powerpc), execution over NFS exhibits SIGILL once in a while as well, though cpu_flush_dcache() hasn't been implemented yet. Doing an explicit D-cache flush as part of the non-DMA based I/O read operation eliminates the need to do it as part of the I-cache coherency operation itself and as such avoids pessimizing the DMA-based I/O read operations for which D-cache are already flushed/invalidated. It also allows future optimizations whereby the bcopy() followed by the D-cache flush can be integrated in a single operation, which could be implemented using on-chips DMA engines, by-passing the D-cache altogether. END K 10 svn:author V 6 marcel K 8 svn:date V 27 2009-05-18T18:44:54.487716Z K 7 svn:log V 88 Rename ia64_invalidate_icache() to ia64_sync_icache(). We're not invalidating anything. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T18:50:38.626958Z K 7 svn:log V 845 merge 177085 - Pass the priority argument from *sleep() into sleepq and down into sched_sleep(). This removes extra thread_lock() acquisition and allows the scheduler to decide what to do with the static boost. - Change the priority arguments to cv_* to match sleepq/msleep/etc. where 0 means no priority change. Catch -1 in cv_broadcastpri() and convert it to 0 for now. - Set a flag when sleeping in a way that is compatible with swapping since direct priority comparisons are meaningless now. - Add a sysctl to ule, kern.sched.static_boost, that defaults to on which controls the boost behavior. Turning it off gives better performance in some workloads but needs more investigation. - While we're modifying sleepq, change signal and broadcast to both return with the lock held as the lock was held on enter. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T18:54:43.587949Z K 7 svn:log V 2262 177491 Fix a race where timeout/untimeout could cause crashes for Giant locked code. The bug: There exists a race condition for timeout/untimeout(9) due to the way that the softclock thread dequeues timeouts. The softclock thread sets the c_func and c_arg of the callout to NULL while holding the callout lock but not Giant. It then drops the callout lock and acquires Giant. It is at this point where untimeout(9) on another cpu/thread could be called. Since c_arg and c_func are cleared, untimeout(9) does not touch the callout and returns as if the callout is canceled. The softclock then tries to acquire Giant and likely blocks due to the other cpu/thread holding it. The other cpu/thread then likely deallocates the backing store that c_arg points to and finishes working and hence drops Giant. Softclock resumes and acquires giant and calls the function with the now free'd c_arg and we have corruption/crash. The fix: We need to track curr_callout even for timeout(9) (LOCAL_ALLOC) callouts. We need to free the callout after the softclock processes it to deal with the race here. 177859 Implement per-cpu callout threads, wheels, and locks. - Move callout thread creation from kern_intr.c to kern_timeout.c - Call callout_tick() on every processor via hardclock_cpu() rather than inspecting callout internal details in kern_clock.c. - Remove callout implementation details from callout.h - Package up all of the global variables into a per-cpu callout structure. - Start one thread per-cpu. Threads are not strictly bound. They prefer to execute on the native cpu but may migrate temporarily if interrupts are starving callout processing. - Run all callouts by default in the thread for cpu0 to maintain current ordering and concurrency guarantees. Many consumers may not properly handle concurrent execution. - The new callout_reset_on() api allows specifying a particular cpu to execute the callout on. This may migrate a callout to a new cpu. callout_reset() schedules on the last assigned cpu while callout_reset_curcpu() schedules on the current cpu. 177949 - Correct a major error introduced in the per-cpu timeout commit. Sleep and wakeup require the same wait channel to function properly. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T19:01:32.505742Z K 7 svn:log V 1002 180608: Fix a race which could result in some timeout buckets being skipped. - When a tick occurs on a cpu, iterate from cs_softticks until ticks. The per-cpu tick processing happens asynchronously with the actual adjustment of the 'ticks' variable. Sometimes the results may be visible before the local call and sometimes after. Previously this could cause a one tick window where we didn't evaluate the bucket. - In softclock fetch curticks before incrementing cc_softticks so we don't skip insertions which were made for the current time. Sponsored by: Nokia 181191 add callout_schedule; besides being useful it also improves compatibility with other systems Reviewed by: ed, battlez 184385 After a machine has been up for a bit more than 20 days with HZ=1000, "ticks" goes negative. This breaks the signed comparison in softclock. This causes sleep() to never wake up, tcp to stop, etc etc. This is bad(TM). Use the SEQ_LT() method from tcp's sequence number comparisons. END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-18T19:02:55.538456Z K 7 svn:log V 23 add block ack frame id END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T19:07:58.848186Z K 7 svn:log V 36 use per-cpu callouts for tcp timers END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T19:26:25.373558Z K 7 svn:log V 25 release sleepq in wakeup END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-18T19:33:59.422376Z K 7 svn:log V 144 Bump CACHE_LINE_SIZE to 128 for x86. Intel's manuals explicitly recommend using 128 byte alignment for locks. (See IA-32 SDM Vol 3A 7.11.6.7) END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T19:47:54.934010Z K 7 svn:log V 1476 merge hwpmc fix and new witness 180794 - Provide kernelname as the name for process with P_KTHREAD set as otherwise their textvp is NULL. 181695 Introduce some WITNESS improvements: - Speedup the lock orderings lookup modifying the witness graph from a linked tree to a matrix. A table lookup caches the lock orderings in order to make a O(1) access for them. Any witness object has an unique index withing this lookup cache table. - Reduce the lock contention on w_mtx acquiring it only when the LOR actually happens and not in a sane case. In order to do this don't totally flush lock lists (per-CPU spinlocks list and per-thread sleeplocks list) but check for ll_count anytime we need to have to verify allocations sanity. - Introduce the function witness_thread_exit() in the witness namespace which should verify a thread doesn't hold any witness occurrence why exiting. - Rename the sysctl debug.witness.graphs into debug.witness.fullgraph and add debug.witness.badstacks which prints out stacks for LOR revealed. This is implemented using the stack(9) support, which makes WITNESS to be dependent by the STACK option or by the DDB (including STACK) option. - Fix style(9) for src/sys/kern/subr_witness.c The hash table approach has been developed by Ilya Maykov on the behalf of Isilon Systems which kindly released the patch. Jeff Roberson, ported the patch to -CURRENT and fixed w_mtx contention, on the behalf of Nokia. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T19:50:18.081722Z K 7 svn:log V 460 merge 187681 - Make the keg abstraction more complete. Permit a zone to have multiple backend kegs so it may source compatible memory from multiple backends. This is useful for cases such as NUMA or different layouts for the same memory type. - Provide a new api for adding new backend kegs to secondary zones. - Provide a new flag for adjusting the layout of zones to stagger allocations better across cache lines. Sponsored by: Nokia END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T19:54:34.848574Z K 7 svn:log V 266 merge 187880 - Allocate apic vectors on a per-cpu basis. This allows us to allocate more irqs as we have more cpus. This is principally useful on systems with msi devices which may want many irqs per-cpu. Discussed with: jhb Sponsored by: Nokia END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T19:56:06.503962Z K 7 svn:log V 477 merge 188904 - Resolve an issue where we may clear an idt while an interrupt on a different cpu is still assigned to that vector by never clearing idt entries. This was only provided as a debugging feature and the bugs are caught by other means. - Drop the sched lock when rebinding to reassign an interrupt vector to a new cpu so that pending interrupts have a chance to be delivered before removing the old vector. Discussed with: tegge, jhb END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-18T20:23:16.619634Z K 7 svn:log V 59 MFC: Convert IPFW_DEFAULT_TO_ACCEPT into a loader tunable. END K 10 svn:author V 8 rmacklem K 8 svn:date V 27 2009-05-18T21:22:03.060718Z K 7 svn:log V 310 Change the experimental NFSv4 client so that it does not do the NFSv4 Close operations until ncl_inactive(). This is necessary so that the Open StateIDs are available for doing I/O on mmap'd files after VOP_CLOSE(). I also changed some indentation for the nfscl_getclose() function. Approved by: kib (mentor) END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T21:23:13.600209Z K 7 svn:log V 13 merge 189789 END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-18T21:36:17.202089Z K 7 svn:log V 31 Fix bug in netisr_setqlimit(). END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-18T21:44:33.470254Z K 7 svn:log V 48 MFC: Remove dependency of vlan.ko on miibus.ko. END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-18T21:46:46.694820Z K 7 svn:log V 132 - Fix typo in description of 'net.inet.ip.fw.autoinc_step'. - Use 'vnet_ipfw' instead of 'vnet_inet' for 'net.inet.ip.fw.one_pass'. END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-18T21:47:32.116203Z K 7 svn:log V 94 Add a read-only sysctl hw.pci.mcfg to mirror the tunable by the same name. MFC after: 1 week END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-18T21:50:06.604514Z K 7 svn:log V 374 - Add a tunable 'hw.mca.enabled' that can be used to enable/disable the machine check code. Disable it by default for now. - When computing the mask of bits that determines a non-restartable event during a machine check exception, or-in the overflow flag rather than replacing the other flags. PR: i386/134586 [2] Submitted by: Andi Kleen andi-fbsd firstfloor.org END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-18T21:57:28.708928Z K 7 svn:log V 327 Implement nh_m2flow for IPv4 -- if no flow ID is available on an mbuf destined for NETISR_IP, calculate a flow ID quickly using the source and destination IP addresses. It's easy to imagine something more complicated here, such as using the same RSS algorithm and parameters programmed into cxgb and other supporting devices. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T21:58:21.352264Z K 7 svn:log V 651 merge 175872 Give MEXTADD() another argument to make both void pointers to the free function controlable, instead of passing the KVA of the buffer storage as the first argument. Fix all conventional users of the API to pass the KVA of the buffer as the first argument, to make this a no-op commit. Likely break the only non-convetional user of the API, after informing the relevant committer. Update the mbuf(9) manual page, which was already out of sync on this point. This paves the way for giving sendfile(9) a way to wait for the passed storage to have been accessed before returning. This does not affect the memory layout or size of mbufs. END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-18T21:58:57.986547Z K 7 svn:log V 48 MFC: Remove dependency of vlan.ko on miibus.ko. END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-18T22:16:51.084974Z K 7 svn:log V 322 Use rmlocks for netisr2 and enable read-locking in the dispatch path. Despite deeply upsetting WITNESS, this should actually be safe as the rmlock is always acquired read-locked in any processing path that could otherwise lead to a lock order reversal, so we'll want to supress these warnings as with pfil lock warnings. END K 10 svn:author V 7 delphij K 8 svn:date V 27 2009-05-18T22:27:42.395781Z K 7 svn:log V 71 Virgin import of Christos Zoulas's FILE 5.03. Security: CVE-2009-1515 END K 10 svn:author V 7 delphij K 8 svn:date V 27 2009-05-18T22:29:44.266074Z K 7 svn:log V 35 "Tag" the file 5.03 vendor import. END K 10 svn:author V 7 delphij K 8 svn:date V 27 2009-05-18T22:34:33.626962Z K 7 svn:log V 89 Merge vendor/file/dist@192348, bringing FILE 5.03 to 8-CURRENT. Security: CVE-2009-1515 END K 10 svn:author V 2 bz K 8 svn:date V 27 2009-05-18T22:34:44.065835Z K 7 svn:log V 277 Revert the logical change of r192341. net.inet.ip.fw.one_pass is a classic ip_input.c variable and is used in the pfil and bridge code as well. As ipfw is loadable we need to always provide it. That is the reason why it lives in struct vnet_inet and not in struct vnet_ipfw. END K 10 svn:author V 6 marcel K 8 svn:date V 27 2009-05-18T22:45:09.805606Z K 7 svn:log V 219 MFC r190834: Fix hangs caused by hardware that signals receive errors (framing, parity, etc), but does not indicate characters being received. Since no chracters have been received, ignore the line errors. PR: 131006 END K 10 svn:author V 7 delphij K 8 svn:date V 27 2009-05-18T22:46:59.153924Z K 7 svn:log V 42 FILE has been upgraded from 4.23 to 5.03. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-18T22:54:47.175881Z K 7 svn:log V 246 merge 191958 - Merge in changes from jeff_mbuf with one addition. - In m_refm() make all referenced data available in the ext area for compatibility with existing code that makes assumptions about referenced sizes. also fix cxgb compile END K 10 svn:author V 5 gonzo K 8 svn:date V 27 2009-05-18T23:20:56.119464Z K 7 svn:log V 52 - Add support for MX25Lxxx SPI flash (readonly atm) END K 10 svn:author V 5 gonzo K 8 svn:date V 27 2009-05-18T23:30:15.873879Z K 7 svn:log V 55 - Unwire spibus from at91_spi in projects/mips as well END K 10 svn:author V 5 gonzo K 8 svn:date V 27 2009-05-18T23:32:04.700144Z K 7 svn:log V 36 - Add SPI bus driver for ar71xx SoC END K 10 svn:author V 5 gonzo K 8 svn:date V 27 2009-05-18T23:36:11.905388Z K 7 svn:log V 30 - Add spibus and mx25l device END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T00:02:58.487715Z K 7 svn:log V 922 merge 174647 Refactor select to reduce contention and hide internal implementation details from consumers. - Track individual selecters on a per-descriptor basis such that there are no longer collisions and after sleeping for events only those descriptors which triggered events must be rescaned. - Protect the selinfo (per descriptor) structure with a mtx pool mutex. mtx pool mutexes were chosen to preserve api compatibility with existing code which does nothing but bzero() to setup selinfo structures. - Use a per-thread wait channel rather than a global wait channel. - Hide select implementation details in a seltd structure which is opaque to the rest of the kernel. - Provide a 'selsocket' interface for those kernel consumers who wish to select on a socket when they have no fd so they no longer have to be aware of select implementation details. Tested by: kris Reviewed on: arch END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T01:14:37.540307Z K 7 svn:log V 52 - back out direct map hack - it is no longer needed END K 10 svn:author V 6 scottl K 8 svn:date V 27 2009-05-19T01:41:11.906334Z K 7 svn:log V 33 Updated PCI ID's from the vendor END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T01:57:26.233916Z K 7 svn:log V 24 fix new witness compile END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T01:58:00.535200Z K 7 svn:log V 24 remove unneeded devices END K 10 svn:author V 5 gonzo K 8 svn:date V 27 2009-05-19T02:43:21.763367Z K 7 svn:log V 332 - Cleanup ticker initialization code. For some MIPS cpu Counter register increments only every second cycle. The only timing references for us is Count value. Therefore it's better to convert frequencies related to it and use them. Besides cleanup this commit fixes twice more then requested sleep interval problem. END K 10 svn:author V 5 gonzo K 8 svn:date V 27 2009-05-19T02:51:30.049583Z K 7 svn:log V 48 - ar71xx increases Count value every two cycles END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T03:09:04.986815Z K 7 svn:log V 41 update arguments to sleepq_resume_thread END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T03:21:51.927096Z K 7 svn:log V 51 fix calls to sleepq_broadcast to pass a 0 priority END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T03:45:36.993485Z K 7 svn:log V 24 fix more sleepq callers END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T04:43:00.902834Z K 7 svn:log V 584 merge 174988 Remove explicit locking of struct file. - Introduce a finit() which is used to initailize the fields of struct file in such a way that the ops vector is only valid after the data, type, and flags are valid. - Protect f_flag and f_count with atomic operations. - Remove the global list of all files and associated accounting. - Rewrite the unp garbage collection such that it no longer requires the global list of all files and instead uses a list of all unp sockets. - Mark sockets in the accept queue so we don't incorrectly gc them. Tested by: kris, pho END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T04:47:30.886362Z K 7 svn:log V 447 merge 175026,175211,175212 - Place the fhold() in unp_internalize_fp to be more consistent with refs. - Clear all of the gc flags before doing a run. Stale flags were causing us to skip some descriptors. - If a unp socket has been marked REF in a gc pass it can't be dead. Found by: rwatson's test tool. Correct typos in comments. MFC after: 3 weeks Remove "lock pushdown" todo item in comment -- I did that for 7.0. MFC after: 3 weeks END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T05:17:41.893830Z K 7 svn:log V 403 merge 177368, 177374, 183297 - Relax requirements for p_numthreads, p_threads, p_swtick, and p_nice from requiring the per-process spinlock to only requiring the process lock. - Reflect these changes in the proc.h documentation and consumers throughout the kernel. This is a substantial reduction in locking cost for these fields and was made possible by recent changes to threading support. END K 10 svn:author V 5 kmacy K 8 svn:date V 27 2009-05-19T05:36:10.248340Z K 7 svn:log V 1509 merge changes to support lockless fget 187677,187682,187693,187996,189708,192080 Fix errors introduced when I rewrote select. - Restructure selscan() and selrescan() to avoid producing extra selfps when we have a fd in multiple sets. As described below multiple selfps may still exist for other reasons. - Make selrescan() tolerate multiple selfds for a given descriptor set since sockets use two selinfos per fd. If an event on each selinfo fires selrescan() will see the descriptor twice. This could result in select() returning 2x the number of fds actually existing in fd sets. Reported by: mgleason@ncftp.com - bit has to be fd_mask to work properly on 64bit platforms. Constants must also be cast even though the result ultimately is promoted to 64bit. - Correct a loop index upper bound in selscan() Fix select on platforms where sizeof(long) != sizeof(int). This used to work by accident before the cleanup done in revision 187693. Approved by: kan (mentor) - Implement a lockless file descriptor lookup algorithm in fget_unlocked(). - Save old file descriptor tables created on expansion until the entire descriptor table is freed so that pointers may be followed without regard for expanders. - Mark the file zone as NOFREE so we may attempt to reference potentially freed files. - Convert several fget_locked() users to fget_unlocked(). This requires us to manage reference counts explicitly but reduces locking overhead in the common case. END K 10 svn:author V 7 dchagin K 8 svn:date V 27 2009-05-19T09:10:53.270980Z K 7 svn:log V 274 Validate user-supplied arguments values. Args argument is a pointer to the structure located in user space in which the socketcall arguments are packed. The structure must be copied to the kernel instead of direct dereferencing. Approved by: kib (mentor) MFC after: 1 week END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-19T09:26:29.911446Z K 7 svn:log V 39 Use NETISR2 in GENERIC in this branch. END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-19T09:28:09.842306Z K 7 svn:log V 58 Minor style tweak for netisr2 sysctls. Submitted by: bde END K 10 svn:author V 5 trasz K 8 svn:date V 27 2009-05-19T12:10:48.466819Z K 7 svn:log V 43 Add links to libgeom(3) where appropriate. END K 10 svn:author V 6 rpaulo K 8 svn:date V 27 2009-05-19T12:42:01.282074Z K 7 svn:log V 73 Handle MBSS in gethtadjustflags(). Sponsored by: The FreeBSD Foundation END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-19T14:01:57.724548Z K 7 svn:log V 105 MFC: Add more PCI-express capability registers and display additional details about PCI-express devices. END K 10 svn:author V 3 avg K 8 svn:date V 27 2009-05-19T14:08:21.612035Z K 7 svn:log V 407 sysctl_rman: report shared resources to devinfo shared uses of a resource are recorded on a sub-list hanging off a main resource object on a main resource list; without this change a shared resource (e.g. irq) is reported only once by devinfo -r/-u; with this change the resource is reported for each driver that allocates it (which is even more than what vmstat -i -a reports). Approved by: jhb (mentor) END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-19T14:10:14.719897Z K 7 svn:log V 1233 Read-locking of netisr2 during dispatch is no longer optional, so update comment. Add an XXXRW because assertions aren't implemented for rmlocks. Break out net.isr2.direct into two sysctls: net.isr2.direct_force to force direct dispatch to always be used even if work should be processed by a workstream, and net.isr2.direct_enable, which allows direct dispatch to be used. Document the three modes of operation we support using these flags. Add a new direct dispatch mode, "hybrid", in which work destined for the current CPU may directly dispatch, but work destined for other CPUs will be queued. Hybrid direct dispatch is only permitted if there is no pending work, the netisr isn't already running, and no other thread is directly dispatching the workstream, as these could lead to misordering. Add a new hybrid direct dispatch counter. Fix a bug in the tunable portion of the net.isr2.maxqlimit sysctl and document it. Make it a writable sysctl. Don't test queue limits during registration with an assert -- instead, cap and printf. Update stats printout in DDB. Default settings remain the same as in the current netisr code: force direct dispatch, create a single worker thread, and don't bind it to a specific CPU. END K 10 svn:author V 3 avg K 8 svn:date V 27 2009-05-19T14:23:54.348789Z K 7 svn:log V 599 find: do not silently disable -L when -delete is used First of all, current behavior is not documented and confusing, and it can be very dangerous in the following sequence: find -L . -type l find -L . -type l -delete (the second line is even suggested by find(1)). Instead simply refuse to proceed when -L and -delete are both used. A descriptive error message is provided. The following command can be safely used to remove broken links: find -L . -type l -print0 | xargs rm -0 To do: update find(1) PR: bin/90687 Obtained from: Anatoli Klassen Approved by: jhb (mentor) END K 10 svn:author V 3 flz K 8 svn:date V 27 2009-05-19T14:26:41.791050Z K 7 svn:log V 74 Skip @pkgdep if there's no argument. Submitted by: pav MFC after: 1 week END K 10 svn:author V 3 avg K 8 svn:date V 27 2009-05-19T14:44:07.080430Z K 7 svn:log V 51 open acpica vendor area Approved by: jhb (mentor) END K 10 svn:author V 3 avg K 8 svn:date V 27 2009-05-19T14:46:41.038192Z K 7 svn:log V 164 flatten acpica vendor area do it for current ('dist') and previous versions; disable keywords expansion at the same time Reviewed by: attilio Approved by: rpaulo END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-19T14:51:48.161413Z K 7 svn:log V 103 AH_SUPPORT_TDMA is gone; ath now honors IEEE80211_SUPPORT_TDMA Noticed by: Jon Loeliger END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-19T15:33:41.252860Z K 7 svn:log V 64 MFC: Add SATA and PCI Advanced Features capabilities reporting. END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-19T15:47:42.180445Z K 7 svn:log V 239 MFC: - Add a void pointer to the ata-pci controller softc to allow chipset-specific code to attach chipset-specific data. - Use chipset-specific data in the acard and promise chipsets rather than changing the ivars of ATA PCI devices. END K 10 svn:author V 3 avg K 8 svn:date V 27 2009-05-19T15:50:36.947748Z K 7 svn:log V 78 bootstrap merge history for acpica contrib code Approved by: rpaulo (mentor) END K 10 svn:author V 6 rpaulo K 8 svn:date V 27 2009-05-19T15:53:07.581996Z K 7 svn:log V 170 Remove the bogus decrease of datalen and pad meshcntl struct instead. This is a temporary hack but makes packets flow in the mesh. Sponsored by: The FreeBSD Foundation END K 10 svn:author V 6 rpaulo K 8 svn:date V 27 2009-05-19T15:56:27.380935Z K 7 svn:log V 110 Add next hop to the node struct in preparation for HWMP implementation. Sponsored by: The FreeBSD Foundation END K 10 svn:author V 6 rpaulo K 8 svn:date V 27 2009-05-19T15:57:05.456382Z K 7 svn:log V 76 Placeholders for HWMP implementation. Sponsored by: The FreeBSD Foundation END K 10 svn:author V 6 rpaulo K 8 svn:date V 27 2009-05-19T15:58:05.925989Z K 7 svn:log V 59 Add ieee80211_hwmp.c Sponsored by: The FreeBSD Foundation END K 10 svn:author V 6 rpaulo K 8 svn:date V 27 2009-05-19T17:10:57.753684Z K 7 svn:log V 120 Parse path replies and path error IEs. Ignore data frames with ni pointing to us. Sponsored by: The FreeBSD Foundation END K 10 svn:author V 6 rpaulo K 8 svn:date V 27 2009-05-19T17:14:52.265640Z K 7 svn:log V 88 Use a similar style for printing mesh link state. Sponsored by: The FreeBSD Foundation END K 10 svn:author V 4 joel K 8 svn:date V 27 2009-05-19T17:17:53.473887Z K 7 svn:log V 58 Add Dell PowerEdge R200 and R300 to the hardware section. END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-19T17:30:13.731180Z K 7 svn:log V 77 add HAL_CAP_INTRMASK to return the set of interrupts supported by the device END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-19T17:35:15.413786Z K 7 svn:log V 145 remove special handling for BNR; it is direct mapped to the harwdare so can be added to HAL_INT_COMMON except on the 5210 where it doesn't exist END K 10 svn:author V 3 gnn K 8 svn:date V 27 2009-05-19T17:40:22.148689Z K 7 svn:log V 183 Add a new program, perror, which takes an errno as a command line argument and outputs the associated textual message in the same way that perror(3) would if called within a program. END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-19T17:43:31.194136Z K 7 svn:log V 14 minor cleanup END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-19T17:53:53.029317Z K 7 svn:log V 183 add TBTT interrupt support; this was added in Griffin so consumers should check HAL_CAP_INTRMASK before using it NB: didn't test 11n parts yet so supported only for 5212-class parts END K 10 svn:author V 3 sam K 8 svn:date V 27 2009-05-19T17:54:32.191230Z K 7 svn:log V 69 correct HAL_INT_BNR comment, this bit is mapped directly the h/w now END K 10 svn:author V 6 scottl K 8 svn:date V 27 2009-05-19T18:14:04.881913Z K 7 svn:log V 35 Merge PCI ID additions from vendor END K 10 svn:author V 6 rpaulo K 8 svn:date V 27 2009-05-19T18:48:10.444624Z K 7 svn:log V 18 MFC 191961:192401 END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-19T19:21:25.198962Z K 7 svn:log V 157 With SMPng, DEVICE_POLLING uses its own idle threads, rather than the system idle loop, to run ether_poll(), so make ether_poll() static. MFC after: 1 week END K 10 svn:author V 6 scottl K 8 svn:date V 27 2009-05-19T19:27:07.571099Z K 7 svn:log V 26 Merge PCI IDs from vendor END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-19T19:45:36.897097Z K 7 svn:log V 155 MFC: Add a MI intr_event_handle() routine for the non-INTR_FILTER case. This allows all the INTR_FILTER #ifdef's to be removed from the MD interrupt code. END K 10 svn:author V 3 gnn K 8 svn:date V 27 2009-05-19T20:16:18.250047Z K 7 svn:log V 46 Fix a few style(9) nits. Submitted by: danfe END K 10 svn:author V 7 rwatson K 8 svn:date V 27 2009-05-19T20:23:32.440602Z K 7 svn:log V 207 Implement a hybrid cross-CPU dispatch model for netisr2, in which a directly dispatching thread is willing to directly dispatch packets normally processed on another CPU if that CPU's netisr thread is idle. END K 10 svn:author V 3 jhb K 8 svn:date V 27 2009-05-19T22:07:54.705838Z K 7 svn:log V 174 MFC: - Add the interrupt vector number to intr_event_create so MI code can lookup hard interrupt events by number. - Add support to cpuset for binding hardware interrupts. END K 10 svn:author V 8 brueffer K 8 svn:date V 27 2009-05-19T22:28:33.649216Z K 7 svn:log V 137 Since audit(4) isn't based on posix1e, remove the commented out audit.h header, xref libbsm(3). Submitted by: rwatson MFC after: 3 days END K 10 svn:author V 3 hrs K 8 svn:date V 27 2009-05-20T00:55:26.123272Z K 7 svn:log V 58 Document fxp(4) TSO issue on 7.2R. Submitted by: delphij END K 10 svn:author V 7 yongari K 8 svn:date V 27 2009-05-20T02:13:01.955200Z K 7 svn:log V 101 MFC r190303: Fix inversed logic. pci_find_extcap() returns 0 when it finds specified capability. END