svn commit: r235773 - head/contrib/ncurses/man
Author: dim Date: Tue May 22 06:28:53 2012 New Revision: 235773 URL: http://svn.freebsd.org/changeset/base/235773 Log: Correct use_screen() and use_window() prototypes in curs_threads(3x). Submitted by: Yanhui Shen shen@gmail.com MFC after:3 days Modified: head/contrib/ncurses/man/curs_threads.3x Modified: head/contrib/ncurses/man/curs_threads.3x == --- head/contrib/ncurses/man/curs_threads.3xTue May 22 05:18:30 2012 (r235772) +++ head/contrib/ncurses/man/curs_threads.3xTue May 22 06:28:53 2012 (r235773) @@ -45,9 +45,9 @@ .br \fBint set_tabsize(int size);\fR .br -\fBint use_screen(SCREEN *scr, NCURSES_WINDOW_CB func, void *data);\fR +\fBint use_screen(SCREEN *scr, NCURSES_SCREEN_CB func, void *data);\fR .br -\fBint use_window(WINDOW *win, NCURSES_SCREEN_CB func, void *data);\fR +\fBint use_window(WINDOW *win, NCURSES_WINDOW_CB func, void *data);\fR .br .SH DESCRIPTION This implementation can be configured to provide rudimentary support ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235774 - head/sys/dev/ath
Author: adrian Date: Tue May 22 06:31:03 2012 New Revision: 235774 URL: http://svn.freebsd.org/changeset/base/235774 Log: Fix up some corner cases with aggregation handling. I've come across a weird scenario in net80211 where two TX streams will happily attempt to setup an aggregation session together. If we're very lucky, it happens concurrently on separate CPUs and the total lack of locking in the net80211 aggregation code causes this stuff to race. Badly. So 1 call would occur to the ath(4) addba start, but only one call would complete to addba complete or timeout. The TID would thus stay paused. The real fix is to implement some proper per-node (or maybe per-TID) locking in net80211, which then could be leveraged by the ath(4) TX aggregation code. Whilst I'm at it, shuffle around the debugging messages a bit. I like to keep people on their toes. Modified: head/sys/dev/ath/if_ath_tx.c head/sys/dev/ath/if_athvar.h Modified: head/sys/dev/ath/if_ath_tx.c == --- head/sys/dev/ath/if_ath_tx.cTue May 22 06:28:53 2012 (r235773) +++ head/sys/dev/ath/if_ath_tx.cTue May 22 06:31:03 2012 (r235774) @@ -1580,21 +1580,21 @@ ath_tx_start(struct ath_softc *sc, struc * reached.) */ if (txq == avp-av_mcastq) { - DPRINTF(sc, ATH_DEBUG_SW_TX_CTRL, + DPRINTF(sc, ATH_DEBUG_SW_TX, %s: bf=%p: mcastq: TX'ing\n, __func__, bf); ATH_TXQ_LOCK(txq); ath_tx_xmit_normal(sc, txq, bf); ATH_TXQ_UNLOCK(txq); } else if (type == IEEE80211_FC0_TYPE_CTL subtype == IEEE80211_FC0_SUBTYPE_BAR) { - DPRINTF(sc, ATH_DEBUG_SW_TX_CTRL, + DPRINTF(sc, ATH_DEBUG_SW_TX, %s: BAR: TX'ing direct\n, __func__); ATH_TXQ_LOCK(txq); ath_tx_xmit_normal(sc, txq, bf); ATH_TXQ_UNLOCK(txq); } else { /* add to software queue */ - DPRINTF(sc, ATH_DEBUG_SW_TX_CTRL, + DPRINTF(sc, ATH_DEBUG_SW_TX, %s: bf=%p: swq: TX'ing\n, __func__, bf); ath_tx_swq(sc, ni, txq, bf); } @@ -3113,7 +3113,7 @@ ath_tx_tid_cleanup(struct ath_softc *sc, struct ath_buf *bf, *bf_next; ath_bufhead bf_cq; - DPRINTF(sc, ATH_DEBUG_SW_TX_CTRL, + DPRINTF(sc, ATH_DEBUG_SW_TX_BAW, %s: TID %d: called\n, __func__, tid); TAILQ_INIT(bf_cq); @@ -4336,7 +4336,15 @@ ath_addba_request(struct ieee80211_node * fall within it. */ ATH_TXQ_LOCK(sc-sc_ac2q[atid-ac]); - ath_tx_tid_pause(sc, atid); + /* +* This is a bit annoying. Until net80211 HT code inherits some +* (any) locking, we may have this called in parallel BUT only +* one response/timeout will be called. Grr. +*/ + if (atid-addba_tx_pending == 0) { + ath_tx_tid_pause(sc, atid); + atid-addba_tx_pending = 1; + } ATH_TXQ_UNLOCK(sc-sc_ac2q[atid-ac]); DPRINTF(sc, ATH_DEBUG_SW_TX_CTRL, @@ -4397,6 +4405,7 @@ ath_addba_response(struct ieee80211_node r = sc-sc_addba_response(ni, tap, status, code, batimeout); ATH_TXQ_LOCK(sc-sc_ac2q[atid-ac]); + atid-addba_tx_pending = 0; /* * XXX dirty! * Slide the BAW left edge to wherever net80211 left it for us. @@ -4500,6 +4509,10 @@ ath_addba_response_timeout(struct ieee80 DPRINTF(sc, ATH_DEBUG_SW_TX_CTRL, %s: called; resuming\n, __func__); + ATH_TXQ_LOCK(sc-sc_ac2q[atid-ac]); + atid-addba_tx_pending = 0; + ATH_TXQ_UNLOCK(sc-sc_ac2q[atid-ac]); + /* Note: This updates the aggregate state to (again) pending */ sc-sc_addba_response_timeout(ni, tap); Modified: head/sys/dev/ath/if_athvar.h == --- head/sys/dev/ath/if_athvar.hTue May 22 06:28:53 2012 (r235773) +++ head/sys/dev/ath/if_athvar.hTue May 22 06:31:03 2012 (r235774) @@ -106,6 +106,7 @@ struct ath_tid { TAILQ_ENTRY(ath_tid)axq_qelem; int sched; int paused; /* 0 if the TID has been paused */ + int addba_tx_pending; /* TX ADDBA pending */ int bar_wait; /* waiting for BAR */ int bar_tx; /* BAR TXed */ ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235775 - head/sys/modules/bwi
Author: uqs Date: Tue May 22 06:33:08 2012 New Revision: 235775 URL: http://svn.freebsd.org/changeset/base/235775 Log: Fix make depend. Modified: head/sys/modules/bwi/Makefile Modified: head/sys/modules/bwi/Makefile == --- head/sys/modules/bwi/Makefile Tue May 22 06:31:03 2012 (r235774) +++ head/sys/modules/bwi/Makefile Tue May 22 06:33:08 2012 (r235775) @@ -5,7 +5,7 @@ KMOD = if_bwi SRCS = if_bwi.c if_bwi_pci.c bwimac.c bwiphy.c bwirf.c -SRCS += device_if.h bus_if.h pci_if.h opt_inet.h opt_bwi.h +SRCS += device_if.h bus_if.h pci_if.h opt_inet.h opt_bwi.h opt_wlan.h opt_bwi.h: echo '#define BWI_DEBUG 1' opt_bwi.h ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235739 - head/lib/libc/gen
On Mon, 21 May 2012, Guy Helmer wrote: Log: Apply style(9) to return and switch/case statements. Reviewed by: delphij (prior version of the patch) Modified: head/lib/libc/gen/getnetgrent.c Modified: head/lib/libc/gen/getnetgrent.c == --- head/lib/libc/gen/getnetgrent.c Mon May 21 19:58:40 2012 (r235738) +++ head/lib/libc/gen/getnetgrent.c Mon May 21 21:04:29 2012 (r235739) ... @@ -311,32 +311,35 @@ _revnetgr_lookup(char* lookupdom, char* for (rot = 0; ; rot++) { switch (rot) { - case(0): snprintf(key, MAXHOSTNAMELEN, %s.%s, - str, dom?dom:lookupdom); -break; - case(1): snprintf(key, MAXHOSTNAMELEN, %s.*, - str); -break; - case(2): snprintf(key, MAXHOSTNAMELEN, *.%s, - dom?dom:lookupdom); -break; - case(3): snprintf(key, MAXHOSTNAMELEN, *.*); -break; - default: return(0); + case(0): + snprintf(key, MAXHOSTNAMELEN, %s.%s, str, + dom ? dom : lookupdom); + break; + case(1): + snprintf(key, MAXHOSTNAMELEN, %s.*, str); + break; + case(2): + snprintf(key, MAXHOSTNAMELEN, *.%s, + dom ? dom : lookupdom); + break; + case(3): + snprintf(key, MAXHOSTNAMELEN, *.*); + break; Thanks, but a fuller application would have removed the obfuscatory parentheses that make case() look like a function call... + default: return (0); ... and split the case statements after : in all cases. } y = yp_match(lookupdom, map, key, strlen(key), result, resultlen); You fixed the continuation indentation in the case statement but not here. if (y == 0) { rv = _listmatch(result, group, resultlen); free(result); - if (rv) return(1); + if (rv) return (1); Another statement not started on a new line. } else if (y != YPERR_KEY) { /* * If we get an error other than 'no * such key in map' then something is * wrong and we should stop the search. */ - return(-1); + return (-1); } } } These style bugs weren't in the CSRG version of course. The YP code added many. The most obvious ones are the case(n) and gnu-style continuation indentation. Bruce ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235776 - head/sys/vm
Author: andrew Date: Tue May 22 07:04:23 2012 New Revision: 235776 URL: http://svn.freebsd.org/changeset/base/235776 Log: Fix booting on ARM. In PHYS_TO_VM_PAGE() when VM_PHYSSEG_DENSE is set the check if we are past the end of vm_page_array was incorrect causing it to return NULL. This value is then used in vm_phys_add_page causing a data abort. Reviewed by: alc, kib, imp Tested by:stas Modified: head/sys/vm/vm_page.c Modified: head/sys/vm/vm_page.c == --- head/sys/vm/vm_page.c Tue May 22 06:33:08 2012(r235775) +++ head/sys/vm/vm_page.c Tue May 22 07:04:23 2012(r235776) @@ -647,7 +647,7 @@ PHYS_TO_VM_PAGE(vm_paddr_t pa) long pi; pi = atop(pa); - if (pi = first_page pi vm_page_array_size) { + if (pi = first_page (pi - first_page) vm_page_array_size) { m = vm_page_array[pi - first_page]; return (m); } ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235777 - head/sys/kern
Author: harti Date: Tue May 22 07:23:41 2012 New Revision: 235777 URL: http://svn.freebsd.org/changeset/base/235777 Log: Make dumptid non-static. It is used by libkvm to detect whether this is a VNET-kernel or not. gcc used to put the static symbol into the symbol table, clang does not. This fixes the 'netstat: no namelist' error seen on clang+VNET systems. Modified: head/sys/kern/kern_shutdown.c Modified: head/sys/kern/kern_shutdown.c == --- head/sys/kern/kern_shutdown.c Tue May 22 07:04:23 2012 (r235776) +++ head/sys/kern/kern_shutdown.c Tue May 22 07:23:41 2012 (r235777) @@ -151,7 +151,7 @@ static struct dumperinfo dumper;/* our /* Context information for dump-debuggers. */ static struct pcb dumppcb; /* Registers. */ -static lwpid_t dumptid;/* Thread ID. */ +lwpid_t dumptid; /* Thread ID. */ static void poweroff_wait(void *, int); static void shutdown_halt(void *junk, int howto); ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235778 - in head/sys: boot/fdt/dts conf dev/fdt geom sys
Author: gber Date: Tue May 22 08:33:14 2012 New Revision: 235778 URL: http://svn.freebsd.org/changeset/base/235778 Log: Add a new geom class which allows to divide NAND Flash chip into partitions. Partitions are created based on data in dts file which are extracted and interpreted by slicer. Obtained from: Semihalf Supported by: FreeBSD Foundation, Juniper Networks Added: head/sys/dev/fdt/fdt_slicer.c (contents, props changed) head/sys/geom/geom_flashmap.c (contents, props changed) head/sys/sys/slicer.h (contents, props changed) Modified: head/sys/boot/fdt/dts/sheevaplug.dts head/sys/conf/files head/sys/dev/fdt/fdt_common.h Modified: head/sys/boot/fdt/dts/sheevaplug.dts == --- head/sys/boot/fdt/dts/sheevaplug.dtsTue May 22 07:23:41 2012 (r235777) +++ head/sys/boot/fdt/dts/sheevaplug.dtsTue May 22 08:33:14 2012 (r235778) @@ -88,7 +88,16 @@ bank-width = 2; device-width = 1; + slice@0 { + reg = 0x0 0x20; + label = u-boot; + read-only; + }; + slice@20 { + reg = 0x20 0x1fe0; + label = root; + }; }; }; Modified: head/sys/conf/files == --- head/sys/conf/files Tue May 22 07:23:41 2012(r235777) +++ head/sys/conf/files Tue May 22 08:33:14 2012(r235778) @@ -1223,6 +1223,7 @@ dev/fatm/if_fatm.coptional fatm pci dev/fb/splash.coptional splash dev/fdt/fdt_common.c optional fdt dev/fdt/fdt_pci.c optional fdt pci +dev/fdt/fdt_slicer.c optional fdt cfi | fdt nand dev/fdt/fdt_static_dtb.S optional fdt fdt_dtb_static dev/fdt/fdtbus.c optional fdt dev/fdt/simplebus.coptional fdt @@ -2388,6 +2389,7 @@ geom/geom_disk.c standard geom/geom_dump.c standard geom/geom_event.c standard geom/geom_fox.coptional geom_fox +geom/geom_flashmap.c optional fdt cfi | fdt nand geom/geom_io.c standard geom/geom_kern.c standard geom/geom_map.coptional geom_map Modified: head/sys/dev/fdt/fdt_common.h == --- head/sys/dev/fdt/fdt_common.h Tue May 22 07:23:41 2012 (r235777) +++ head/sys/dev/fdt/fdt_common.h Tue May 22 08:33:14 2012 (r235778) @@ -32,6 +32,7 @@ #ifndef _FDT_COMMON_H_ #define _FDT_COMMON_H_ +#include sys/slicer.h #include contrib/libfdt/libfdt_env.h #include dev/ofw/ofw_bus.h #include machine/fdt.h Added: head/sys/dev/fdt/fdt_slicer.c == --- /dev/null 00:00:00 1970 (empty, because file is newly added) +++ head/sys/dev/fdt/fdt_slicer.c Tue May 22 08:33:14 2012 (r235778) @@ -0,0 +1,115 @@ +/*- + * Copyright (c) 2012 Semihalf. + * All rights reserved. + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions + * are met: + * 1. Redistributions of source code must retain the above copyright + *notice, this list of conditions and the following disclaimer. + * 2. Redistributions in binary form must reproduce the above copyright + *notice, this list of conditions and the following disclaimer in the + *documentation and/or other materials provided with the distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. + */ + +#include sys/cdefs.h +__FBSDID($FreeBSD$); + +#include sys/param.h +#include sys/systm.h +#include sys/kernel.h +#include sys/module.h +#include sys/slicer.h + +#include dev/fdt/fdt_common.h + +#define DEBUG +#undef DEBUG + +#ifdef DEBUG +#define debugf(fmt, args...) do { printf(%s(): , __func__);
svn commit: r235779 - head/sys/boot/fdt/dts
Author: gber Date: Tue May 22 09:27:57 2012 New Revision: 235779 URL: http://svn.freebsd.org/changeset/base/235779 Log: Divide nand flash for DB6281 into two partitions. One for u-boot and second one for general use. Obtained from: Semihalf Supported by: FreeBSD Foundation, Juniper Networks Modified: head/sys/boot/fdt/dts/db88f6281.dts Modified: head/sys/boot/fdt/dts/db88f6281.dts == --- head/sys/boot/fdt/dts/db88f6281.dts Tue May 22 08:33:14 2012 (r235778) +++ head/sys/boot/fdt/dts/db88f6281.dts Tue May 22 09:27:57 2012 (r235779) @@ -89,7 +89,16 @@ bank-width = 2; device-width = 1; + slice@0 { + reg = 0x0 0x20; + label = u-boot; + read-only; + }; + slice@20 { + reg = 0x20 0x7e0; + label = root; + }; }; }; ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235780 - head/include
Author: harti Date: Tue May 22 09:59:49 2012 New Revision: 235780 URL: http://svn.freebsd.org/changeset/base/235780 Log: Fix a compilation error with some compilers: __attribute__ requires two parenthesis for its argument, but instead of using __attribute__ directly, use the appropriate __nonnull macro from cdefs.h. Modified: head/include/malloc_np.h Modified: head/include/malloc_np.h == --- head/include/malloc_np.hTue May 22 09:27:57 2012(r235779) +++ head/include/malloc_np.hTue May 22 09:59:49 2012(r235780) @@ -55,13 +55,11 @@ int mallctlbymib(const size_t *mib, size #defineALLOCM_ERR_OOM 1 #defineALLOCM_ERR_NOT_MOVED2 -intallocm(void **ptr, size_t *rsize, size_t size, int flags) -__attribute__(nonnull(1)); +intallocm(void **ptr, size_t *rsize, size_t size, int flags) __nonnull(1); intrallocm(void **ptr, size_t *rsize, size_t size, size_t extra, -int flags) __attribute__(nonnull(1)); -intsallocm(const void *ptr, size_t *rsize, int flags) -__attribute__(nonnull(1)); -intdallocm(void *ptr, int flags) __attribute__(nonnull(1)); +int flags) __nonnull(1); +intsallocm(const void *ptr, size_t *rsize, int flags) __nonnull(1); +intdallocm(void *ptr, int flags) __nonnull(1); intnallocm(size_t *rsize, size_t size, int flags); __END_DECLS ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235723 - in head: contrib/byacc tools/regression/usr.bin/yacc usr.bin/yacc usr.bin/yacc/test
Baptiste Daroussin b...@freebsd.org writes: Log: Import byacc from invisible island, it brings us lots of compatibilities with bison, keeping full compatibility with our previous yacc implementation. This commit broke the build, in large part because Baptiste tested with Clang instead of GCC, and GCC generates a warning when compiling the generated code. Baptiste has a patch ready to commit, currently undergoing review by myself and other interested parties. I chose not to ask Baptiste to revert the commit, because it is a complex commit that copies some files from a vendor branch into contrib and removes others from the usr.bin/yacc, and I wasn't sure that reverting it wouldn't cause trouble with merge tracking. Lessons learned: 1) always test a *stock* build before committing: % cd /usr/src env __MAKE_CONF=/dev/null make buildworld 2) the contrib part and the usr.bin part should probably have been committed separately. DES -- Dag-Erling Smørgrav - d...@des.no ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235767 - head/lib/libc/net
On Mon, 21 May 2012, Garrett Cooper wrote: On Mon, May 21, 2012 at 6:28 PM, Kevin Lo ke...@freebsd.org wrote: Author: kevlo Date: Tue May 22 01:28:32 2012 New Revision: 235767 URL: http://svn.freebsd.org/changeset/base/235767 Log: ?Add missing header needed by free() ?Reported by: ?tinderbox Please don't use binary characters in mail. Thank you for saving my mailbox :)!!! FWIW, it's weird because my STABLE-9 box didn't reproduce this. This seems to be a bug in both nsparser.y and the new yacc: - new yacc: it is incompatible. Old yacc includes stdlib.h and string.h as the first thing in the generated file, so the C code copied from the .y part sees them too. New yacc puts more its code including its all includes (which still involve stdlib.h) at the end of the generated file. Including stdlib.h nearly first is from Lite2 (skeleton.c 1.7 in 1997). Much later, FreeBSD moved it to the very first thing in the generated file, so as to use namespace pollution (__unused) from it early. __unused was ifdefed and was defined by the yacc skeleton if stdlib.h didn't define it. It took much churn to produce this, but it seems to have been garbage, since old yacc didn't actually generate any use of __unused. The definition was apparently compatibilty cruft to hide unportabilities in .y files that use __unused (I guess these failed for bootstrapping). New yacc doesn't define __unused. - nsparser.y: it uses free() but never declared it. It depended on namespace pollution in the yacc code, and this pollution being in a particular order. The yyparse() incompatibilities seem to be bugs on both sides too: - Old yacc didn't declare yyparse() automatically, so .y files had to do it, but probably shouldn't have. - New yacc still doesn't declare yylex() or yyerror() automatically, so applications still have to declare them. - Some of these may be more the responsibility of the yacc library (to supply defaults) than others. I couldn't see who is required to declare them in POSIX.1-2001. However, a bad example in POSIX.1 has extern int yyparse(); in main() just before calling yyparse(). Badness in this example include several style bugs and the declaration being incomplete (missing void). Old yacc has about 20 lines of ifdefs half just to supply this void for STDC but not for KR. New yacc no longer supports KR, at least here, but it still has 11 lines of ifdefs to get the declaration of yyparse() right for bison compatibility). I don't know if the badness includes having the home made declaration. Nearby bugs: all CSRG and FreeBSD changes including copyrights and history seem to have been lost. Just after the old include of stdlib.h, skeleton.c used to produce a string yyrcsid[] with __FBSDID() for skeleton.c itself in it. New yacc generates a string yysccsid with a hard-coded berkeley 1993 id in it, and the name in this id doesn't even match the file name (it is yaccpar). Another FreeBSD change was to remove this bogus id from the generated file. Some relevant FreeBSD changes: % Index: skeleton.c % === % RCS file: /home/ncvs/src/usr.bin/yacc/skeleton.c,v % retrieving revision 1.1.1.2 % retrieving revision 1.37 % diff -u -2 -r1.1.1.2 -r1.37 % --- skeleton.c6 Jan 1997 03:03:46 - 1.1.1.2 % +++ skeleton.c12 Feb 2003 18:03:55 - 1.37 % @@ -35,7 +35,12 @@ % */ % % +#if 0 % #ifndef lint % static char sccsid[] = @(#)skeleton.c 5.8 (Berkeley) 4/29/95; % -#endif /* not lint */ % +#endif % +#endif FreeBSD normally comments out vendor ids like this. 5.8 is from Lite2; Lite1 had 5.7. New yacc doesn't have any of this. It apparently started from older yacc (before Lite1; indeed, it agrees with the FreeBSD-1 == Net/2 version in not having any CSRG copyrights or this sccsid). % + % +#include sys/cdefs.h % +__FBSDID($FreeBSD: src/usr.bin/yacc/skeleton.c,v 1.37 2003/02/12 18:03:55 davidc Exp $); FreeBSD normally adds this to source files. Except in contrib. yacc is now mostly in contrib, although it hasn't changed all that much. % % #include defs.h % @@ -53,31 +58,42 @@ % /* are conditional. */ % % -char *banner[] = % +const char *banner[] = New yacc has this change too. % { % +#include stdlib.h, See above. % #ifndef lint, % -static char yysccsid[] = \@(#)yaccpar1.9 (Berkeley) 02/21/93\;, Was in Net/2 (?), FreeBSD-1, Lite1 and Lite2. Came back with new yacc. It's interesting that yaccpar 1.9 (Berkeley) 1.9 02/21/93 is only 3 months before the correct id skeleton.c 5.7 (Berkeley) 5/24/93. Hmm, the FreeBSD-1 history is even more instructive than I first thought. FreeBSD-1 started with the CSRG version of skeleton.c on 1993/06/12. This had CSRG copyrights and ids, and yaccpar was at 1.8 01/20/90. But this was upgraded to the newest version on
svn commit: r235781 - head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs
Author: trasz Date: Tue May 22 10:54:42 2012 New Revision: 235781 URL: http://svn.freebsd.org/changeset/base/235781 Log: Fix enforcement of file size limit with O_APPEND on ZFS. vn_rlimit_fsize takes uio-uio_offset and uio-uio_resid into account when determining whether given write would exceed RLIMIT_FSIZE. When APPEND flag is specified, ZFS updates uio-uio_offset to point to the end of file. But this happens after a call to vn_rlimit_fsize, so vn_rlimit_fsize check can be rendered ineffective by thread that opens some file with O_APPEND and lseeks below RLIMIT_FSIZE before calling write. Submitted by: Mateusz Guzik mjguzik at gmail dot com MFC after:2 weeks Modified: head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c Modified: head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c == --- head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c Tue May 22 09:59:49 2012(r235780) +++ head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c Tue May 22 10:54:42 2012(r235781) @@ -838,6 +838,12 @@ zfs_write(vnode_t *vp, uio_t *uio, int i rl = zfs_range_lock(zp, woff, n, RL_WRITER); } + if (vn_rlimit_fsize(vp, uio, uio-uio_td)) { + zfs_range_unlock(rl); + ZFS_EXIT(zfsvfs); + return (EFBIG); + } + if (woff = limit) { zfs_range_unlock(rl); ZFS_EXIT(zfsvfs); @@ -5696,9 +5702,6 @@ zfs_freebsd_write(ap) } */ *ap; { - if (vn_rlimit_fsize(ap-a_vp, ap-a_uio, ap-a_uio-uio_td)) - return (EFBIG); - return (zfs_write(ap-a_vp, ap-a_uio, ioflags(ap-a_ioflag), ap-a_cred, NULL)); } ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235782 - in head/sys: dev/agp modules/agp sys
Author: kib Date: Tue May 22 10:59:26 2012 New Revision: 235782 URL: http://svn.freebsd.org/changeset/base/235782 Log: A rewrite of the i810 bits of the agp(4) driver. New driver supports operations required by GEMified i915.ko. It also attaches to SandyBridge and IvyBridge CPU northbridges now. Sponsored by: The FreeBSD Foundation MFC after:1 month Added: head/sys/dev/agp/agp_i810.h (contents, props changed) Modified: head/sys/dev/agp/agp.c head/sys/dev/agp/agp_i810.c head/sys/dev/agp/agp_if.m head/sys/dev/agp/agppriv.h head/sys/dev/agp/agpreg.h head/sys/dev/agp/agpvar.h head/sys/modules/agp/Makefile head/sys/sys/agpio.h Modified: head/sys/dev/agp/agp.c == --- head/sys/dev/agp/agp.c Tue May 22 10:54:42 2012(r235781) +++ head/sys/dev/agp/agp.c Tue May 22 10:59:26 2012(r235782) @@ -239,7 +239,8 @@ agp_generic_attach(device_t dev) if (memsize = agp_max[i][0]) break; } - if (i == agp_max_size) i = agp_max_size - 1; + if (i == agp_max_size) + i = agp_max_size - 1; sc-as_maxmem = agp_max[i][1] 20U; /* @@ -803,6 +804,13 @@ agp_unbind_user(device_t dev, agp_unbind } static int +agp_chipset_flush(device_t dev) +{ + + return (AGP_CHIPSET_FLUSH(dev)); +} + +static int agp_open(struct cdev *kdev, int oflags, int devtype, struct thread *td) { device_t dev = kdev-si_drv1; @@ -869,6 +877,8 @@ agp_ioctl(struct cdev *kdev, u_long cmd, case AGPIOC_UNBIND: return agp_unbind_user(dev, (agp_unbind *)data); + case AGPIOC_CHIPSET_FLUSH: + return agp_chipset_flush(dev); } return EINVAL; Modified: head/sys/dev/agp/agp_i810.c == --- head/sys/dev/agp/agp_i810.c Tue May 22 10:54:42 2012(r235781) +++ head/sys/dev/agp/agp_i810.c Tue May 22 10:59:26 2012(r235782) @@ -1,8 +1,12 @@ /*- * Copyright (c) 2000 Doug Rabson * Copyright (c) 2000 Ruslan Ermilov + * Copyright (c) 2011 The FreeBSD Foundation * All rights reserved. * + * Portions of this software were developed by Konstantin Belousov + * under sponsorship from the FreeBSD Foundation. + * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: @@ -28,6 +32,9 @@ /* * Fixes for 830/845G support: David Dawes da...@xfree86.org * 852GM/855GM/865G support added by David Dawes da...@xfree86.org + * + * This is generic Intel GTT handling code, morphed from the AGP + * bridge code. */ #include sys/cdefs.h @@ -35,10 +42,17 @@ __FBSDID($FreeBSD$); #include opt_bus.h +#if 0 +#defineKTR_AGP_I810KTR_DEV +#else +#defineKTR_AGP_I8100 +#endif + #include sys/param.h #include sys/systm.h #include sys/malloc.h #include sys/kernel.h +#include sys/ktr.h #include sys/module.h #include sys/bus.h #include sys/lock.h @@ -47,8 +61,10 @@ __FBSDID($FreeBSD$); #include dev/agp/agppriv.h #include dev/agp/agpreg.h +#include dev/agp/agp_i810.h #include dev/pci/pcivar.h #include dev/pci/pcireg.h +#include dev/pci/pci_private.h #include vm/vm.h #include vm/vm_object.h @@ -63,6 +79,88 @@ __FBSDID($FreeBSD$); MALLOC_DECLARE(M_AGP); +struct agp_i810_match; + +static int agp_i810_check_active(device_t bridge_dev); +static int agp_i830_check_active(device_t bridge_dev); +static int agp_i915_check_active(device_t bridge_dev); +static int agp_sb_check_active(device_t bridge_dev); + +static void agp_82852_set_desc(device_t dev, +const struct agp_i810_match *match); +static void agp_i810_set_desc(device_t dev, const struct agp_i810_match *match); + +static void agp_i810_dump_regs(device_t dev); +static void agp_i830_dump_regs(device_t dev); +static void agp_i855_dump_regs(device_t dev); +static void agp_i915_dump_regs(device_t dev); +static void agp_i965_dump_regs(device_t dev); +static void agp_sb_dump_regs(device_t dev); + +static int agp_i810_get_stolen_size(device_t dev); +static int agp_i830_get_stolen_size(device_t dev); +static int agp_i915_get_stolen_size(device_t dev); +static int agp_sb_get_stolen_size(device_t dev); + +static int agp_i810_get_gtt_mappable_entries(device_t dev); +static int agp_i830_get_gtt_mappable_entries(device_t dev); +static int agp_i915_get_gtt_mappable_entries(device_t dev); + +static int agp_i810_get_gtt_total_entries(device_t dev); +static int agp_i965_get_gtt_total_entries(device_t dev); +static int agp_gen5_get_gtt_total_entries(device_t dev); +static int agp_sb_get_gtt_total_entries(device_t dev); + +static int agp_i810_install_gatt(device_t dev); +static int agp_i830_install_gatt(device_t dev); + +static void agp_i810_deinstall_gatt(device_t dev); +static void agp_i830_deinstall_gatt(device_t dev); + +static
svn commit: r235783 - in head/sys/dev/drm2: . i915
Author: kib Date: Tue May 22 11:07:44 2012 New Revision: 235783 URL: http://svn.freebsd.org/changeset/base/235783 Log: Add the code for new Intel GPU driver, which supports GEM, KMS and works with new generations of GPUs (IronLake, SandyBridge and supposedly IvyBridge). The driver is not connected to the build yet. Sponsored by: The FreeBSD Foundation MFC after:1 week Added: head/sys/dev/drm2/ head/sys/dev/drm2/drm.h (contents, props changed) head/sys/dev/drm2/drmP.h (contents, props changed) head/sys/dev/drm2/drm_agpsupport.c (contents, props changed) head/sys/dev/drm2/drm_atomic.h (contents, props changed) head/sys/dev/drm2/drm_auth.c (contents, props changed) head/sys/dev/drm2/drm_bufs.c (contents, props changed) head/sys/dev/drm2/drm_context.c (contents, props changed) head/sys/dev/drm2/drm_crtc.c (contents, props changed) head/sys/dev/drm2/drm_crtc.h (contents, props changed) head/sys/dev/drm2/drm_crtc_helper.c (contents, props changed) head/sys/dev/drm2/drm_crtc_helper.h (contents, props changed) head/sys/dev/drm2/drm_dma.c (contents, props changed) head/sys/dev/drm2/drm_dp_helper.h (contents, props changed) head/sys/dev/drm2/drm_dp_iic_helper.c (contents, props changed) head/sys/dev/drm2/drm_drawable.c (contents, props changed) head/sys/dev/drm2/drm_drv.c (contents, props changed) head/sys/dev/drm2/drm_edid.c (contents, props changed) head/sys/dev/drm2/drm_edid.h (contents, props changed) head/sys/dev/drm2/drm_edid_modes.h (contents, props changed) head/sys/dev/drm2/drm_fb_helper.c (contents, props changed) head/sys/dev/drm2/drm_fb_helper.h (contents, props changed) head/sys/dev/drm2/drm_fops.c (contents, props changed) head/sys/dev/drm2/drm_fourcc.h (contents, props changed) head/sys/dev/drm2/drm_gem.c (contents, props changed) head/sys/dev/drm2/drm_gem_names.c (contents, props changed) head/sys/dev/drm2/drm_gem_names.h (contents, props changed) head/sys/dev/drm2/drm_hashtab.c (contents, props changed) head/sys/dev/drm2/drm_hashtab.h (contents, props changed) head/sys/dev/drm2/drm_internal.h (contents, props changed) head/sys/dev/drm2/drm_ioctl.c (contents, props changed) head/sys/dev/drm2/drm_irq.c (contents, props changed) head/sys/dev/drm2/drm_linux_list.h (contents, props changed) head/sys/dev/drm2/drm_linux_list_sort.c (contents, props changed) head/sys/dev/drm2/drm_lock.c (contents, props changed) head/sys/dev/drm2/drm_memory.c (contents, props changed) head/sys/dev/drm2/drm_mm.c (contents, props changed) head/sys/dev/drm2/drm_mm.h (contents, props changed) head/sys/dev/drm2/drm_mode.h (contents, props changed) head/sys/dev/drm2/drm_modes.c (contents, props changed) head/sys/dev/drm2/drm_pci.c (contents, props changed) head/sys/dev/drm2/drm_pciids.h (contents, props changed) head/sys/dev/drm2/drm_sarea.h (contents, props changed) head/sys/dev/drm2/drm_scatter.c (contents, props changed) head/sys/dev/drm2/drm_sman.c (contents, props changed) head/sys/dev/drm2/drm_sman.h (contents, props changed) head/sys/dev/drm2/drm_stub.c (contents, props changed) head/sys/dev/drm2/drm_sysctl.c (contents, props changed) head/sys/dev/drm2/drm_vm.c (contents, props changed) head/sys/dev/drm2/i915/ head/sys/dev/drm2/i915/i915_debug.c (contents, props changed) head/sys/dev/drm2/i915/i915_dma.c (contents, props changed) head/sys/dev/drm2/i915/i915_drm.h (contents, props changed) head/sys/dev/drm2/i915/i915_drv.c (contents, props changed) head/sys/dev/drm2/i915/i915_drv.h (contents, props changed) head/sys/dev/drm2/i915/i915_gem.c (contents, props changed) head/sys/dev/drm2/i915/i915_gem_evict.c (contents, props changed) head/sys/dev/drm2/i915/i915_gem_execbuffer.c (contents, props changed) head/sys/dev/drm2/i915/i915_gem_gtt.c (contents, props changed) head/sys/dev/drm2/i915/i915_gem_tiling.c (contents, props changed) head/sys/dev/drm2/i915/i915_irq.c (contents, props changed) head/sys/dev/drm2/i915/i915_reg.h (contents, props changed) head/sys/dev/drm2/i915/i915_suspend.c (contents, props changed) head/sys/dev/drm2/i915/intel_bios.c (contents, props changed) head/sys/dev/drm2/i915/intel_bios.h (contents, props changed) head/sys/dev/drm2/i915/intel_crt.c (contents, props changed) head/sys/dev/drm2/i915/intel_display.c (contents, props changed) head/sys/dev/drm2/i915/intel_dp.c (contents, props changed) head/sys/dev/drm2/i915/intel_drv.h (contents, props changed) head/sys/dev/drm2/i915/intel_fb.c (contents, props changed) head/sys/dev/drm2/i915/intel_hdmi.c (contents, props changed) head/sys/dev/drm2/i915/intel_iic.c (contents, props changed) head/sys/dev/drm2/i915/intel_lvds.c (contents, props changed) head/sys/dev/drm2/i915/intel_modes.c (contents, props changed)
Re: svn commit: r235777 - head/sys/kern
On Tue, 22 May 2012, Hartmut Brandt wrote: Log: Make dumptid non-static. It is used by libkvm to detect whether this is a VNET-kernel or not. gcc used to put the static symbol into the symbol table, clang does not. This fixes the 'netstat: no namelist' error seen on clang+VNET systems. Modified: head/sys/kern/kern_shutdown.c That would be a bug in clang if it were done for static symbols generally, but here the bug seems to be that the symbol is not declared as __used. gcc does the same for a file containing only static int x;, but it is apparently confused by dumptid being initialized non-statically, although the initialization has no side effects. If dumptid were a local variable, then clang would probably warn about the variable being unused, but gcc-4.2.1 never detects such unused variables (thus code that compiles with gcc -Wunused -Werror often fails with clang). Here the initialization is to curthread-td_tid, so it isn't clear if the compiler can tell if it has no side effects. curthread() is actually __curthread(). __curthread() is now declared as __pure2, but that never worked for me with older compilers (its result wasn't cached). If the compilers can tell that the expression has no side effects, then it is another bug that they don't warn about it having no effect when it is only assigned to the apparently-unused variable dumptid. Modified: head/sys/kern/kern_shutdown.c == --- head/sys/kern/kern_shutdown.c Tue May 22 07:04:23 2012 (r235776) +++ head/sys/kern/kern_shutdown.c Tue May 22 07:23:41 2012 (r235777) @@ -151,7 +151,7 @@ static struct dumperinfo dumper;/* our /* Context information for dump-debuggers. */ static struct pcb dumppcb; /* Registers. */ -static lwpid_t dumptid;/* Thread ID. */ +lwpid_t dumptid; /* Thread ID. */ static void poweroff_wait(void *, int); static void shutdown_halt(void *junk, int howto); Now there are 3 bugs instead of 1: - the variable is declared (implicit) extern instead of static - the extern declaration is in a section for static declaration - the variable is still not declared as __used. If the compiler did a more extensive usage analysis, that looked at all object files but not at the libkvm API, then it should remove this variable anyway when it is not declared as __used. Bruce ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235780 - head/include
On Tue, 22 May 2012, Hartmut Brandt wrote: Log: Fix a compilation error with some compilers: __attribute__ requires two parenthesis for its argument, but instead of using __attribute__ directly, use the appropriate __nonnull macro from cdefs.h. This also fixes: - the style bug of hard-coding __attribute__(()) - the namespace pollution of using nonnull() instead of __nonnull__(). Modified: head/include/malloc_np.h == --- head/include/malloc_np.hTue May 22 09:27:57 2012(r235779) +++ head/include/malloc_np.hTue May 22 09:59:49 2012(r235780) @@ -55,13 +55,11 @@ int mallctlbymib(const size_t *mib, size #define ALLOCM_ERR_OOM 1 #define ALLOCM_ERR_NOT_MOVED2 -intallocm(void **ptr, size_t *rsize, size_t size, int flags) -__attribute__(nonnull(1)); +intallocm(void **ptr, size_t *rsize, size_t size, int flags) __nonnull(1); int rallocm(void **ptr, size_t *rsize, size_t size, size_t extra, -int flags) __attribute__(nonnull(1)); -intsallocm(const void *ptr, size_t *rsize, int flags) -__attribute__(nonnull(1)); -intdallocm(void *ptr, int flags) __attribute__(nonnull(1)); +int flags) __nonnull(1); +intsallocm(const void *ptr, size_t *rsize, int flags) __nonnull(1); +intdallocm(void *ptr, int flags) __nonnull(1); int nallocm(size_t *rsize, size_t size, int flags); __END_DECLS Many unfixed bugs are visible nearby, starting with the namespace pollution of parameter names in the application namespace. Bruce ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235777 - head/sys/kern
On Tue, 22 May 2012, Bruce Evans wrote: BEOn Tue, 22 May 2012, Hartmut Brandt wrote: BE BE Log: BE Make dumptid non-static. It is used by libkvm to detect whether BE this is a VNET-kernel or not. gcc used to put the static symbol into BE the symbol table, clang does not. This fixes the 'netstat: no namelist' BE error seen on clang+VNET systems. BE BE Modified: BE head/sys/kern/kern_shutdown.c BE BEThat would be a bug in clang if it were done for static symbols generally, BEbut here the bug seems to be that the symbol is not declared as __used. I don't get this. Why should a symbol declared static be in the symbol table (except for debugging purposes) ? It has internal linkage and so has a meaning only in the given file. What is the linker supposed to do with several static symbols with the same name from several object files? If several files declared static dumptids, which one would kldsym be supposed to return? harti BE BEgcc does the same for a file containing only static int x;, but it BEis apparently confused by dumptid being initialized non-statically, BEalthough the initialization has no side effects. If dumptid were a BElocal variable, then clang would probably warn about the variable being BEunused, but gcc-4.2.1 never detects such unused variables (thus code BEthat compiles with gcc -Wunused -Werror often fails with clang). Here BEthe initialization is to curthread-td_tid, so it isn't clear if the BEcompiler can tell if it has no side effects. curthread() is actually BE__curthread(). __curthread() is now declared as __pure2, but that BEnever worked for me with older compilers (its result wasn't cached). BEIf the compilers can tell that the expression has no side effects, BEthen it is another bug that they don't warn about it having no effect BEwhen it is only assigned to the apparently-unused variable dumptid. BE BE Modified: head/sys/kern/kern_shutdown.c BE == BE --- head/sys/kern/kern_shutdown.c Tue May 22 07:04:23 2012 BE (r235776) BE +++ head/sys/kern/kern_shutdown.c Tue May 22 07:23:41 2012 BE (r235777) BE @@ -151,7 +151,7 @@ static struct dumperinfo dumper; /* our BE BE /* Context information for dump-debuggers. */ BE static struct pcb dumppcb; /* Registers. */ BE -static lwpid_t dumptid; /* Thread ID. */ BE +lwpid_t dumptid; /* Thread ID. */ BE BE static void poweroff_wait(void *, int); BE static void shutdown_halt(void *junk, int howto); BE BENow there are 3 bugs instead of 1: BE- the variable is declared (implicit) extern instead of static BE- the extern declaration is in a section for static declaration BE- the variable is still not declared as __used. If the compiler did BE a more extensive usage analysis, that looked at all object files but BE not at the libkvm API, then it should remove this variable anyway BE when it is not declared as __used. BE BEBruce BE BE ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235723 - in head: contrib/byacc tools/regression/usr.bin/yacc usr.bin/yacc usr.bin/yacc/test
On 2012-05-22 12:22, Dag-Erling Smørgrav wrote: Baptiste Daroussin b...@freebsd.org writes: Log: Import byacc from invisible island, it brings us lots of compatibilities with bison, keeping full compatibility with our previous yacc implementation. This commit broke the build, in large part because Baptiste tested with Clang instead of GCC, and GCC generates a warning when compiling the generated code. It doesn't seem to compile with clang either, at least not as of r235777: /usr/src/bin/expr/expr.y:291:2: error: implicit declaration of function 'yyparse' is invalid in C99 [-Werror,-Wimplicit-function-declaration] yyparse(); ^ Note that it looks like the expr.c file is still generated with the system yacc, so I'm not entirely sure what's going on here... ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235739 - head/lib/libc/gen
On May 22, 2012, at 1:48 AM, Bruce Evans wrote: On Mon, 21 May 2012, Guy Helmer wrote: Log: Apply style(9) to return and switch/case statements. Reviewed by: delphij (prior version of the patch) Modified: head/lib/libc/gen/getnetgrent.c Modified: head/lib/libc/gen/getnetgrent.c == --- head/lib/libc/gen/getnetgrent.c Mon May 21 19:58:40 2012 (r235738) +++ head/lib/libc/gen/getnetgrent.c Mon May 21 21:04:29 2012 (r235739) ... @@ -311,32 +311,35 @@ _revnetgr_lookup(char* lookupdom, char* for (rot = 0; ; rot++) { switch (rot) { -case(0): snprintf(key, MAXHOSTNAMELEN, %s.%s, - str, dom?dom:lookupdom); - break; -case(1): snprintf(key, MAXHOSTNAMELEN, %s.*, - str); - break; -case(2): snprintf(key, MAXHOSTNAMELEN, *.%s, - dom?dom:lookupdom); - break; -case(3): snprintf(key, MAXHOSTNAMELEN, *.*); - break; -default: return(0); +case(0): +snprintf(key, MAXHOSTNAMELEN, %s.%s, str, +dom ? dom : lookupdom); +break; +case(1): +snprintf(key, MAXHOSTNAMELEN, %s.*, str); +break; +case(2): +snprintf(key, MAXHOSTNAMELEN, *.%s, +dom ? dom : lookupdom); +break; +case(3): +snprintf(key, MAXHOSTNAMELEN, *.*); +break; Thanks, but a fuller application would have removed the obfuscatory parentheses that make case() look like a function call... +default: return (0); ... and split the case statements after : in all cases. } y = yp_match(lookupdom, map, key, strlen(key), result, resultlen); You fixed the continuation indentation in the case statement but not here. if (y == 0) { rv = _listmatch(result, group, resultlen); free(result); -if (rv) return(1); +if (rv) return (1); Another statement not started on a new line. } else if (y != YPERR_KEY) { /* * If we get an error other than 'no * such key in map' then something is * wrong and we should stop the search. */ -return(-1); +return (-1); } } } These style bugs weren't in the CSRG version of course. The YP code added many. The most obvious ones are the case(n) and gnu-style continuation indentation. I am not sure how to best resolve the long lines in the block of code to free grp-ng_str[] elements. The indentation is quite deep at that point, and breaking short statements over multiple lines would make the code quite ugly. Would this resolve the other issues you pointed out? Index: lib/libc/gen/getnetgrent.c === --- lib/libc/gen/getnetgrent.c (revision 235784) +++ lib/libc/gen/getnetgrent.c (working copy) @@ -309,28 +309,30 @@ for (rot = 0; ; rot++) { switch (rot) { - case(0): + case 0: snprintf(key, MAXHOSTNAMELEN, %s.%s, str, dom ? dom : lookupdom); break; - case(1): + case 1: snprintf(key, MAXHOSTNAMELEN, %s.*, str); break; - case(2): + case 2: snprintf(key, MAXHOSTNAMELEN, *.%s, dom ? dom : lookupdom); break; - case(3): + case 3: snprintf(key, MAXHOSTNAMELEN, *.*); break; - default: return (0); + default: + return (0); } y = yp_match(lookupdom, map, key, strlen(key), result, -resultlen); + resultlen); if (y == 0) { rv = _listmatch(result, group, resultlen); free(result); - if (rv) return (1); + if (rv) + return (1); } else if (y != YPERR_KEY) { /* * If we get an error other than 'no @@ -418,7
Re: svn commit: r235723 - in head: contrib/byacc tools/regression/usr.bin/yacc usr.bin/yacc usr.bin/yacc/test
On Tue, May 22, 2012 at 6:32 AM, Dimitry Andric d...@freebsd.org wrote: On 2012-05-22 12:22, Dag-Erling Smørgrav wrote: Baptiste Daroussin b...@freebsd.org writes: Log: Import byacc from invisible island, it brings us lots of compatibilities with bison, keeping full compatibility with our previous yacc implementation. This commit broke the build, in large part because Baptiste tested with Clang instead of GCC, and GCC generates a warning when compiling the generated code. It doesn't seem to compile with clang either, at least not as of r235777: /usr/src/bin/expr/expr.y:291:2: error: implicit declaration of function 'yyparse' is invalid in C99 [-Werror,-Wimplicit-function-declaration] yyparse(); ^ Note that it looks like the expr.c file is still generated with the system yacc, so I'm not entirely sure what's going on here... I mentioned it in a private email, but the issue is that the yacc used during the build seems to be coming from the base system (despite the fact that it's a bootstrap tool). Once I installed yacc onto my stable-9 system at work, all of the changes that should have been checked in with this commit became apparent. Thanks, -Garrett 1022 .if ${BOOTSTRAPPING} 96 1023 _lex= usr.bin/lex 1024 _yacc= usr.bin/yacc 1025 .endif # ... 1068 bootstrap-tools: 1069 .for _tool in \ 1070 ${_clang_tblgen} \ 1071 ${_kerberos5_bootstrap_tools} \ 1072 ${_dtrace_tools} \ 1073 ${_strfile} \ 1074 ${_gperf} \ 1075 ${_groff} \ 1076 ${_ar} \ 1077 ${_dtc} \ 1078 ${_awk} \ 1079 usr.bin/lorder \ 1080 usr.bin/makewhatis \ 1081 ${_mklocale} \ 1082 usr.bin/rpcgen \ 1083 ${_sed} \ 1084 ${_lex} \ 1085 ${_yacc} \ ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235723 - in head: contrib/byacc tools/regression/usr.bin/yacc usr.bin/yacc usr.bin/yacc/test
On 2012-05-22 16:44, Garrett Cooper wrote: ... I mentioned it in a private email, but the issue is that the yacc used during the build seems to be coming from the base system (despite the fact that it's a bootstrap tool). Once I installed yacc onto my stable-9 system at work, all of the changes that should have been checked in with this commit became apparent. Thanks, -Garrett 1022 .if ${BOOTSTRAPPING} 96 Yes, Dag-Erling mentioned he will fix this particular test for head, which needs a version bump. Though I still think it might have been less churn if yacc had just continued to not declare yyparse()... :) ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235723 - in head: contrib/byacc tools/regression/usr.bin/yacc usr.bin/yacc usr.bin/yacc/test
On 2012.05.22. 17:03, Dimitry Andric wrote: ... I mentioned it in a private email, but the issue is that the yacc used during the build seems to be coming from the base system (despite the fact that it's a bootstrap tool). Once I installed yacc onto my stable-9 system at work, all of the changes that should have been checked in with this commit became apparent. Thanks, -Garrett 1022 .if ${BOOTSTRAPPING} 96 Yes, Dag-Erling mentioned he will fix this particular test for head, which needs a version bump. Maybe I'm missing some point but what if we always use yacc and lex from the source tree instead of from the base system? That would be the most logical way of doing this and whether they are built early or just in the world phase does not make that much difference, does it? Gabor ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235787 - head/sys/kern
Author: trasz Date: Tue May 22 15:58:27 2012 New Revision: 235787 URL: http://svn.freebsd.org/changeset/base/235787 Log: Fix panic with RACCT that could occur in low memory (or out of swap) situations, due to fork1() calling racct_proc_exit() without calling racct_proc_fork() first. Submitted by: Mateusz Guzik mjguzik at gmail dot com (earlier version) Reviewed by: Mateusz Guzik mjguzik at gmail dot com Modified: head/sys/kern/kern_fork.c head/sys/kern/kern_racct.c Modified: head/sys/kern/kern_fork.c == --- head/sys/kern/kern_fork.c Tue May 22 15:26:55 2012(r235786) +++ head/sys/kern/kern_fork.c Tue May 22 15:58:27 2012(r235787) @@ -939,8 +939,8 @@ fail: #ifdef MAC mac_proc_destroy(newproc); #endif -fail1: racct_proc_exit(newproc); +fail1: if (vm2 != NULL) vmspace_free(vm2); uma_zfree(proc_zone, newproc); Modified: head/sys/kern/kern_racct.c == --- head/sys/kern/kern_racct.c Tue May 22 15:26:55 2012(r235786) +++ head/sys/kern/kern_racct.c Tue May 22 15:58:27 2012(r235787) @@ -573,6 +573,9 @@ out: PROC_UNLOCK(child); PROC_UNLOCK(parent); + if (error != 0) + racct_proc_exit(child); + return (error); } ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235788 - in head: . sys/sys
Author: des Date: Tue May 22 15:59:07 2012 New Revision: 235788 URL: http://svn.freebsd.org/changeset/base/235788 Log: Bump __FreeBSD_version for the byacc import, and update _bootstrap_tools. Modified: head/Makefile.inc1 head/sys/sys/param.h Modified: head/Makefile.inc1 == --- head/Makefile.inc1 Tue May 22 15:58:27 2012(r235787) +++ head/Makefile.inc1 Tue May 22 15:59:07 2012(r235788) @@ -1021,6 +1021,9 @@ _sed= usr.bin/sed .if ${BOOTSTRAPPING} 96 _lex= usr.bin/lex +.endif + +.if ${BOOTSTRAPPING} 113 _yacc= usr.bin/yacc .endif @@ -1081,8 +1084,8 @@ bootstrap-tools: ${_mklocale} \ usr.bin/rpcgen \ ${_sed} \ -${_lex} \ ${_yacc} \ +${_lex} \ usr.bin/xinstall \ ${_gensnmptree} \ usr.sbin/config Modified: head/sys/sys/param.h == --- head/sys/sys/param.hTue May 22 15:58:27 2012(r235787) +++ head/sys/sys/param.hTue May 22 15:59:07 2012(r235788) @@ -58,7 +58,7 @@ * in the range 5 to 9. */ #undef __FreeBSD_version -#define __FreeBSD_version 112 /* Master, propagated to newvers */ +#define __FreeBSD_version 113 /* Master, propagated to newvers */ /* * __FreeBSD_kernel__ indicates that this system uses the kernel of FreeBSD, ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235723 - in head: contrib/byacc tools/regression/usr.bin/yacc usr.bin/yacc usr.bin/yacc/test
Gabor Kovesdan ga...@freebsd.org writes: Maybe I'm missing some point but what if we always use yacc and lex from the source tree instead of from the base system? That would be the most logical way of doing this and whether they are built early or just in the world phase does not make that much difference, does it? Bootstrap tools are built twice, so we try to only build them when we know that we need them, i.e. when cross-building (for machine-dependent tools) or when building on a system that has an older, incompatible version of the tool. For lex and yacc, the latter case applies, since they are machine-independent. My final test build just completed. I just committed the version bump and the bootstrap change; Baptiste will commit the yyparse() fix later today. DES -- Dag-Erling Smørgrav - d...@des.no ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235723 - in head: contrib/byacc tools/regression/usr.bin/yacc usr.bin/yacc usr.bin/yacc/test
On Tue, May 22, 2012 at 06:01:31PM +0200, Dag-Erling Smørgrav wrote: Gabor Kovesdan ga...@freebsd.org writes: Maybe I'm missing some point but what if we always use yacc and lex from the source tree instead of from the base system? That would be the most logical way of doing this and whether they are built early or just in the world phase does not make that much difference, does it? Bootstrap tools are built twice, so we try to only build them when we know that we need them, i.e. when cross-building (for machine-dependent tools) or when building on a system that has an older, incompatible version of the tool. For lex and yacc, the latter case applies, since they are machine-independent. My final test build just completed. I just committed the version bump and the bootstrap change; Baptiste will commit the yyparse() fix later today. Should be done, sorry all for the mess, lots of lessons learned. regards, Bapt pgpbHHdOUzwaK.pgp Description: PGP signature
Re: svn commit: r235777 - head/sys/kern
On 22. May 2012, at 07:23 , Hartmut Brandt wrote: Author: harti Date: Tue May 22 07:23:41 2012 New Revision: 235777 URL: http://svn.freebsd.org/changeset/base/235777 Log: Make dumptid non-static. It is used by libkvm to detect whether this is a VNET-kernel or not. Just for clarifications - it's used to detect whether we are operating on a crash dump or not in the vnet case. gcc used to put the static symbol into the symbol table, clang does not. This fixes the 'netstat: no namelist' error seen on clang+VNET systems. Modified: head/sys/kern/kern_shutdown.c Modified: head/sys/kern/kern_shutdown.c == --- head/sys/kern/kern_shutdown.c Tue May 22 07:04:23 2012 (r235776) +++ head/sys/kern/kern_shutdown.c Tue May 22 07:23:41 2012 (r235777) @@ -151,7 +151,7 @@ static struct dumperinfo dumper; /* our /* Context information for dump-debuggers. */ static struct pcb dumppcb;/* Registers. */ -static lwpid_t dumptid; /* Thread ID. */ +lwpid_t dumptid; /* Thread ID. */ static void poweroff_wait(void *, int); static void shutdown_halt(void *junk, int howto); -- Bjoern A. Zeeb You have to have visions! It does not matter how good you are. It matters what good you do! ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235794 - head/contrib/gcc
Author: obrien Date: Tue May 22 17:11:18 2012 New Revision: 235794 URL: http://svn.freebsd.org/changeset/base/235794 Log: Record that r235793 (-objc) has been merged from vendor/gcc into HEAD. Modified: Directory Properties: head/contrib/gcc/ (props changed) ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235795 - head/sys/kern
Author: trasz Date: Tue May 22 17:30:02 2012 New Revision: 235795 URL: http://svn.freebsd.org/changeset/base/235795 Log: Don't leak locks in prison_racct_modify(). Submitted by: Mateusz Guzik mjguzik at gmail dot com MFC after:2 weeks Modified: head/sys/kern/kern_jail.c Modified: head/sys/kern/kern_jail.c == --- head/sys/kern/kern_jail.c Tue May 22 17:11:18 2012(r235794) +++ head/sys/kern/kern_jail.c Tue May 22 17:30:02 2012(r235795) @@ -4491,8 +4491,11 @@ prison_racct_modify(struct prison *pr) sx_slock(allproc_lock); sx_xlock(allprison_lock); - if (strcmp(pr-pr_name, pr-pr_prison_racct-prr_name) == 0) + if (strcmp(pr-pr_name, pr-pr_prison_racct-prr_name) == 0) { + sx_xunlock(allprison_lock); + sx_sunlock(allproc_lock); return; + } oldprr = pr-pr_prison_racct; pr-pr_prison_racct = NULL; ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235777 - head/sys/kern
On Tue, 22 May 2012, Hartmut Brandt wrote: On Tue, 22 May 2012, Bruce Evans wrote: BEOn Tue, 22 May 2012, Hartmut Brandt wrote: BE BE Log: BE Make dumptid non-static. It is used by libkvm to detect whether BE this is a VNET-kernel or not. gcc used to put the static symbol into BE the symbol table, clang does not. This fixes the 'netstat: no namelist' BE error seen on clang+VNET systems. BE BE Modified: BE head/sys/kern/kern_shutdown.c BE BEThat would be a bug in clang if it were done for static symbols generally, BEbut here the bug seems to be that the symbol is not declared as __used. I don't get this. Why should a symbol declared static be in the symbol table (except for debugging purposes) ? It must be there for debugging purposes and historical compatibility (mainly other debugging uses, including kvm). Static symbols are not really special here. The C standard doesn't even require a symbol table, and it is only implementation details and debugging and historical compatibility that require putting global symbols in symbol tables. It has internal linkage and so has a meaning only in the given file. What is the linker supposed to do with several static symbols with the same name from several object files? Same as it always did. The names are in per-object-file namespaces for the linker, so they don't conflict for linking, but they mess up primitive debuggers starting with nm. If several files declared static dumptids, which one would kldsym be supposed to return? libkvm and kldsym would be broken. I don't know of any even non-primitive debuggers that can handle this. In gdb the only way that I know of to print a non-unique variable is to display the source file that declares the variable using something like `l' on a function in that file; the variable scope is then that of the selected file. The address file.c:foo only works for displaying functions. The latter probably depends on a full symbol table. When there is no symbol table, `l' of course doesn't work; `disass foo works. disass file.c:foo of course doesn't work. disass file.o:foo should work, but doesn't, even when there is a full symbol table. When the file.*: address doesn't work, this is is initially because it misparsed as the symbol file. Actually, I know of the primitive way which works even using ddb with all symbols broken in ddb, and have had to use this several times: you find the address of things using nm or something less primitive on another system, and guess which address to use if several variables have the same name, and type it in. BTW, kernels have lots of conflicting symbols which mess up debugging using primitive debuggers like ddb. /usr/src/tools/tools/kernxref/kernxref.sh was supposed to be used to find and fix these as well as finding and fixing public variables that should be static, but it is rarely used. uniq reports removing 382 out of 28443 (non-unique) symbols in a fairly current kernel with not many features configured. Bruce ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235797 - head/contrib/gcc
Author: obrien Date: Tue May 22 18:18:06 2012 New Revision: 235797 URL: http://svn.freebsd.org/changeset/base/235797 Log: Do not incorrectly warn when printing a quad_t using %qd on 64-bit platforms. Modified: head/contrib/gcc/c-format.c Modified: head/contrib/gcc/c-format.c == --- head/contrib/gcc/c-format.c Tue May 22 17:44:01 2012(r235796) +++ head/contrib/gcc/c-format.c Tue May 22 18:18:06 2012(r235797) @@ -287,7 +287,11 @@ static const format_length_info printf_l { { h, FMT_LEN_h, STD_C89, hh, FMT_LEN_hh, STD_C99 }, { l, FMT_LEN_l, STD_C89, ll, FMT_LEN_ll, STD_C9L }, +#ifdef __LP64__ + { q, FMT_LEN_l, STD_EXT, NULL, 0, 0 }, +#else { q, FMT_LEN_ll, STD_EXT, NULL, 0, 0 }, +#endif { L, FMT_LEN_L, STD_C89, NULL, 0, 0 }, { z, FMT_LEN_z, STD_C99, NULL, 0, 0 }, { Z, FMT_LEN_z, STD_EXT, NULL, 0, 0 }, ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235799 - head/lib/libjail
Author: jamie Date: Tue May 22 18:30:32 2012 New Revision: 235799 URL: http://svn.freebsd.org/changeset/base/235799 Log: The fix in r235291 re-broke the allow.nomount case. Re-fix it by testing for the right parameter name. Modified: head/lib/libjail/jail.c Modified: head/lib/libjail/jail.c == --- head/lib/libjail/jail.c Tue May 22 18:30:14 2012(r235798) +++ head/lib/libjail/jail.c Tue May 22 18:30:32 2012(r235799) @@ -853,7 +853,7 @@ jailparam_free(struct jailparam *jp, uns static int jailparam_type(struct jailparam *jp) { - char *p, *nname; + char *p, *name, *nname; size_t miblen, desclen; int i, isarray; struct { @@ -863,7 +863,8 @@ jailparam_type(struct jailparam *jp) int mib[CTL_MAXNAME]; /* The lastjid parameter isn't real. */ - if (!strcmp(jp-jp_name, lastjid)) { + name = jp-jp_name; + if (!strcmp(name, lastjid)) { jp-jp_valuelen = sizeof(int); jp-jp_ctltype = CTLTYPE_INT | CTLFLAG_WR; return (0); @@ -872,19 +873,19 @@ jailparam_type(struct jailparam *jp) /* Find the sysctl that describes the parameter. */ mib[0] = 0; mib[1] = 3; - snprintf(desc.s, sizeof(desc.s), SJPARAM .%s, jp-jp_name); + snprintf(desc.s, sizeof(desc.s), SJPARAM .%s, name); miblen = sizeof(mib) - 2 * sizeof(int); if (sysctl(mib, 2, mib + 2, miblen, desc.s, strlen(desc.s)) 0) { if (errno != ENOENT) { snprintf(jail_errmsg, JAIL_ERRMSGLEN, - sysctl(0.3.%s): %s, jp-jp_name, strerror(errno)); + sysctl(0.3.%s): %s, name, strerror(errno)); return (-1); } /* * The parameter probably doesn't exist. But it might be * the no counterpart to a boolean. */ - nname = nononame(jp-jp_name); + nname = nononame(name); if (nname == NULL) { unknown_parameter: snprintf(jail_errmsg, JAIL_ERRMSGLEN, @@ -892,8 +893,10 @@ jailparam_type(struct jailparam *jp) errno = ENOENT; return (-1); } - snprintf(desc.s, sizeof(desc.s), SJPARAM .%s, nname); + name = alloca(strlen(nname) + 1); + strcpy(name, nname); free(nname); + snprintf(desc.s, sizeof(desc.s), SJPARAM .%s, name); miblen = sizeof(mib) - 2 * sizeof(int); if (sysctl(mib, 2, mib + 2, miblen, desc.s, strlen(desc.s)) 0) @@ -906,7 +909,7 @@ jailparam_type(struct jailparam *jp) if (sysctl(mib, (miblen / sizeof(int)) + 2, desc, desclen, NULL, 0) 0) { snprintf(jail_errmsg, JAIL_ERRMSGLEN, - sysctl(0.4.%s): %s, jp-jp_name, strerror(errno)); + sysctl(0.4.%s): %s, name, strerror(errno)); return (-1); } jp-jp_ctltype = desc.i; @@ -952,7 +955,7 @@ jailparam_type(struct jailparam *jp) if (sysctl(mib + 2, miblen / sizeof(int), desc.s, desclen, NULL, 0) 0) { snprintf(jail_errmsg, JAIL_ERRMSGLEN, - sysctl( SJPARAM .%s): %s, jp-jp_name, + sysctl( SJPARAM .%s): %s, name, strerror(errno)); return (-1); } @@ -970,7 +973,7 @@ jailparam_type(struct jailparam *jp) if (sysctl(mib + 2, miblen / sizeof(int), NULL, jp-jp_valuelen, NULL, 0) 0) { snprintf(jail_errmsg, JAIL_ERRMSGLEN, - sysctl( SJPARAM .%s): %s, jp-jp_name, + sysctl( SJPARAM .%s): %s, name, strerror(errno)); return (-1); } @@ -995,10 +998,9 @@ jailparam_type(struct jailparam *jp) sysctl(0.1): %s, strerror(errno)); return (-1); } - if (desclen == - sizeof(SJPARAM) + strlen(jp-jp_name) + 2 + if (desclen == sizeof(SJPARAM) + strlen(name) + 2 memcmp(SJPARAM ., desc.s, sizeof(SJPARAM)) == 0 - memcmp(jp-jp_name, desc.s + sizeof(SJPARAM), + memcmp(name, desc.s + sizeof(SJPARAM), desclen - sizeof(SJPARAM) - 2) == 0 desc.s[desclen - 2] == '.') goto mib_desc;
Re: svn commit: r235623 - in head/contrib/gcc: . config/rs6000 config/sparc
On Fri, May 18, 2012 at 07:02:40PM +, Pedro F. Giffuni wrote: Log: Bring in a subset of gcc fixes that were back ported to the GCC 4.1 branch and are available under GPLv2. I do not understand this -- We're using GCC 4.2.1 @r127959, but you are bringing in patches that are against 4.1. What version of GCC were the changes back ported from if not 4.2.0 or 4.2.1? BTW, these also should have been imported into ^/vendor/gcc/ and then merged into ^/head/contrib/gcc as there are commercial users of FreeBSD that have to modify their compiler to be able to consume FreeBSD source code. And provenance and licensing matters to these folks. -- -- David (obr...@freebsd.org) ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235623 - in head/contrib/gcc: . config/rs6000 config/sparc
Hi David; On 05/22/12 13:27, David O'Brien wrote: On Fri, May 18, 2012 at 07:02:40PM +, Pedro F. Giffuni wrote: Log: Bring in a subset of gcc fixes that were back ported to the GCC 4.1 branch and are available under GPLv2. I do not understand this -- We're using GCC 4.2.1 @r127959, but you are bringing in patches that are against 4.1. What version of GCC were the changes back ported from if not 4.2.0 or 4.2.1? Most of them originated in gcc 4.3 branch and were backported to gcc 4.2.2 under the GPLv3. Those patches were further backported to gcc 4.1.x which continued to be released under GPLv2. Technically I cannot take patches from the gcc 4.2.x where x =2 so I took them from the gcc 4.1 branch. BTW, these also should have been imported into ^/vendor/gcc/ and then merged into ^/head/contrib/gcc as there are commercial users of FreeBSD that have to modify their compiler to be able to consume FreeBSD source code. And provenance and licensing matters to these folks. This was discussed in private with mm@, as he brought some GPLv2 changes from the gcc 4.3 branch. We don't carry gcc 4.3 or 4.1 in the vendors area and the code from those branches that we can merge is relatively small. (We do keep a changelog for gcc4.3 changes for reference.) FWIW, we also merged the complete libstdc++ from gcc 4.2.4 as libstdc++ remained under LGPLv2 for all the gcc 4.2 branch. Pedro. ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235801 - head/sys/net80211
Author: adrian Date: Tue May 22 19:37:12 2012 New Revision: 235801 URL: http://svn.freebsd.org/changeset/base/235801 Log: Fix some corner cases in the ieee80211_send_bar() handling. * If the first call succeeded but failed to transmit, a timer would reschedule it via bar_timeout(). Unfortunately bar_timeout() didn't check the return value from the ieee80211_send_bar() reattempt and if that failed (eg the driver ic_raw_xmit() failed), it would never re-arm the timer. * If BARPEND is cleared (which ieee80211_send_bar() will do if it can't TX), then re-arming the timer isn't enough - once bar_timeout() occurs, it'll see BARPEND is 0 and not run through the rest of the routine. So when rearming the timer, also set that flag. * If the TX wasn't occuring, bar_tx_complete() wouldn't be called and the driver callback wouldn't be called either. So the driver had no idea that the BAR TX attempt had failed. In the ath(4) case, TX would stay paused. (There's no callback to indicate that BAR TX had failed or not; only a BAR TX was attempted. That's a separate, later problem.) So call the driver callback (ic_bar_response()) before the ADDBA session is torn down, so it has a chance of being notified that things didn't quite go to plan. I've verified that yes, this does suspend traffic for ath(4), retry BAR TX even if the driver is failing ic_raw_xmit(), and then eventually giving up and sending a DELBA. I'll address the out of ath_buf issue in ath(4) in a subsequent commit - this commit just fixes the edge case where any driver is (way) out of internal buffers/descriptors and fails frame TX. PR: kern/168170 Reviewed by: bschmidt MFC after:1 month Modified: head/sys/net80211/ieee80211_ht.c Modified: head/sys/net80211/ieee80211_ht.c == --- head/sys/net80211/ieee80211_ht.cTue May 22 18:31:56 2012 (r235800) +++ head/sys/net80211/ieee80211_ht.cTue May 22 19:37:12 2012 (r235801) @@ -2166,6 +2166,9 @@ ieee80211_ampdu_stop(struct ieee80211_no } } +/* XXX */ +static void bar_start_timer(struct ieee80211_tx_ampdu *tap); + static void bar_timeout(void *arg) { @@ -2184,11 +2187,34 @@ bar_timeout(void *arg) return; /* XXX ? */ if (tap-txa_attempts = ieee80211_bar_maxtries) { + struct ieee80211com *ic = ni-ni_ic; + ni-ni_vap-iv_stats.is_ampdu_bar_tx_fail++; + /* +* If (at least) the last BAR TX timeout was due to +* an ieee80211_send_bar() failures, then we need +* to make sure we notify the driver that a BAR +* TX did occur and fail. This gives the driver +* a chance to undo any queue pause that may +* have occured. +*/ + ic-ic_bar_response(ni, tap, 1); ieee80211_ampdu_stop(ni, tap, IEEE80211_REASON_TIMEOUT); } else { ni-ni_vap-iv_stats.is_ampdu_bar_tx_retry++; - ieee80211_send_bar(ni, tap, tap-txa_seqpending); + if (ieee80211_send_bar(ni, tap, tap-txa_seqpending) != 0) { + /* +* If ieee80211_send_bar() fails here, the +* timer may have stopped and/or the pending +* flag may be clear. Because of this, +* fake the BARPEND and reset the timer. +* A retransmission attempt will then occur +* during the next timeout. +*/ + /* XXX locking */ + tap-txa_flags |= IEEE80211_AGGR_BARPEND; + bar_start_timer(tap); + } } } ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235802 - head/usr.bin/minigzip
Author: delphij Date: Tue May 22 19:40:54 2012 New Revision: 235802 URL: http://svn.freebsd.org/changeset/base/235802 Log: commandline - command line MFC after:1 week Modified: head/usr.bin/minigzip/minigzip.1 Modified: head/usr.bin/minigzip/minigzip.1 == --- head/usr.bin/minigzip/minigzip.1Tue May 22 19:37:12 2012 (r235801) +++ head/usr.bin/minigzip/minigzip.1Tue May 22 19:40:54 2012 (r235802) @@ -24,7 +24,7 @@ .\ .\ $FreeBSD$ .\ -.Dd October 3, 2002 +.Dd May 22, 2012 .Dt MINIGZIP 1 .Os .Sh NAME @@ -48,7 +48,7 @@ output. The default operation is compression, decompression can be selected by supplying the .Fl d -flag on the commandline. +flag on the command line. .Pp If any .Ar file ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235803 - head/sys/kern
Author: trasz Date: Tue May 22 19:43:20 2012 New Revision: 235803 URL: http://svn.freebsd.org/changeset/base/235803 Log: Fix use-after-free in kern_jail_set() triggered e.g. by attempts to clear persist flag from empty persistent jail, like this: jail -c persist=1 jail -n 1 -m persist=0 Submitted by: Mateusz Guzik mjguzik at gmail dot com MFC after:2 weeks Modified: head/sys/kern/kern_jail.c Modified: head/sys/kern/kern_jail.c == --- head/sys/kern/kern_jail.c Tue May 22 19:40:54 2012(r235802) +++ head/sys/kern/kern_jail.c Tue May 22 19:43:20 2012(r235803) @@ -1811,6 +1811,16 @@ kern_jail_set(struct thread *td, struct } } +#ifdef RACCT + if (!created) { + sx_sunlock(allprison_lock); + prison_racct_modify(pr); + sx_slock(allprison_lock); + } +#endif + + td-td_retval[0] = pr-pr_id; + /* * Now that it is all there, drop the temporary reference from existing * prisons. Or add a reference to newly created persistent prisons @@ -1832,12 +1842,6 @@ kern_jail_set(struct thread *td, struct sx_sunlock(allprison_lock); } -#ifdef RACCT - if (!created) - prison_racct_modify(pr); -#endif - - td-td_retval[0] = pr-pr_id; goto done_errmsg; done_deref_locked: ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235804 - head/sys/dev/ath
Author: adrian Date: Tue May 22 19:50:21 2012 New Revision: 235804 URL: http://svn.freebsd.org/changeset/base/235804 Log: Re-up the TX ath_buf limit from 128 to 512. I'll have to leave this high for now, until I've done some significant surgery with how ath_bufs (and descriptors) are handled. This should significantly cut down on the opportunities for a full TX queue hanging traffic. I'll continue making things work though; I'm mostly doing this for users. :) Modified: head/sys/dev/ath/if_athvar.h Modified: head/sys/dev/ath/if_athvar.h == --- head/sys/dev/ath/if_athvar.hTue May 22 19:43:20 2012 (r235803) +++ head/sys/dev/ath/if_athvar.hTue May 22 19:50:21 2012 (r235804) @@ -47,7 +47,7 @@ * 802.11n requires more TX and RX buffers to do AMPDU. */ #ifdef ATH_ENABLE_11N -#defineATH_TXBUF 128 +#defineATH_TXBUF 512 #defineATH_RXBUF 512 #endif ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r234482 - in head/sys: fs/msdosfs fs/nfsserver kern sys
2012/4/22 Pawel Jakub Dawidek p...@freebsd.org: On Fri, Apr 20, 2012 at 06:50:44AM +, Kirk McKusick wrote: Author: mckusick Date: Fri Apr 20 06:50:44 2012 New Revision: 234482 URL: http://svn.freebsd.org/changeset/base/234482 Log: This change creates a new list of active vnodes associated with a mount point. Active vnodes are those with a non-zero use or hold count, e.g., those vnodes that are not on the free list. Note that this list is in addition to the list of all the vnodes associated with a mount point. To avoid adding another set of linkage pointers to the vnode structure, the active list uses the existing linkage pointers used by the free list (previously named v_freelist, now renamed v_actfreelist). This update adds the MNT_VNODE_FOREACH_ACTIVE interface that loops over just the active vnodes associated with a mount point (typically less than 1% of the vnodes associated with the mount point). [...] @@ -1099,6 +1128,14 @@ insmntque1(struct vnode *vp, struct moun VNASSERT(mp-mnt_nvnodelistsize = 0, vp, (neg mount point vnode list size)); mp-mnt_nvnodelistsize++; + KASSERT((vp-v_iflag VI_ACTIVE) == 0, + (Activating already active vnode)); + vp-v_iflag |= VI_ACTIVE; + mtx_lock(vnode_free_list_mtx); + TAILQ_INSERT_HEAD(mp-mnt_activevnodelist, vp, v_actfreelist); + mp-mnt_activevnodelistsize++; + mtx_unlock(vnode_free_list_mtx); + VI_UNLOCK(vp); MNT_IUNLOCK(mp); return (0); } Now, for every vnode that is activated, it has to go through global mutex, which seems like scalability issue to me. With ZFS it is typical to have a lot of file systems and this global mutex was not needed before (well, it was needed, but only to get vnode from the free list). If we require vnode interlock to be held during v_actfreelist manipulation then why can't we use interlock+vnode_free_list_mtx when operating on the free list and interlock+per-mountpoint-lock when operating on mnt_activevnodelist? I think this is the better idea for this case and it should really be fixed. However, note that a per-mount lock here is far from being ideal as you would contest a lot on the mountpoint if you have a lot of activated vnodes. Anyway the approach you propose doesn't introduce any scalability difference than what we already had in place for pre-234482. Also, it would be good to implement things like: mtx_assert(MNT_MTX(mp), MA_OWNED); by providing appropriate wrappers as we do in vnode interface. Thanks, Attilio -- Peace can only be achieved by understanding - A. Einstein ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235797 - head/contrib/gcc
On Tue, 22 May 2012, David E. O'Brien wrote: Log: Do not incorrectly warn when printing a quad_t using %qd on 64-bit platforms. I think I like this, since it is technically correct, and will find a different set of type mismatches. Modified: head/contrib/gcc/c-format.c == --- head/contrib/gcc/c-format.c Tue May 22 17:44:01 2012(r235796) +++ head/contrib/gcc/c-format.c Tue May 22 18:18:06 2012(r235797) @@ -287,7 +287,11 @@ static const format_length_info printf_l { { h, FMT_LEN_h, STD_C89, hh, FMT_LEN_hh, STD_C99 }, { l, FMT_LEN_l, STD_C89, ll, FMT_LEN_ll, STD_C9L }, +#ifdef __LP64__ + { q, FMT_LEN_l, STD_EXT, NULL, 0, 0 }, +#else { q, FMT_LEN_ll, STD_EXT, NULL, 0, 0 }, +#endif { L, FMT_LEN_L, STD_C89, NULL, 0, 0 }, { z, FMT_LEN_z, STD_C99, NULL, 0, 0 }, { Z, FMT_LEN_z, STD_EXT, NULL, 0, 0 }, Of course, %qd should never be used. On LP32, quad_t is long long, while on LP64, quad_t is long, so any use of %qd required messy ifdefs or casting a quad_t arg to long long to work on LP64. Now, %qd actually matches quad_t on LP64, so casting to long long is no longer needed, but anything that does it is broken and would require changing to cast to quad_t, or perhaps to omit the cast. You might find too much bugware that (1) uses %qd (2) uses it on args that don't always have type quad_t (3) casts to long long. Casting to quad_t didn't work on LP64 before, so probably nothing in FreeBSD does it. Grepping for %q in /sys finds only a few uses of %q with a few type mismatches (different mismatches before and after this commit). (I didn't grep for more compicated formats with stuff between the % and the q.): % ./compat/ndis/subr_ntoskrnl.c:printf(timer sets: %qu\n, ntoskrnl_timer_sets); % ./compat/ndis/subr_ntoskrnl.c:printf(timer reloads: %qu\n, ntoskrnl_timer_reloads); % ./compat/ndis/subr_ntoskrnl.c:printf(timer cancels: %qu\n, ntoskrnl_timer_cancels); % ./compat/ndis/subr_ntoskrnl.c:printf(timer fires: %qu\n, ntoskrnl_timer_fires); This was broken before on LP64. It now works accidentally. All these %qu formats are bogus, since the variables don't have type quad_t; they have type uint64_t. Now that %q works, quad_t's are actually easier to print that int64_t's since there is a format letter just for them. But this only helps if the variables actually have type quad_t. % ./dev/esp/ncr53c9x.c: panic(%s: lun %qx for ecb %p does not exist, __func__, % ./dev/esp/ncr53c9x.c: panic(%s: slot %d for lun %qx has %p instead of ecb This is under DIAGNOSTIC and is now broken. Again the variable doesn't have type [u_]quad_t. It has type int64_t. This is printed using the doubly- incompatible format %qx (signed type but unsigned format; int64 type but quad format letter). Then to be bug for bug compatible with old gcc, this incompatible format was cast to a different doubly-incompatible type. % ./fs/msdosfs/msdosfs_vnops.c: printf(va_blocksize %lx, va_rdev %x, va_bytes %qx, va_gen %lx\n, This is under the non-option MSDOSFS_DEBUG. It was broken before, but now works, since va_bytes actually has type u_quad_t and was not bogusly cast to unsigned long long). % ./fs/nfsclient/nfs_clstate.c: printf(lck typ=%d fst=%qd end=%qd\n, % ./fs/nfsclient/nfs_clstate.c: printf(lck typ=%d fst=%qd end=%qd\n, These are under !__FreeBSD__ ifdefs. The ifdefs are related to the type errors. The types are u_int64_t. Under FreeBSD, they are printed correctly with only 1 type error for each: they are bogusly cast to intmax_t (sign error), then printed with %ju (this matches the sign of the variables but is a sign error relative to the cast. The errors normally cancel). Under !__FreeBSD__, they are printed using %qd, without any casts. There is now a sign error in the format, and type mismatches. This would now compiler under FreeBSD despite all the type mismatches -- the sign error doesn't matter in practice; u_int64_t matches u_quad_t on all supported arches, and after your changes the logical mismatch is no longer detected by gcc. % ./geom/geom_map.c:ret = sscanf(line, search:%qi:%qi:%63c, This has more than the usual density of bugs: - scanf() is unusable but is used - %q format is used - % the variables are further from being quad_t's than usual. They are off_t's. off_t happens to have type int64_t, so this works accidentally on all supported arches. - %qi is an unusual spelling of %qd. You didn't change gcc for scanf. q for it still maps to FMT_LEN_ll. The above should never have compiled on LP64. % ./gnu/fs/xfs/FreeBSD/xfs_buf.c: printf(bread failed specvp %p blkno %qd BBTOB(len) %ld\n, The variable has type xfs_daddr_t, which is __s64, which is signed long long int in xfs/FreeBSD. This used to be compatible with %qd, but no longer is. xfs shouldn't
Re: svn commit: r235797 - head/contrib/gcc
At Wed, 23 May 2012 06:05:06 +1000 (EST), Bruce Evans wrote: On Tue, 22 May 2012, David E. O'Brien wrote: Log: Do not incorrectly warn when printing a quad_t using %qd on 64-bit platforms. I think I like this, since it is technically correct, and will find a different set of type mismatches. Modified: head/contrib/gcc/c-format.c == --- head/contrib/gcc/c-format.c Tue May 22 17:44:01 2012 (r235796) +++ head/contrib/gcc/c-format.c Tue May 22 18:18:06 2012 (r235797) @@ -287,7 +287,11 @@ static const format_length_info printf_l { { h, FMT_LEN_h, STD_C89, hh, FMT_LEN_hh, STD_C99 }, { l, FMT_LEN_l, STD_C89, ll, FMT_LEN_ll, STD_C9L }, +#ifdef __LP64__ + { q, FMT_LEN_l, STD_EXT, NULL, 0, 0 }, +#else { q, FMT_LEN_ll, STD_EXT, NULL, 0, 0 }, +#endif { L, FMT_LEN_L, STD_C89, NULL, 0, 0 }, { z, FMT_LEN_z, STD_C99, NULL, 0, 0 }, { Z, FMT_LEN_z, STD_EXT, NULL, 0, 0 }, Of course, %qd should never be used. On LP32, quad_t is long long, while on LP64, quad_t is long, so any use of %qd required messy ifdefs or casting a quad_t arg to long long to work on LP64. Now, %qd actually matches quad_t on LP64, so casting to long long is no longer needed, but anything that does it is broken and would require changing to cast to quad_t, or perhaps to omit the cast. You might find too much bugware that (1) uses %qd (2) uses it on args that don't always have type quad_t (3) casts to long long. Casting to quad_t didn't work on LP64 before, so probably nothing in FreeBSD does it. Grepping for %q in /sys finds only a few uses of %q with a few type mismatches (different mismatches before and after this commit). (I didn't grep for more compicated formats with stuff between the % and the q.): % ./compat/ndis/subr_ntoskrnl.c: printf(timer sets: %qu\n, ntoskrnl_timer_sets); % ./compat/ndis/subr_ntoskrnl.c: printf(timer reloads: %qu\n, ntoskrnl_timer_reloads); % ./compat/ndis/subr_ntoskrnl.c: printf(timer cancels: %qu\n, ntoskrnl_timer_cancels); % ./compat/ndis/subr_ntoskrnl.c: printf(timer fires: %qu\n, ntoskrnl_timer_fires); This was broken before on LP64. It now works accidentally. All these %qu formats are bogus, since the variables don't have type quad_t; they have type uint64_t. Now that %q works, quad_t's are actually easier to print that int64_t's since there is a format letter just for them. But this only helps if the variables actually have type quad_t. % ./dev/esp/ncr53c9x.c: panic(%s: lun %qx for ecb %p does not exist, __func__, % ./dev/esp/ncr53c9x.c: panic(%s: slot %d for lun %qx has %p instead of ecb This is under DIAGNOSTIC and is now broken. Again the variable doesn't have type [u_]quad_t. It has type int64_t. This is printed using the doubly- incompatible format %qx (signed type but unsigned format; int64 type but quad format letter). Then to be bug for bug compatible with old gcc, this incompatible format was cast to a different doubly-incompatible type. % ./fs/msdosfs/msdosfs_vnops.c: printf(va_blocksize %lx, va_rdev %x, va_bytes %qx, va_gen %lx\n, This is under the non-option MSDOSFS_DEBUG. It was broken before, but now works, since va_bytes actually has type u_quad_t and was not bogusly cast to unsigned long long). % ./fs/nfsclient/nfs_clstate.c: printf(lck typ=%d fst=%qd end=%qd\n, % ./fs/nfsclient/nfs_clstate.c: printf(lck typ=%d fst=%qd end=%qd\n, These are under !__FreeBSD__ ifdefs. The ifdefs are related to the type errors. The types are u_int64_t. Under FreeBSD, they are printed correctly with only 1 type error for each: they are bogusly cast to intmax_t (sign error), then printed with %ju (this matches the sign of the variables but is a sign error relative to the cast. The errors normally cancel). Under !__FreeBSD__, they are printed using %qd, without any casts. There is now a sign error in the format, and type mismatches. This would now compiler under FreeBSD despite all the type mismatches -- the sign error doesn't matter in practice; u_int64_t matches u_quad_t on all supported arches, and after your changes the logical mismatch is no longer detected by gcc. % ./geom/geom_map.c: ret = sscanf(line, search:%qi:%qi:%63c, This has more than the usual density of bugs: - scanf() is unusable but is used - %q format is used - % the variables are further from being quad_t's than usual. They are off_t's. off_t happens to have type int64_t, so this works accidentally on all supported arches. - %qi is an unusual spelling of %qd. You didn't change gcc for scanf. q for it still maps to FMT_LEN_ll. The above should never have compiled on LP64. % ./gnu/fs/xfs/FreeBSD/xfs_buf.c:
Re: svn commit: r235797 - head/contrib/gcc
On Tue, May 22, 2012 at 1:05 PM, Bruce Evans b...@optusnet.com.au wrote: On Tue, 22 May 2012, David E. O'Brien wrote: Log: Do not incorrectly warn when printing a quad_t using %qd on 64-bit platforms. I think I like this, since it is technically correct, and will find a different set of type mismatches. We run with the following at Isilon, which is somewhat bogus because it allows a bit of sloppiness in types, but is also terribly convenient since it means no casting on printf arguments is needed: --- contrib/gcc/c-format.c 2012-05-22 14:08:23.538266746 -0700 +++ /data/sb/head/src/contrib/gcc/c-format.c2012-05-16 12:59:40.937016702 -0700 @@ -2298,10 +2570,20 @@ check_format_types (format_wanted_type * equivalent but the above test won't consider them equivalent. */ if (wanted_type == char_type_node (!pedantic || i 2) char_type_flag) continue; + + /* Isilon: FreeBSD defines int64_t (and others) as one type (e.g. long +long) on i386 and another type (e.g. long) on amd64. This prevents +the use of a common format specifier. Treat equal sized integer types +as equivalent. */ + if (TREE_CODE (wanted_type) == INTEGER_TYPE + TREE_CODE (cur_type) == INTEGER_TYPE + int_size_in_bytes (wanted_type) == int_size_in_bytes (cur_type)) +continue; + /* Now we have a type mismatch. */ format_type_warning (types-name, format_start, format_length, wanted_type, types-pointer_count, types-wanted_type_name, orig_cur_type, arg_num); } If there's no objections, I (or David or anyone else) can commit to the FreeBSD repository. Cheers, matthew ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235797 - head/contrib/gcc
On May 22, 2012, at 2:07 PM, m...@freebsd.org wrote: On Tue, May 22, 2012 at 1:05 PM, Bruce Evans b...@optusnet.com.au wrote: On Tue, 22 May 2012, David E. O'Brien wrote: Log: Do not incorrectly warn when printing a quad_t using %qd on 64-bit platforms. I think I like this, since it is technically correct, and will find a different set of type mismatches. We run with the following at Isilon, which is somewhat bogus because it allows a bit of sloppiness in types, but is also terribly convenient since it means no casting on printf arguments is needed: --- contrib/gcc/c-format.c 2012-05-22 14:08:23.538266746 -0700 +++ /data/sb/head/src/contrib/gcc/c-format.c2012-05-16 12:59:40.937016702 -0700 @@ -2298,10 +2570,20 @@ check_format_types (format_wanted_type * equivalent but the above test won't consider them equivalent. */ if (wanted_type == char_type_node (!pedantic || i 2) char_type_flag) continue; + + /* Isilon: FreeBSD defines int64_t (and others) as one type (e.g. long +long) on i386 and another type (e.g. long) on amd64. This prevents +the use of a common format specifier. Treat equal sized integer types +as equivalent. */ + if (TREE_CODE (wanted_type) == INTEGER_TYPE + TREE_CODE (cur_type) == INTEGER_TYPE + int_size_in_bytes (wanted_type) == int_size_in_bytes (cur_type)) +continue; + /* Now we have a type mismatch. */ format_type_warning (types-name, format_start, format_length, wanted_type, types-pointer_count, types-wanted_type_name, orig_cur_type, arg_num); } If there's no objections, I (or David or anyone else) can commit to the FreeBSD repository. I think such would make support for external (i.e. non-FreeBSD) toolchains even more problematic. Our format extensions are a big hurdle already. I think we should not tweak the compiler. -- Marcel Moolenaar mar...@xcllnt.net ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235816 - head/sys/dev/bce
Author: yongari Date: Wed May 23 01:20:25 2012 New Revision: 235816 URL: http://svn.freebsd.org/changeset/base/235816 Log: Make IPMI work in the bce driver even when the interface is configured down. Formerly, IPMI communication was lost whenever the interface was not up. The reason was that the BCE_EMAC_MODE register was not configured with the correct media settings. There are two parts to the fix. First, resetting the chip in bce_reset() causes the BCE_EMAC_MODE register to be initialized to a default value that does not necessarily correspond to the actual media settings. The fix implemented here is a bit of a hack. Ideally, at the end of bce_reset() we would poll the PHY to determine the negotiated media, and then we would set the BCE_EMAC_MODE register accordingly. That is difficult, since the PHY is abstracted behind the MII layer and is not supposed to be queried directly from the MAC driver. Instead, we read the BCE_EMAC_MODE register at the beginning of bce_reset() and then restore its media bits to their original values before returning. If IPMI is up and running, then the link is already established and the BCE_EMAC_MODE register is already set appropriately when bce_reset() is called. If IPMI is not running, no harm is done by preserving the BCE_EMAC_MODE settings. The driver will set the register properly once the interface is configured up and link is established. Second, bce_miibus_statchg() is sometimes called when the link is down. In that case, the reported media settings are invalid. Formerly, the driver used them anyway to setup the BCE_EMAC_MODE register. We now avoid changing any MAC registers unless link is active and the reported media settings are valid. Submitted by: jdp Tested by:jdp MFC after:5 days Modified: head/sys/dev/bce/if_bce.c Modified: head/sys/dev/bce/if_bce.c == --- head/sys/dev/bce/if_bce.c Wed May 23 00:46:19 2012(r235815) +++ head/sys/dev/bce/if_bce.c Wed May 23 01:20:25 2012(r235816) @@ -2062,6 +2062,11 @@ bce_miibus_statchg(device_t dev) media_status = mii-mii_media_status; } + /* Ignore invalid media status. */ + if ((media_status (IFM_ACTIVE | IFM_AVALID)) != + (IFM_ACTIVE | IFM_AVALID)) + goto bce_miibus_statchg_exit; + val = REG_RD(sc, BCE_EMAC_MODE); val = ~(BCE_EMAC_MODE_PORT | BCE_EMAC_MODE_HALF_DUPLEX | BCE_EMAC_MODE_MAC_LOOP | BCE_EMAC_MODE_FORCE_LINK | @@ -2131,6 +2136,7 @@ bce_miibus_statchg(device_t dev) /* ToDo: Update watermarks in bce_init_rx_context(). */ +bce_miibus_statchg_exit: DBEXIT(BCE_VERBOSE_PHY); } @@ -4997,14 +5003,25 @@ bce_stop(struct bce_softc *sc) static int bce_reset(struct bce_softc *sc, u32 reset_code) { - u32 val; + u32 emac_mode_save, val; int i, rc = 0; + static const u32 emac_mode_mask = BCE_EMAC_MODE_PORT | + BCE_EMAC_MODE_HALF_DUPLEX | BCE_EMAC_MODE_25G; DBENTER(BCE_VERBOSE_RESET); DBPRINT(sc, BCE_VERBOSE_RESET, %s(): reset_code = 0x%08X\n, __FUNCTION__, reset_code); + /* +* If ASF/IPMI is operational, then the EMAC Mode register already +* contains appropriate values for the link settings that have +* been auto-negotiated. Resetting the chip will clobber those +* values. Save the important bits so we can restore them after +* the reset. +*/ + emac_mode_save = REG_RD(sc, BCE_EMAC_MODE) emac_mode_mask; + /* Wait for pending PCI transactions to complete. */ REG_WR(sc, BCE_MISC_ENABLE_CLR_BITS, BCE_MISC_ENABLE_CLR_BITS_TX_DMA_ENABLE | @@ -5094,6 +5111,11 @@ bce_reset(struct bce_softc *sc, u32 rese bce_fw_cap_init(sc); bce_reset_exit: + /* Restore EMAC Mode bits needed to keep ASF/IPMI running. */ + val = REG_RD(sc, BCE_EMAC_MODE); + val = (val ~emac_mode_mask) | emac_mode_save; + REG_WR(sc, BCE_EMAC_MODE, val); + DBEXIT(BCE_VERBOSE_RESET); return (rc); } ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
Re: svn commit: r235601 - head/include/protocols
On Mon, 21 May 2012, John Baldwin wrote: On Friday, May 18, 2012 11:24:36 am Gleb Kurtsou wrote: On (18/05/2012 09:41), John Baldwin wrote: On Friday, May 18, 2012 6:01:31 am Gleb Kurtsou wrote: Author: gleb Date: Fri May 18 10:01:31 2012 New Revision: 235601 URL: http://svn.freebsd.org/changeset/base/235601 Log: Don't use ino_t in dumprestore protocol definition. Since ino_t size is about to change to 64-bits, replace ino_t used in dump protocol definition with 32-bit dump_ino_t to preserve backward compatibility. At some point, it may be necessary to use spare fields in struct in order to fully support 64-bit inode numbers. Sponsored by: Google Summer of Code 2011 A question about your stat changes: did you expand dev_t to 32 bits for the AFS folks, or did you leave it as 16 bits? dev_t is already 32-bit. Changing it to 64-bit was discussed at some point and from what I recall no decision was made: http://marc.info/?t=12911947875r=1w=2 I'm going to commit preparatory changes only for now. Then publish diff for testing. We can still change dev_t to 64-bit if needed. Although I didn't work on it. Ah, it was 64-bit they asked for. If it is easy to do so, I'd favor changing it since you've already done all the hard work of rolling a new stat structure. I'd rather err on wasting 32-bits for dev_t than having to do all this over again. Hi John, Thanks for remembering this -- I'm still pretty swamped with other stuff and missed it. It would indeed be nice for us to have a 64-bit dev_t to work with. -Ben ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org
svn commit: r235821 - head/sys/dev/bge
Author: yongari Date: Wed May 23 03:35:08 2012 New Revision: 235821 URL: http://svn.freebsd.org/changeset/base/235821 Log: Don't force max payload size to 128. Root complex and Endpoint will negotiate with each other on the TLP payload size so blindly forcing the size to 128 can cause a completion error which in turn will stop device. Reported by: Geans Pin geanspin broadcom dot com MFC after:5 days Modified: head/sys/dev/bge/if_bge.c Modified: head/sys/dev/bge/if_bge.c == --- head/sys/dev/bge/if_bge.c Wed May 23 02:43:28 2012(r235820) +++ head/sys/dev/bge/if_bge.c Wed May 23 03:35:08 2012(r235821) @@ -3638,8 +3638,6 @@ bge_reset(struct bge_softc *sc) /* Clear enable no snoop and disable relaxed ordering. */ devctl = ~(PCIM_EXP_CTL_RELAXED_ORD_ENABLE | PCIM_EXP_CTL_NOSNOOP_ENABLE); - /* Set PCIE max payload size to 128. */ - devctl = ~PCIM_EXP_CTL_MAX_PAYLOAD; pci_write_config(dev, sc-bge_expcap + PCIR_EXPRESS_DEVICE_CTL, devctl, 2); /* Clear error status. */ ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to svn-src-head-unsubscr...@freebsd.org