Re: svn commit: r328914 - in head/sys: kern ufs/ffs
Hi Kirk, On Mon, Feb 5, 2018 at 4:19 PM, Kirk McKusick wrote: > Author: mckusick > Date: Tue Feb 6 00:19:46 2018 > New Revision: 328914 > URL: https://svnweb.freebsd.org/changeset/base/328914 > > Log: > ... > The second problem was that the check hash computed at the end of the > read was incorrect because the calculation of the check hash on > completion of the read was being done too soon. > > - When a read completes we had the following sequence: > > - bufdone() > -- b_ckhashcalc (calculates check hash) > -- bufdone_finish() > --- vfs_vmio_iodone() (replaces bogus pages with the cached ones) > > - When we are reading a buffer where one or more pages are already > in memory (but not all pages, or we wouldn't be doing the read), > the I/O is done with bogus_page mapped in for the pages that exist > in the VM cache. This mapping is done to avoid corrupting the > cached pages if there is any I/O overrun. The vfs_vmio_iodone() > function is responsible for replacing the bogus_page(s) with the > cached ones. But we were calculating the check hash before the > bogus_page(s) were replaced. Hence, when we were calculating the > check hash, we were partly reading from bogus_page, which means > we calculated a bad check hash (e.g., because multiple pages have > been mapped to bogus_page, so its contents are indeterminate). > > The second fix is to move the check-hash calculation from bufdone() > to bufdone_finish() after the call to vfs_vmio_iodone() so that it > computes the check hash over the correct set of pages. Does the b_iodone callback have a very similar potential problem? It is invoked in bufdone(), before bufdone_finish() and vfs_vmio_iodone(). One example b_iodone is the bdone() callback. It seems that b_iodone -> bdone() can then wake bwait()ers before any VMIO cached content has been filled in. I don't know that any specific consumers of b_iodone are currently broken, but it seems like maybe the b_iodone callback should really be in the later location as well. Best, Conrad ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328922 - head/sys/dev/etherswitch
Author: adrian
Date: Tue Feb 6 08:34:50 2018
New Revision: 328922
URL: https://svnweb.freebsd.org/changeset/base/328922
Log:
[etherswitch] add initial support for potentially configuring and fetching
the switch MAC address.
Switches that originate their own frames (eg obvious ones like Pause frames)
need a MAC address to use to send those frames from.
This API will hopefully begin to allow that to be configurable.
Modified:
head/sys/dev/etherswitch/etherswitch.h
Modified: head/sys/dev/etherswitch/etherswitch.h
==
--- head/sys/dev/etherswitch/etherswitch.h Tue Feb 6 07:50:30 2018
(r328921)
+++ head/sys/dev/etherswitch/etherswitch.h Tue Feb 6 08:34:50 2018
(r328922)
@@ -48,10 +48,12 @@ typedef struct etherswitch_info etherswitch_info_t;
#defineETHERSWITCH_CONF_FLAGS (1 << 0)
#defineETHERSWITCH_CONF_MIRROR (1 << 1)
#defineETHERSWITCH_CONF_VLAN_MODE (1 << 2)
+#defineETHERSWITCH_CONF_SWITCH_MACADDR (1 << 3)
struct etherswitch_conf {
uint32_tcmd;/* What to configure */
uint32_tvlan_mode; /* Switch VLAN mode */
+ struct ether_addr switch_macaddr; /* Switch MAC address */
};
typedef struct etherswitch_conf etherswitch_conf_t;
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328923 - head/sbin/etherswitchcfg
Author: adrian
Date: Tue Feb 6 08:35:09 2018
New Revision: 328923
URL: https://svnweb.freebsd.org/changeset/base/328923
Log:
[etherswitchcfg] print the switch MAC address if provided.
Modified:
head/sbin/etherswitchcfg/etherswitchcfg.c
Modified: head/sbin/etherswitchcfg/etherswitchcfg.c
==
--- head/sbin/etherswitchcfg/etherswitchcfg.c Tue Feb 6 08:34:50 2018
(r328922)
+++ head/sbin/etherswitchcfg/etherswitchcfg.c Tue Feb 6 08:35:09 2018
(r328923)
@@ -556,6 +556,13 @@ print_config(struct cfg *cfg)
printf("none\n");
}
}
+
+ /* Print switch MAC address. */
+ if (cfg->conf.cmd & ETHERSWITCH_CONF_SWITCH_MACADDR) {
+ printf("%s: Switch MAC address: %s\n",
+ c,
+ ether_ntoa(&cfg->conf.switch_macaddr));
+ }
}
static void
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328924 - head/sys/dev/etherswitch/arswitch
Author: adrian
Date: Tue Feb 6 08:35:49 2018
New Revision: 328924
URL: https://svnweb.freebsd.org/changeset/base/328924
Log:
[arswitch] Implement the switch MAC address fetch API.
The placeholders are here for some future "set" MAC address API.
Tested:
* AR9340 switch
* AR8327 switch
Modified:
head/sys/dev/etherswitch/arswitch/arswitch.c
head/sys/dev/etherswitch/arswitch/arswitchreg.h
head/sys/dev/etherswitch/arswitch/arswitchvar.h
Modified: head/sys/dev/etherswitch/arswitch/arswitch.c
==
--- head/sys/dev/etherswitch/arswitch/arswitch.cTue Feb 6 08:35:09
2018(r328923)
+++ head/sys/dev/etherswitch/arswitch/arswitch.cTue Feb 6 08:35:49
2018(r328924)
@@ -484,6 +484,41 @@ ar8xxx_atu_learn_default(struct arswitch_softc *sc)
*/
/*
+ * Fetch the configured switch MAC address.
+ */
+static int
+ar8xxx_hw_get_switch_macaddr(struct arswitch_softc *sc, struct ether_addr *ea)
+{
+ uint32_t ret0, ret1;
+ char *s;
+
+ s = (void *) ea;
+
+ ret0 = arswitch_readreg(sc->sc_dev, AR8X16_REG_SW_MAC_ADDR0);
+ ret1 = arswitch_readreg(sc->sc_dev, AR8X16_REG_SW_MAC_ADDR1);
+
+ s[5] = MS(ret0, AR8X16_REG_SW_MAC_ADDR0_BYTE5);
+ s[4] = MS(ret0, AR8X16_REG_SW_MAC_ADDR0_BYTE4);
+ s[3] = MS(ret1, AR8X16_REG_SW_MAC_ADDR1_BYTE3);
+ s[2] = MS(ret1, AR8X16_REG_SW_MAC_ADDR1_BYTE2);
+ s[1] = MS(ret1, AR8X16_REG_SW_MAC_ADDR1_BYTE1);
+ s[0] = MS(ret1, AR8X16_REG_SW_MAC_ADDR1_BYTE0);
+
+ return (0);
+}
+
+/*
+ * Set the switch mac address.
+ */
+static int
+ar8xxx_hw_set_switch_macaddr(struct arswitch_softc *sc,
+const struct ether_addr *ea)
+{
+
+ return (ENXIO);
+}
+
+/*
* XXX TODO: this attach routine does NOT free all memory, resources
* upon failure!
*/
@@ -527,6 +562,8 @@ arswitch_attach(device_t dev)
sc->hal.arswitch_port_vlan_setup = ar8xxx_port_vlan_setup;
sc->hal.arswitch_port_vlan_get = ar8xxx_port_vlan_get;
sc->hal.arswitch_vlan_init_hw = ar8xxx_reset_vlans;
+ sc->hal.arswitch_hw_get_switch_macaddr = ar8xxx_hw_get_switch_macaddr;
+ sc->hal.arswitch_hw_set_switch_macaddr = ar8xxx_hw_set_switch_macaddr;
sc->hal.arswitch_vlan_getvgroup = ar8xxx_getvgroup;
sc->hal.arswitch_vlan_setvgroup = ar8xxx_setvgroup;
@@ -1115,6 +1152,7 @@ static int
arswitch_getconf(device_t dev, etherswitch_conf_t *conf)
{
struct arswitch_softc *sc;
+ int ret;
sc = device_get_softc(dev);
@@ -1122,6 +1160,13 @@ arswitch_getconf(device_t dev, etherswitch_conf_t *con
conf->cmd = ETHERSWITCH_CONF_VLAN_MODE;
conf->vlan_mode = sc->vlan_mode;
+ /* Return the switch ethernet address. */
+ ret = sc->hal.arswitch_hw_get_switch_macaddr(sc,
+ &conf->switch_macaddr);
+ if (ret == 0) {
+ conf->cmd |= ETHERSWITCH_CONF_SWITCH_MACADDR;
+ }
+
return (0);
}
@@ -1139,6 +1184,8 @@ arswitch_setconf(device_t dev, etherswitch_conf_t *con
if (err != 0)
return (err);
}
+
+ /* TODO: Set the switch ethernet address. */
return (0);
}
Modified: head/sys/dev/etherswitch/arswitch/arswitchreg.h
==
--- head/sys/dev/etherswitch/arswitch/arswitchreg.h Tue Feb 6 08:35:09
2018(r328923)
+++ head/sys/dev/etherswitch/arswitch/arswitchreg.h Tue Feb 6 08:35:49
2018(r328924)
@@ -88,7 +88,20 @@
#defineAR8X16_REG_IMR 0x0014
#defineAR8X16_REG_SW_MAC_ADDR0 0x0020
+#defineAR8X16_REG_SW_MAC_ADDR0_BYTE4 BITS(8, 8)
+#defineAR8X16_REG_SW_MAC_ADDR0_BYTE4_S 8
+#defineAR8X16_REG_SW_MAC_ADDR0_BYTE5 BITS(0, 8)
+#defineAR8X16_REG_SW_MAC_ADDR0_BYTE5_S 0
+
#defineAR8X16_REG_SW_MAC_ADDR1 0x0024
+#defineAR8X16_REG_SW_MAC_ADDR1_BYTE0 BITS(24, 8)
+#defineAR8X16_REG_SW_MAC_ADDR1_BYTE0_S 24
+#defineAR8X16_REG_SW_MAC_ADDR1_BYTE1 BITS(16, 8)
+#defineAR8X16_REG_SW_MAC_ADDR1_BYTE1_S 16
+#defineAR8X16_REG_SW_MAC_ADDR1_BYTE2 BITS(8, 8)
+#defineAR8X16_REG_SW_MAC_ADDR1_BYTE2_S 8
+#defineAR8X16_REG_SW_MAC_ADDR1_BYTE3 BITS(0, 8)
+#defineAR8X16_REG_SW_MAC_ADDR1_BYTE3_S 0
#defineAR8X16_REG_FLOOD_MASK 0x002c
#defineAR8X16_FLOOD_MASK_BCAST_TO_CPU (1 << 26)
Modified: head/sys/dev/etherswitch/arswitch/arswitchvar.h
==
--- head/sys/dev/etherswitch/arswitch/arswitchvar.h Tue Feb 6 08:35:09
2018(r328923)
+++ head/sys/dev/etherswitch/arswitch/arswitchvar.h Tue Feb 6 08:35:49
2018(r328924)
@@ -99,6 +99,11 @@ struct arswitch_softc {
int (* arswitch_hw_setup) (struct arswitch
Re: svn commit: r328916 - in head/sys: kern vm
> On 6 Feb 2018, at 04:16, Gleb Smirnoff wrote: > > Author: glebius > Date: Tue Feb 6 04:16:00 2018 > New Revision: 328916 > URL: https://svnweb.freebsd.org/changeset/base/328916 > > Log: > Followup on r302393 by cperciva, improving calculation of boot pages required > for UMA startup. > > o Introduce another stage of UMA startup, which is entered after >vm_page_startup() finishes. After this stage we don't yet enable buckets, >but we can ask VM for pages. Rename stages to meaningful names while here. >New list of stages: BOOT_COLD, BOOT_STRAPPED, BOOT_PAGEALLOC, BOOT_BUCKETS, >BOOT_RUNNING. >Enabling page alloc earlier allows us to dramatically reduce number of >boot pages required. What is more important number of zones becomes >consistent across different machines, as no MD allocations are done before >the BOOT_PAGEALLOC stage. Now only UMA internal zones actually need to use >startup_alloc(), however that may change, so vm_page_startup() provides >its need for early zones as argument. > o Introduce uma_startup_count() function, to avoid code duplication. The >functions calculates sizes of zones zone and kegs zone, and calculates how >many pages UMA will need to bootstrap. >It counts not only of zone structures, but also of kegs, slabs and hashes. > o Hide uma_startup_foo() declarations from public file. > o Provide several DIAGNOSTIC printfs on boot_pages usage. > o Bugfix: when calculating zone of zones size use (mp_maxid + 1) instead of >mp_ncpus. Use resulting number not only in the size argument to zone_ctor() >but also as args.size. With this I’m getting "panic: UMA: Increase vm.boot_pages” on an arm64 simulator with 4GB of memory. I can increase vm.boot_pages manually, however this isn’t useful in the long term. Andrew ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
On Tue, Feb 06, 2018 at 04:16:00AM +, Gleb Smirnoff wrote: > Author: glebius > Date: Tue Feb 6 04:16:00 2018 > New Revision: 328916 > URL: https://svnweb.freebsd.org/changeset/base/328916 > > Log: > Followup on r302393 by cperciva, improving calculation of boot pages > required > for UMA startup. > > o Introduce another stage of UMA startup, which is entered after > vm_page_startup() finishes. After this stage we don't yet enable buckets, > but we can ask VM for pages. Rename stages to meaningful names while here. > New list of stages: BOOT_COLD, BOOT_STRAPPED, BOOT_PAGEALLOC, > BOOT_BUCKETS, > BOOT_RUNNING. > Enabling page alloc earlier allows us to dramatically reduce number of > boot pages required. What is more important number of zones becomes > consistent across different machines, as no MD allocations are done before > the BOOT_PAGEALLOC stage. Now only UMA internal zones actually need to use > startup_alloc(), however that may change, so vm_page_startup() provides > its need for early zones as argument. > o Introduce uma_startup_count() function, to avoid code duplication. The > functions calculates sizes of zones zone and kegs zone, and calculates how > many pages UMA will need to bootstrap. > It counts not only of zone structures, but also of kegs, slabs and hashes. > o Hide uma_startup_foo() declarations from public file. > o Provide several DIAGNOSTIC printfs on boot_pages usage. > o Bugfix: when calculating zone of zones size use (mp_maxid + 1) instead of > mp_ncpus. Use resulting number not only in the size argument to > zone_ctor() > but also as args.size. > > Reviewed by:imp, gallatin (earlier version) > Differential Revision: https://reviews.freebsd.org/D14054 > > Modified: > head/sys/kern/kern_malloc.c > head/sys/vm/uma.h > head/sys/vm/uma_core.c > head/sys/vm/uma_int.h > head/sys/vm/vm_page.c > > Modified: head/sys/kern/kern_malloc.c > == > --- head/sys/kern/kern_malloc.c Tue Feb 6 02:13:44 2018 > (r328915) i386 also doesn't boot: FreeBSD 12.0-CURRENT #1 r328924: Tue Feb 6 11:51:17 CET 2018 p...@x4.osted.lan:/usr/src/sys/i386/compile/PHO i386 FreeBSD clang version 6.0.0 (branches/release_60 321788) (based on LLVM 6.0.0) WARNING: WITNESS option enabled, expect reduced performance. WARNING: DIAGNOSTIC option enabled, expect reduced performance. Entering uma_startup with 4 boot pages configured startup_alloc from "UMA Kegs", 3 boot pages left startup_alloc from "UMA Zones", 2 boot pages left startup_alloc from "UMA Hash", 1 boot pages left Entering uma_startup1 with 0 boot pages left kernel trap 12 with interrupts disabled Fatal trap 12: page fault while in kernel mode cpuid = 0; apic id = 00 fault virtual address = 0x19c fault code = supervisor read, page not present instruction pointer = 0x20:0xc0cda3f8 stack pointer = 0x28:0xc2022aac frame pointer = 0x28:0xc2022adc code segment= base 0x0, limit 0xf, type 0x1b = DPL 0, pres 1, def32 1, gran 1 processor eflags= resume, IOPL = 0 current process = 0 () [ thread pid 0 tid 0 ] Stopped at vmem_alloc+0x88:cmpl%esi,0x19c(%edi) db> x/s version version:FreeBSD 12.0-CURRENT #1 r328924: Tue Feb 6 11:51:17 CET 2018\012p...@x4.osted.lan:/usr/src/sys/i386/compile/PHO\012 db> show registers cs0x20 ds0x28 es0x28 fs 0x8 gs0x28 ss0x28 eax 0xc2022af8 ecx 0xc2022af8 edx 0xc23b5000 ebx 0x2002 esp 0xc2022aac ebp 0xc2022adc esi 0x1000 edi 0 eip 0xc0cda3f8 vmem_alloc+0x88 efl0x10002 vmem_alloc+0x88:cmpl%esi,0x19c(%edi) db> bt Tracing pid 0 tid 0 td 0xc1be3f40 vmem_alloc(0,1000,2102,c2022af8,450,...) at vmem_alloc+0x88/frame 0xc2022adc kmem_malloc_domain(0,1000,102,450,c23b5000,...) at kmem_malloc_domain+0x44/frame 0xc2022b08 startup_alloc(c23b4000,1000,0,c2022b63,102,...) at startup_alloc+0xeb/frame 0xc2022b30 keg_alloc_slab(0,2,c169eb10,984,c169eb10,...) at keg_alloc_slab+0xf6/frame 0xc2022b78 keg_fetch_slab(,2,c169eb10,9e6,c23b4000,...) at keg_fetch_slab+0x10e/frame 0xc2022bd0 zone_fetch_slab(c23b4000,0,,2,ef0,...) at zone_fetch_slab+0x61/frame 0xc2022bf4 zone_import(c23b4000,c2022c50,1,,2,...) at zone_import+0x3b/frame 0xc2022c2c zone_alloc_item(,2,c1655850,c2022c94,c165585f,...) at zone_alloc_item+0x3d/frame 0xc2022c60 uma_zcreate(c165585f,1c,0,0,0,...) at uma_zcreate+0xce/frame 0xc2022cbc vmem_startup(c2157000,c0cc49fb,c12afc4b,c2022d0c,c2022d0c,...) at vmem_startup+0xb7/frame 0xc2022ce4 vm_mem_init(0,0,0,0,c18689c0,...) at vm_mem_init+0x24/frame 0xc2022d10 mi_startup
svn commit: r328925 - in head: . gnu/usr.bin/gdb gnu/usr.bin/gdb/gdbtui targets/pseudo/userland/gnu tools/build/mk
Author: bapt
Date: Tue Feb 6 11:54:20 2018
New Revision: 328925
URL: https://svnweb.freebsd.org/changeset/base/328925
Log:
Remove gdbtui, it was already not installed on every arches
only installed on arm and sparc64.
It is the only bits that keeps us having libreadline in base
The rest of gdb can be switched to libedit and will be in another
commit
Deleted:
head/gnu/usr.bin/gdb/gdbtui/
Modified:
head/ObsoleteFiles.inc
head/gnu/usr.bin/gdb/Makefile
head/targets/pseudo/userland/gnu/Makefile.depend
head/tools/build/mk/OptionalObsoleteFiles.inc
Modified: head/ObsoleteFiles.inc
==
--- head/ObsoleteFiles.inc Tue Feb 6 08:35:49 2018(r328924)
+++ head/ObsoleteFiles.inc Tue Feb 6 11:54:20 2018(r328925)
@@ -38,10 +38,11 @@
# xargs -n1 | sort | uniq -d;
# done
+# 20180206: remove gdbtui
+OLD_FILES+=usr/bin/gdbtui
# 20180201: Obsolete forth files
OLD_FILES+=boot/efi.4th
OLD_FILES+=boot/pcibios.4th
-
# 20180114: new clang import which bumps version from 5.0.1 to 6.0.0.
OLD_FILES+=usr/lib/clang/5.0.1/include/sanitizer/allocator_interface.h
OLD_FILES+=usr/lib/clang/5.0.1/include/sanitizer/asan_interface.h
Modified: head/gnu/usr.bin/gdb/Makefile
==
--- head/gnu/usr.bin/gdb/Makefile Tue Feb 6 08:35:49 2018
(r328924)
+++ head/gnu/usr.bin/gdb/Makefile Tue Feb 6 11:54:20 2018
(r328925)
@@ -5,8 +5,6 @@
SUBDIR=libgdb gdb kgdb
.if ${MK_GDB_LIBEXEC} == "no"
-SUBDIR+= gdbtui
-
.if exists(${.CURDIR}/gdbserver/reg-${MACHINE_CPUARCH}.c)
SUBDIR+=gdbserver
.endif
Modified: head/targets/pseudo/userland/gnu/Makefile.depend
==
--- head/targets/pseudo/userland/gnu/Makefile.dependTue Feb 6 08:35:49
2018(r328924)
+++ head/targets/pseudo/userland/gnu/Makefile.dependTue Feb 6 11:54:20
2018(r328925)
@@ -32,7 +32,6 @@ DIRDEPS = \
gnu/usr.bin/gdb/doc \
gnu/usr.bin/gdb/gdb \
gnu/usr.bin/gdb/gdbserver \
- gnu/usr.bin/gdb/gdbtui \
gnu/usr.bin/gdb/kgdb \
gnu/usr.bin/gdb/libgdb \
gnu/usr.bin/gperf \
Modified: head/tools/build/mk/OptionalObsoleteFiles.inc
==
--- head/tools/build/mk/OptionalObsoleteFiles.inc Tue Feb 6 08:35:49
2018(r328924)
+++ head/tools/build/mk/OptionalObsoleteFiles.inc Tue Feb 6 11:54:20
2018(r328925)
@@ -2460,7 +2460,6 @@ OLD_FILES+=usr/share/man/man1/gcov.1.gz
.if ${MK_GDB} == no || ${MK_GDB_LIBEXEC} == yes
OLD_FILES+=usr/bin/gdb
OLD_FILES+=usr/bin/gdbserver
-OLD_FILES+=usr/bin/gdbtui
OLD_FILES+=usr/bin/kgdb
OLD_FILES+=usr/share/man/man1/gdb.1.gz
OLD_FILES+=usr/share/man/man1/gdbserver.1.gz
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328926 - in head/gnu/usr.bin/gdb: . gdb kgdb
Author: bapt
Date: Tue Feb 6 12:12:44 2018
New Revision: 328926
URL: https://svnweb.freebsd.org/changeset/base/328926
Log:
Switch to use libedit instead of readline
Modified:
head/gnu/usr.bin/gdb/Makefile.inc
head/gnu/usr.bin/gdb/gdb/Makefile
head/gnu/usr.bin/gdb/kgdb/Makefile
Modified: head/gnu/usr.bin/gdb/Makefile.inc
==
--- head/gnu/usr.bin/gdb/Makefile.inc Tue Feb 6 11:54:20 2018
(r328925)
+++ head/gnu/usr.bin/gdb/Makefile.inc Tue Feb 6 12:12:44 2018
(r328926)
@@ -13,11 +13,9 @@ BMAKE_BU= ${BMAKE_ROOT}/binutils
CNTRB_BU= ${SRCTOP}/contrib/binutils
CNTRB_GDB= ${SRCTOP}/contrib/gdb
-CNTRB_RL= ${SRCTOP}/contrib/libreadline
OBJ_BU= ${OBJTOP}/gnu/usr.bin/binutils
OBJ_GDB= ${OBJTOP}/gnu/usr.bin/gdb
-OBJ_RL= ${OBJTOP}/gnu/lib/libreadline/readline
# These assignments duplicate much of the functionality of
# MACHINE_CPUARCH, but there's no easy way to export make functions...
@@ -37,7 +35,7 @@ GDB_CROSS_DEBUGGER=
.PATH: ${CNTRB_GDB}/gdb ${CNTRB_GDB}/gdb/cli ${CNTRB_GDB}/gdb/mi \
${CNTRB_GDB}/gdb/signals ${CNTRB_GDB}/gdb/tui ${TARGET_SUBDIR}
-CFLAGS+= -DHAVE_CONFIG_H -DRL_NO_COMPAT -DMI_OUT=1 -DTUI=1
+CFLAGS+= -DHAVE_CONFIG_H -DRL_NO_COMPAT -DMI_OUT=1
CFLAGS+= -DDEBUGDIR=\"${DEBUGDIR}\"
CFLAGS+= -I.
CFLAGS+= -I${TARGET_SUBDIR}
@@ -47,7 +45,7 @@ CFLAGS+= -I${CNTRB_GDB}/gdb/config
CFLAGS+= -I${CNTRB_BU}/include
CFLAGS+= -I${CNTRB_GDB}/include
CFLAGS+= -I${CNTRB_BU}/bfd
-CFLAGS+= -I${OBJ_RL:H}
+CFLAGS+= -I${SRCTOP}/lib/libedit/edit
GENSRCS+= nm.h tm.h
Modified: head/gnu/usr.bin/gdb/gdb/Makefile
==
--- head/gnu/usr.bin/gdb/gdb/Makefile Tue Feb 6 11:54:20 2018
(r328925)
+++ head/gnu/usr.bin/gdb/gdb/Makefile Tue Feb 6 12:12:44 2018
(r328926)
@@ -13,7 +13,7 @@ LDFLAGS+= -Wl,-E
DPADD= ${GDBLIBS} ${BULIBS}
LDADD= ${GDBLIBS} ${BULIBS}
-LIBADD+= m readline ncursesw gnuregex
+LIBADD+= m edit ncursesw gnuregex
.include
CFLAGS+= -DDEBUGDIR=\"${DEBUGDIR}\"
Modified: head/gnu/usr.bin/gdb/kgdb/Makefile
==
--- head/gnu/usr.bin/gdb/kgdb/Makefile Tue Feb 6 11:54:20 2018
(r328925)
+++ head/gnu/usr.bin/gdb/kgdb/Makefile Tue Feb 6 12:12:44 2018
(r328926)
@@ -10,6 +10,6 @@ GDBLIBS= ${OBJ_GDB}/libgdb/libgdb.a
DPADD= ${GDBLIBS} ${BULIBS}
LDADD= ${GDBLIBS} ${BULIBS}
-LIBADD+= m readline ncursesw gnuregex kvm
+LIBADD+= m edit ncursesw gnuregex kvm
.include
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328489 - head/sys/conf
On Mon, Feb 05, 2018 at 01:32:04PM -0800, Jason Harmening wrote: > > Can you try downloading the raw diff from > https://reviews.freebsd.org/D14143 and > applying that to src? > For this morning's update of head (from r328876 to r328924), I reverted an earlier patch to src/sys/conf/kern.post.mk and applied the above-cited D14143 -- both on my laptop (which has PORTS_MODULES=x11/nvidia-driver-340 in /etc/src.conf) and my build machine (which runs mostly-headless, and has no ports modules). No issues; everything Just Worked. Peace, david -- David H. Wolfskill da...@catwhisker.org The circus around that memo helps confirm that Mr. Trump is unfit for office. See http://www.catwhisker.org/~david/publickey.gpg for my public key. signature.asc Description: PGP signature
svn commit: r328927 - head/contrib/gdb/gdb/tui
Author: bapt
Date: Tue Feb 6 12:17:03 2018
New Revision: 328927
URL: https://svnweb.freebsd.org/changeset/base/328927
Log:
Commit forgotten change in gdb allowing to use libedit
Modified:
head/contrib/gdb/gdb/tui/tui-io.c
Modified: head/contrib/gdb/gdb/tui/tui-io.c
==
--- head/contrib/gdb/gdb/tui/tui-io.c Tue Feb 6 12:12:44 2018
(r328926)
+++ head/contrib/gdb/gdb/tui/tui-io.c Tue Feb 6 12:17:03 2018
(r328927)
@@ -397,7 +397,7 @@ static void
tui_rl_display_match_list (char **matches, int len, int max)
{
typedef int QSFUNC (const void *, const void *);
- extern int _rl_qsort_string_compare (const void*, const void*);
+ extern int _rl_qsort_string_compare (char **, char **);
extern int _rl_print_completions_horizontally;
int count, limit, printed_len;
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328928 - in head: contrib/libreadline gnu/lib gnu/lib/libreadline share/mk targets/pseudo/userland/gnu
Author: bapt
Date: Tue Feb 6 12:22:42 2018
New Revision: 328928
URL: https://svnweb.freebsd.org/changeset/base/328928
Log:
Remove libreadline from the source tree, all consumers but gdb
has been switched to libedit long ago, libreadline was built as an
internallib for a while and kept only for gdbtui which was broken using
libreadline.
Since gdb has been mostly deorbitted in all arches, gdbtui was only installed
on arm and sparc64, given it has been removed, gdb has been switched to use
libedit, no consumers are left for libreadline. Thus this removal
Deleted:
head/contrib/libreadline/
head/gnu/lib/libreadline/
Modified:
head/gnu/lib/Makefile
head/share/mk/src.libnames.mk
head/targets/pseudo/userland/gnu/Makefile.depend
Modified: head/gnu/lib/Makefile
==
--- head/gnu/lib/Makefile Tue Feb 6 12:17:03 2018(r328927)
+++ head/gnu/lib/Makefile Tue Feb 6 12:22:42 2018(r328928)
@@ -7,7 +7,6 @@ SUBDIR.${MK_DIALOG}+= libdialog
SUBDIR.${MK_GCC}+= libgcov libgomp
SUBDIR.${MK_SSP}+= libssp
SUBDIR.${MK_TESTS}+= tests
-SUBDIR.${MK_GDB}+= libreadline
.if ${MK_GNU_GREP} != "no" || ${MK_GNU_GREP_COMPAT} != "no" || \
${MK_GDB} != "no"
Modified: head/share/mk/src.libnames.mk
==
--- head/share/mk/src.libnames.mk Tue Feb 6 12:17:03 2018
(r328927)
+++ head/share/mk/src.libnames.mk Tue Feb 6 12:22:42 2018
(r328928)
@@ -44,7 +44,6 @@ _INTERNALLIBS=\
parse \
pe \
pmcstat \
- readline \
sl \
sm \
smdb \
@@ -147,7 +146,6 @@ _LIBRARIES= \
procstat \
pthread \
radius \
- readline \
regex \
roken \
rpcsec_gss \
@@ -290,7 +288,6 @@ _DP_pam+= ssh
.if ${MK_NIS} != "no"
_DP_pam+= ypclnt
.endif
-_DP_readline= ncursesw
_DP_roken= crypt
_DP_kadm5clnt= com_err krb5 roken
_DP_kadm5srv= com_err hdb krb5 roken
@@ -413,9 +410,6 @@ LIBELFTC?= ${LIBELFTCDIR}/libelftc.a
LIBPEDIR= ${OBJTOP}/lib/libpe
LIBPE?=${LIBPEDIR}/libpe.a
-
-LIBREADLINEDIR=${OBJTOP}/gnu/lib/libreadline/readline
-LIBREADLINE?= ${LIBREADLINEDIR}/libreadline.a
LIBOPENBSDDIR= ${OBJTOP}/lib/libopenbsd
LIBOPENBSD?= ${LIBOPENBSDDIR}/libopenbsd.a
Modified: head/targets/pseudo/userland/gnu/Makefile.depend
==
--- head/targets/pseudo/userland/gnu/Makefile.dependTue Feb 6 12:17:03
2018(r328927)
+++ head/targets/pseudo/userland/gnu/Makefile.dependTue Feb 6 12:22:42
2018(r328928)
@@ -11,8 +11,6 @@ DIRDEPS = \
gnu/lib/libdialog \
gnu/lib/libgcov \
gnu/lib/libgomp \
- gnu/lib/libreadline/history/doc \
- gnu/lib/libreadline/readline/doc \
gnu/lib/libregex/doc \
gnu/lib/libssp \
gnu/lib/libssp/libssp_nonshared \
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328930 - head/sys/modules/dtb/allwinner
Author: kevans Date: Tue Feb 6 14:57:03 2018 New Revision: 328930 URL: https://svnweb.freebsd.org/changeset/base/328930 Log: dtb/allwinner: Add sun7i-a20-lamobo-r1.dts (Banana Pi R1) FreeBSD boots on this board, but the ethernet switch is not currently supported, resulting in no ethernet. A U-Boot port will be added once the ethernet switch is at least basically supported, but we add its DTS to the build here to lower the barrier-to-boot while work is underway. Modified: head/sys/modules/dtb/allwinner/Makefile Modified: head/sys/modules/dtb/allwinner/Makefile == --- head/sys/modules/dtb/allwinner/Makefile Tue Feb 6 14:04:39 2018 (r328929) +++ head/sys/modules/dtb/allwinner/Makefile Tue Feb 6 14:57:03 2018 (r328930) @@ -8,6 +8,7 @@ DTS=\ sun5i-r8-chip.dts \ sun7i-a20-bananapi.dts \ sun7i-a20-cubieboard2.dts \ + sun7i-a20-lamobo-r1.dts \ sun7i-a20-olimex-som-evb.dts \ sun7i-a20-pcduino3.dts \ sun8i-a83t-bananapi-m3.dts \ ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328914 - in head/sys: kern ufs/ffs
On Tue, Feb 6, 2018 at 1:09 AM, Conrad Meyer wrote:
> Hi Kirk,
>
> On Mon, Feb 5, 2018 at 4:19 PM, Kirk McKusick
> wrote:
> > Author: mckusick
> > Date: Tue Feb 6 00:19:46 2018
> > New Revision: 328914
> > URL: https://svnweb.freebsd.org/changeset/base/328914
> >
> > Log:
> > ...
> > The second problem was that the check hash computed at the end of the
> > read was incorrect because the calculation of the check hash on
> > completion of the read was being done too soon.
> >
> > - When a read completes we had the following sequence:
> >
> > - bufdone()
> > -- b_ckhashcalc (calculates check hash)
> > -- bufdone_finish()
> > --- vfs_vmio_iodone() (replaces bogus pages with the cached ones)
> >
> > - When we are reading a buffer where one or more pages are already
> > in memory (but not all pages, or we wouldn't be doing the read),
> > the I/O is done with bogus_page mapped in for the pages that exist
> > in the VM cache. This mapping is done to avoid corrupting the
> > cached pages if there is any I/O overrun. The vfs_vmio_iodone()
> > function is responsible for replacing the bogus_page(s) with the
> > cached ones. But we were calculating the check hash before the
> > bogus_page(s) were replaced. Hence, when we were calculating the
> > check hash, we were partly reading from bogus_page, which means
> > we calculated a bad check hash (e.g., because multiple pages have
> > been mapped to bogus_page, so its contents are indeterminate).
> >
> > The second fix is to move the check-hash calculation from bufdone()
> > to bufdone_finish() after the call to vfs_vmio_iodone() so that it
> > computes the check hash over the correct set of pages.
>
> Does the b_iodone callback have a very similar potential problem? It
> is invoked in bufdone(), before bufdone_finish() and
> vfs_vmio_iodone().
>
> One example b_iodone is the bdone() callback. It seems that b_iodone
> -> bdone() can then wake bwait()ers before any VMIO cached content has
> been filled in.
>
> I don't know that any specific consumers of b_iodone are currently
> broken, but it seems like maybe the b_iodone callback should really be
> in the later location as well.
>
It looks to me like this part of bufdone_finish()
if (bp->b_flags & B_VMIO) {
/*
* Set B_CACHE if the op was a normal read and no error
* occurred. B_CACHE is set for writes in the b*write()
* routines.
*/
if (bp->b_iocmd == BIO_READ &&
!(bp->b_flags & (B_INVAL|B_NOCACHE)) &&
!(bp->b_ioflags & BIO_ERROR))
bp->b_flags |= B_CACHE;
vfs_vmio_iodone(bp);
}
belongs before the callback to b_iodone() rather than in bufdone_finish().
It appears that bufdone_finish() isn't called elsewhere, despite being
non-static.
I'm curious why this is...
Warner
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328931 - head/share/man/man8
Author: amdmi3 (ports committer)
Date: Tue Feb 6 15:30:17 2018
New Revision: 328931
URL: https://svnweb.freebsd.org/changeset/base/328931
Log:
- Document new ${name}_limits rc.conf option
Approved by: cy
MFC after:2 weeks (along with 328331 which introduced this option)
Differential Revision:https://reviews.freebsd.org/D14028
Modified:
head/share/man/man8/rc.subr.8
Modified: head/share/man/man8/rc.subr.8
==
--- head/share/man/man8/rc.subr.8 Tue Feb 6 14:57:03 2018
(r328930)
+++ head/share/man/man8/rc.subr.8 Tue Feb 6 15:30:17 2018
(r328931)
@@ -551,7 +551,7 @@ is mounted.
A list of environment variables to run
.Va command
with.
-This will be passed as arguments to
+This will be passed as arguments to the
.Xr env 1
utility.
.It Va ${name}_fib
@@ -583,6 +583,13 @@ as.
Only supported after
.Pa /usr
is mounted.
+.It Va ${name}_limits
+.Xr limits 1
+to apply to
+.Va command .
+This will be passed as arguments to the
+.Xr limits 1
+utility.
.It Va ${name}_oomprotect
.Xr protect 1
.Va command
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328933 - head/usr.bin/find
Author: arichardson
Date: Tue Feb 6 15:41:26 2018
New Revision: 328933
URL: https://svnweb.freebsd.org/changeset/base/328933
Log:
Allow compiling usr.bin/find on Linux and Mac
When building FreeBSD the makefiles invoke find with various flags such as
`-s` that aren't supported in the native /usr/bin/find. To fix this I
build the FreeBSD version of find and use that when crossbuilding.
Inserting lots if #ifdefs in the code is rather ugly but I don't see a
better solution.
Reviewed By: brooks (mentor)
Approved By: jhb (mentor)
Differential Revision: https://reviews.freebsd.org/D13306
Modified:
head/usr.bin/find/Makefile
head/usr.bin/find/find.h
head/usr.bin/find/function.c
head/usr.bin/find/ls.c
head/usr.bin/find/operator.c
head/usr.bin/find/option.c
Modified: head/usr.bin/find/Makefile
==
--- head/usr.bin/find/Makefile Tue Feb 6 15:41:15 2018(r328932)
+++ head/usr.bin/find/Makefile Tue Feb 6 15:41:26 2018(r328933)
@@ -7,6 +7,7 @@ PROG= find
SRCS= find.c function.c ls.c main.c misc.c operator.c option.c \
getdate.y
YFLAGS=
+CFLAGS.clang+= -Werror=undef
NO_WMISSING_VARIABLE_DECLARATIONS=
Modified: head/usr.bin/find/find.h
==
--- head/usr.bin/find/find.hTue Feb 6 15:41:15 2018(r328932)
+++ head/usr.bin/find/find.hTue Feb 6 15:41:26 2018(r328933)
@@ -36,7 +36,32 @@
*/
#include
+#include
+#include
+/*
+ * We need to build find during the bootstrap stage when building on a
+ * non-FreeBSD system. Linux does not have the st_flags and st_birthtime
+ * members in struct stat so we need to omit support for tests that depend
+ * on these members. This works fine since none of these flags are used
+ * during the build of world and kernel.
+ */
+#ifdef UF_SETTABLE
+#define HAVE_STRUCT_STAT_ST_FLAGS 1
+#else
+#define HAVE_STRUCT_STAT_ST_FLAGS 0
+#endif
+#if defined(st_birthtime) || defined(st_birthtimespec)
+#define HAVE_STRUCT_STAT_ST_BIRTHTIME 1
+#else
+#define HAVE_STRUCT_STAT_ST_BIRTHTIME 0
+#endif
+#if defined(MFSNAMELEN) || defined(MFSTYPENAMELEN)
+#define HAVE_STRUCT_STATFS_F_FSTYPENAME 1
+#else
+#define HAVE_STRUCT_STATFS_F_FSTYPENAME 0
+#endif
+
/* forward declarations */
struct _plandata;
struct _option;
@@ -70,8 +95,10 @@ typedef struct _plandata *creat_f(struct _option *, ch
#defineF_IGNCASE 0x0001 /* iname ipath iregex */
#defineF_EXACTTIME F_IGNCASE /* -[acm]time units syntax */
#define F_EXECPLUS 0x0002 /* -exec ... {} + */
+#if HAVE_STRUCT_STAT_ST_BIRTHTIME
#defineF_TIME_B0x0004 /* one of -Btime, -Bnewer,
-newerB* */
#defineF_TIME2_B 0x0008 /* one of -newer?B */
+#endif
#define F_LINK 0x0010 /* lname or ilname */
/* node definition */
Modified: head/usr.bin/find/function.c
==
--- head/usr.bin/find/function.cTue Feb 6 15:41:15 2018
(r328932)
+++ head/usr.bin/find/function.cTue Feb 6 15:41:26 2018
(r328933)
@@ -261,9 +261,11 @@ f_Xmin(PLAN *plan, FTSENT *entry)
} else if (plan->flags & F_TIME_A) {
COMPARE((now - entry->fts_statp->st_atime +
60 - 1) / 60, plan->t_data.tv_sec);
+#if HAVE_STRUCT_STAT_ST_BIRTHTIME
} else if (plan->flags & F_TIME_B) {
COMPARE((now - entry->fts_statp->st_birthtime +
60 - 1) / 60, plan->t_data.tv_sec);
+#endif
} else {
COMPARE((now - entry->fts_statp->st_mtime +
60 - 1) / 60, plan->t_data.tv_sec);
@@ -304,8 +306,10 @@ f_Xtime(PLAN *plan, FTSENT *entry)
if (plan->flags & F_TIME_A)
xtime = entry->fts_statp->st_atime;
+#if HAVE_STRUCT_STAT_ST_BIRTHTIME
else if (plan->flags & F_TIME_B)
xtime = entry->fts_statp->st_birthtime;
+#endif
else if (plan->flags & F_TIME_C)
xtime = entry->fts_statp->st_ctime;
else
@@ -362,6 +366,7 @@ c_mXXdepth(OPTION *option, char ***argvp)
return new;
}
+#ifdef ACL_TYPE_NFS4
/*
* -acl function --
*
@@ -412,6 +417,7 @@ f_acl(PLAN *plan __unused, FTSENT *entry)
return (0);
return (1);
}
+#endif
PLAN *
c_acl(OPTION *option, char ***argvp __unused)
@@ -448,12 +454,14 @@ f_delete(PLAN *plan __unused, FTSENT *entry)
errx(1, "-delete: %s: relative path potentially not safe",
entry->fts_accpath);
+#if HAVE_STRUCT_STAT_ST_FLAGS
/* Turn off user immutable bits if running as root */
if ((entry->fts_statp->st_flags & (UF_APPEND|UF_IMMUTABLE)) &&
!(entry->fts_statp->st_flags & (SF_APPEND|SF_IMMUTABLE)) &&
geteuid() == 0)
svn commit: r328932 - head/sys/mips/mips
Author: arichardson
Date: Tue Feb 6 15:41:15 2018
New Revision: 328932
URL: https://svnweb.freebsd.org/changeset/base/328932
Log:
Make mips_postboot_fixup work when building the kernel with clang+lld
The compiler/linker can align fake_preload anyway it would like. When
building the kernel with gcc+bfd this always happened to be a multiple of 8.
When I built the kernel with clang and linked with lld fake_preload
happened to only be aligned to 4 bytes which caused a an ADDRS trap because
the compiler will emit sd instructions to store to this buffer.
Reviewed By: jhb, imp
Approved By: jhb (mentor)
Differential Revision: https://reviews.freebsd.org/D14018
Modified:
head/sys/mips/mips/machdep.c
Modified: head/sys/mips/mips/machdep.c
==
--- head/sys/mips/mips/machdep.cTue Feb 6 15:30:17 2018
(r328931)
+++ head/sys/mips/mips/machdep.cTue Feb 6 15:41:15 2018
(r328932)
@@ -383,7 +383,11 @@ mips_vector_init(void)
void
mips_postboot_fixup(void)
{
- static char fake_preload[256];
+ /*
+* We store u_long sized objects into the reload area, so the array
+* must be so aligned. The standard allows any alignment for char data.
+*/
+ static char fake_preload[256] _Alignas(_Alignof(u_long));
caddr_t preload_ptr = (caddr_t)&fake_preload[0];
size_t size = 0;
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328934 - in head: . bin/sh
Author: arichardson
Date: Tue Feb 6 15:41:35 2018
New Revision: 328934
URL: https://svnweb.freebsd.org/changeset/base/328934
Log:
Don't hardcode /usr/bin as the path for mktemp in build tools
It won't work e.g. when crossbuilding from Ubuntu Linux as mktemp is in
/bin there.
Reviewed By: bdrewery
Approved By: jhb (mentor)
Differential Revision: https://reviews.freebsd.org/D13937
Modified:
head/Makefile.inc1
head/bin/sh/mkbuiltins
head/bin/sh/mktokens
Modified: head/Makefile.inc1
==
--- head/Makefile.inc1 Tue Feb 6 15:41:26 2018(r328933)
+++ head/Makefile.inc1 Tue Feb 6 15:41:35 2018(r328934)
@@ -463,7 +463,7 @@ TMPPATH=${STRICTTMPPATH}:${PATH}
# when in the middle of installing over this system.
#
.if make(distributeworld) || make(installworld) || make(stageworld)
-INSTALLTMP!= /usr/bin/mktemp -d -u -t install
+INSTALLTMP!= mktemp -d -u -t install
.endif
.if make(stagekernel) || make(distributekernel)
Modified: head/bin/sh/mkbuiltins
==
--- head/bin/sh/mkbuiltins Tue Feb 6 15:41:26 2018(r328933)
+++ head/bin/sh/mkbuiltins Tue Feb 6 15:41:35 2018(r328934)
@@ -34,7 +34,7 @@
# @(#)mkbuiltins 8.2 (Berkeley) 5/4/95
# $FreeBSD$
-temp=`/usr/bin/mktemp -t ka`
+temp=`mktemp -t ka`
havehist=1
if [ "X$1" = "X-h" ]; then
havehist=0
Modified: head/bin/sh/mktokens
==
--- head/bin/sh/mktokensTue Feb 6 15:41:26 2018(r328933)
+++ head/bin/sh/mktokensTue Feb 6 15:41:35 2018(r328934)
@@ -38,7 +38,7 @@
# token marks the end of a list. The third column is the name to print in
# error messages.
-temp=`/usr/bin/mktemp -t ka`
+temp=`mktemp -t ka`
cat > $temp <<\!
TEOF 1 end of file
TNL0 newline
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328935 - head
Author: arichardson
Date: Tue Feb 6 15:41:45 2018
New Revision: 328935
URL: https://svnweb.freebsd.org/changeset/base/328935
Log:
crossbuild: Make the CHECK_TIME variable work on Linux
Linux /usr/bin/find doesn't understand the -mtime -0s flag.
Instead create a temporary file and compare that file's mtime to
sys/sys/param.h to check whether the clock is correct.
Reviewed By: jhb, imp
Approved By: jhb (mentor)
Differential Revision: https://reviews.freebsd.org/D14157
Modified:
head/Makefile
Modified: head/Makefile
==
--- head/Makefile Tue Feb 6 15:41:35 2018(r328934)
+++ head/Makefile Tue Feb 6 15:41:45 2018(r328935)
@@ -352,7 +352,7 @@ _guard: .PHONY
@false
STARTTIME!= LC_ALL=C date
-CHECK_TIME!= find ${.CURDIR}/sys/sys/param.h -mtime -0s ; echo
+CHECK_TIME!= cmp=`mktemp`; find ${.CURDIR}/sys/sys/param.h -newer "$$cmp" &&
rm "$$cmp"; echo
.if !empty(CHECK_TIME)
.error check your date/time: ${STARTTIME}
.endif
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328936 - in head/sys/dev: mpr mps
Author: ken
Date: Tue Feb 6 15:58:22 2018
New Revision: 328936
URL: https://svnweb.freebsd.org/changeset/base/328936
Log:
Diagnostic buffer fixes for the mps(4) and mpr(4) drivers.
In mp{r,s}_diag_register(), which is used to register diagnostic
buffers with the mp{r,s}(4) firmware, we allocate DMAable memory.
There were several issues here:
o No checking of the bus_dmamap_load() return value. If the load
failed or got deferred, mp{r,s}_diag_register() continued on as if
nothing had happened. We now check the return value and bail
out if it fails.
o No waiting for a deferred load callback. bus_dmamap_load()
calls a supplied callback when the mapping is done. This is
generally done immediately, but it can be deferred.
mp{r,s}_diag_register() did not check to see whether the callback
was already done before proceeding on. We now sleep until the
callback is done if it is deferred.
o No call to bus_dmamap_sync(... BUS_DMASYNC_PREREAD) after the
memory is allocated and loaded. This is necessary on some
platforms to synchronize host memory that is going to be updated
by a device.
Both drivers would also panic if the firmware was reinitialized while
a diagnostic buffer operation was in progress. This fixes that problem
as well. (The driver will reinitialize the firmware in various
circumstances, but the problem I ran into was that the firmware would
generate an IOC Fault due to a PCIe error.)
mp{r,s}var.h:
Add a new structure, struct mpr_busdma_context, that is
used for deferred busdma load callbacks.
Add a prototype for mp{r,s}_memaddr_wait_cb().
mp{r,s}.c:
Add a new busdma callback function, mp{r,s}_memaddr_wait_cb().
This provides synchronization for callers that want to
wait on a deferred bus_dmamap_load() callback.
mp{r,s}_user.c:
In bus_dmamap_register(), add a call to bus_dmamap_sync()
with the BUS_DMASYNC_PREREAD flag set after an allocation
is loaded.
Also, check the return value of bus_dmamap_load(). If it
fails, bail out. If it is EINPROGRESS, wait for the
callback to happen. We use an interruptible sleep (msleep
with PCATCH) and let the callback clean things up if we get
interrupted.
In mpr_diag_read_buffer() and mps_diag_read_buffer(), call
bus_dmamap_sync(..., BUS_DMASYNC_POSTREAD) before copying
the data out to make sure the data is in stable storage.
In mp{r,s}_post_fw_diag_buffer() and
mp{r,s}_release_fw_diag_buffer(), check the reply to see
whether it is NULL. It can be NULL (and the command non-NULL)
if the controller gets reinitialized while we're waiting for
the command to complete but the driver structures aren't
reallocated. The driver structures generally won't be
reallocated unless there is a firmware upgrade that changes
one of the IOCFacts.
When freeing diagnostic buffers in mp{r,s}_diag_register()
and mp{r,s}_diag_unregister(), zero/NULL out the buffer after
freeing it. This will prevent a duplicate free in some
situations.
Sponsored by: Spectra Logic
Reviewed by: mav, scottl
MFC after:1 week
Differential Revision:D13453
Modified:
head/sys/dev/mpr/mpr.c
head/sys/dev/mpr/mpr_user.c
head/sys/dev/mpr/mprvar.h
head/sys/dev/mps/mps.c
head/sys/dev/mps/mps_user.c
head/sys/dev/mps/mpsvar.h
Modified: head/sys/dev/mpr/mpr.c
==
--- head/sys/dev/mpr/mpr.c Tue Feb 6 15:41:45 2018(r328935)
+++ head/sys/dev/mpr/mpr.c Tue Feb 6 15:58:22 2018(r328936)
@@ -1183,6 +1183,42 @@ mpr_memaddr_cb(void *arg, bus_dma_segment_t *segs, int
*addr = segs[0].ds_addr;
}
+void
+mpr_memaddr_wait_cb(void *arg, bus_dma_segment_t *segs, int nsegs, int error)
+{
+ struct mpr_busdma_context *ctx;
+ int need_unload, need_free;
+
+ ctx = (struct mpr_busdma_context *)arg;
+ need_unload = 0;
+ need_free = 0;
+
+ mpr_lock(ctx->softc);
+ ctx->error = error;
+ ctx->completed = 1;
+ if ((error == 0) && (ctx->abandoned == 0)) {
+ *ctx->addr = segs[0].ds_addr;
+ } else {
+ if (nsegs != 0)
+ need_unload = 1;
+ if (ctx->abandoned != 0)
+ need_free = 1;
+ }
+ if (need_free == 0)
+ wakeup(ctx);
+
+ mpr_unlock(ctx->softc);
+
+ if (need_unload != 0) {
+ bus_dmamap_unload(ctx->buffer_dmat,
+ ctx->buffer_dmamap);
+ *ctx->addr = 0;
+ }
+
+ if (need_free != 0)
+ free(ctx, M_MPR);
+}
+
static int
mpr_alloc_queues(struct mpr_softc *sc)
{
Modified: head/sys/dev/mp
svn commit: r328937 - in head/sys/dev: mpr mps
Author: mav
Date: Tue Feb 6 16:02:25 2018
New Revision: 328937
URL: https://svnweb.freebsd.org/changeset/base/328937
Log:
Fix queue length reporting in mps(4) and mpr(4).
Both drivers were found to report CAM bigger queue depth then they really
can handle. It made them later under high load with many disks return
some of submitted requests back with CAM_REQUEUE_REQ status for later
resubmission.
Reviewed by: scottl
MFC after:1 week
Sponsored by: iXsystems, Inc.
Differential Revision:https://reviews.freebsd.org/D14215
Modified:
head/sys/dev/mpr/mpr.c
head/sys/dev/mpr/mpr_sas.c
head/sys/dev/mpr/mprvar.h
head/sys/dev/mps/mps.c
head/sys/dev/mps/mps_sas.c
head/sys/dev/mps/mpsvar.h
Modified: head/sys/dev/mpr/mpr.c
==
--- head/sys/dev/mpr/mpr.c Tue Feb 6 15:58:22 2018(r328936)
+++ head/sys/dev/mpr/mpr.c Tue Feb 6 16:02:25 2018(r328937)
@@ -397,6 +397,7 @@ mpr_resize_queues(struct mpr_softc *sc)
reqcr = MIN(reqcr, sc->facts->RequestCredit);
sc->num_reqs = prireqcr + reqcr;
+ sc->num_prireqs = prireqcr;
sc->num_replies = MIN(sc->max_replyframes + sc->max_evtframes,
sc->facts->MaxReplyDescriptorPostQueueDepth) - 1;
@@ -1507,7 +1508,7 @@ mpr_alloc_requests(struct mpr_softc *sc)
/* XXX Is a failure here a critical problem? */
if (bus_dmamap_create(sc->buffer_dmat, 0, &cm->cm_dmamap)
== 0) {
- if (i <= sc->facts->HighPriorityCredit)
+ if (i <= sc->num_prireqs)
mpr_free_high_priority_command(sc, cm);
else
mpr_free_command(sc, cm);
Modified: head/sys/dev/mpr/mpr_sas.c
==
--- head/sys/dev/mpr/mpr_sas.c Tue Feb 6 15:58:22 2018(r328936)
+++ head/sys/dev/mpr/mpr_sas.c Tue Feb 6 16:02:25 2018(r328937)
@@ -728,7 +728,7 @@ mpr_attach_sas(struct mpr_softc *sc)
{
struct mprsas_softc *sassc;
cam_status status;
- int unit, error = 0;
+ int unit, error = 0, reqs;
MPR_FUNCTRACE(sc);
mpr_dprint(sc, MPR_INIT, "%s entered\n", __func__);
@@ -758,7 +758,8 @@ mpr_attach_sas(struct mpr_softc *sc)
sc->sassc = sassc;
sassc->sc = sc;
- if ((sassc->devq = cam_simq_alloc(sc->num_reqs)) == NULL) {
+ reqs = sc->num_reqs - sc->num_prireqs - 1;
+ if ((sassc->devq = cam_simq_alloc(reqs)) == NULL) {
mpr_dprint(sc, MPR_INIT|MPR_ERROR, "Cannot allocate SIMQ\n");
error = ENOMEM;
goto out;
@@ -766,7 +767,7 @@ mpr_attach_sas(struct mpr_softc *sc)
unit = device_get_unit(sc->mpr_dev);
sassc->sim = cam_sim_alloc(mprsas_action, mprsas_poll, "mpr", sassc,
- unit, &sc->mpr_mtx, sc->num_reqs, sc->num_reqs, sassc->devq);
+ unit, &sc->mpr_mtx, reqs, reqs, sassc->devq);
if (sassc->sim == NULL) {
mpr_dprint(sc, MPR_INIT|MPR_ERROR, "Cannot allocate SIM\n");
error = EINVAL;
Modified: head/sys/dev/mpr/mprvar.h
==
--- head/sys/dev/mpr/mprvar.h Tue Feb 6 15:58:22 2018(r328936)
+++ head/sys/dev/mpr/mprvar.h Tue Feb 6 16:02:25 2018(r328937)
@@ -357,6 +357,7 @@ struct mpr_softc {
MPI2_IOC_FACTS_REPLY*facts;
int num_reqs;
+ int num_prireqs;
int num_replies;
int fqdepth;/* Free queue */
int pqdepth;/* Post queue */
Modified: head/sys/dev/mps/mps.c
==
--- head/sys/dev/mps/mps.c Tue Feb 6 15:58:22 2018(r328936)
+++ head/sys/dev/mps/mps.c Tue Feb 6 16:02:25 2018(r328937)
@@ -394,6 +394,7 @@ mps_resize_queues(struct mps_softc *sc)
reqcr = MIN(reqcr, sc->facts->RequestCredit);
sc->num_reqs = prireqcr + reqcr;
+ sc->num_prireqs = prireqcr;
sc->num_replies = MIN(sc->max_replyframes + sc->max_evtframes,
sc->facts->MaxReplyDescriptorPostQueueDepth) - 1;
@@ -1453,7 +1454,7 @@ mps_alloc_requests(struct mps_softc *sc)
/* XXX Is a failure here a critical problem? */
if (bus_dmamap_create(sc->buffer_dmat, 0, &cm->cm_dmamap) == 0)
- if (i <= sc->facts->HighPriorityCredit)
+ if (i <= sc->num_prireqs)
mps_free_high_priority_command(sc, cm);
else
mps_free_command(sc, cm);
Modified: head/sys/de
svn commit: r328938 - head/sys/geom/mirror
Author: markj
Date: Tue Feb 6 16:02:33 2018
New Revision: 328938
URL: https://svnweb.freebsd.org/changeset/base/328938
Log:
Simplify synchronization read error handling.
Since synchronization reads are performed by submitting a request to
the external mirror provider, we know that the request returns with an
error only when gmirror was unable to read a copy of the block from any
mirror. Thus, there is no need to retry the request from the
synchronization error handler.
Tested by:pho
MFC after:2 weeks
Sponsored by: Dell EMC Isilon
Modified:
head/sys/geom/mirror/g_mirror.c
Modified: head/sys/geom/mirror/g_mirror.c
==
--- head/sys/geom/mirror/g_mirror.c Tue Feb 6 16:02:25 2018
(r328937)
+++ head/sys/geom/mirror/g_mirror.c Tue Feb 6 16:02:33 2018
(r328938)
@@ -1336,9 +1336,7 @@ g_mirror_sync_request(struct g_mirror_softc *sc, struc
*/
switch (bp->bio_cmd) {
case BIO_READ: {
- struct g_mirror_disk *d;
struct g_consumer *cp;
- int readable;
KFAIL_POINT_ERROR(DEBUG_FP, g_mirror_sync_request_read,
bp->bio_error);
@@ -1349,31 +1347,17 @@ g_mirror_sync_request(struct g_mirror_softc *sc, struc
bp->bio_error);
/*
-* If there's at least one other disk from which we can
-* read the block, retry the request.
-*/
- readable = 0;
- LIST_FOREACH(d, &sc->sc_disks, d_next)
- if (d->d_state == G_MIRROR_DISK_STATE_ACTIVE &&
- !(d->d_flags & G_MIRROR_DISK_FLAG_BROKEN))
- readable++;
-
- /*
* The read error will trigger a syncid bump, so there's
* no need to do that here.
*
-* If we can retry the read from another disk, do so.
-* Otherwise, all we can do is kick out the new disk.
+* The read error handling for regular requests will
+* retry the read from all active mirrors before passing
+* the error back up, so there's no need to retry here.
*/
- if (readable == 0) {
- g_mirror_sync_request_free(disk, bp);
- g_mirror_event_send(disk,
- G_MIRROR_DISK_STATE_DISCONNECTED,
- G_MIRROR_EVENT_DONTWAIT);
- } else {
- g_mirror_sync_reinit(disk, bp, bp->bio_offset);
- goto retry_read;
- }
+ g_mirror_sync_request_free(disk, bp);
+ g_mirror_event_send(disk,
+ G_MIRROR_DISK_STATE_DISCONNECTED,
+ G_MIRROR_EVENT_DONTWAIT);
return;
}
G_MIRROR_LOGREQ(3, bp,
@@ -1429,7 +1413,6 @@ g_mirror_sync_request(struct g_mirror_softc *sc, struc
g_mirror_sync_reinit(disk, bp, sync->ds_offset);
sync->ds_offset += bp->bio_length;
-retry_read:
G_MIRROR_LOGREQ(3, bp, "Sending synchronization request.");
sync->ds_consumer->index++;
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328933 - head/usr.bin/find
On 02/06/2018 09:41, Alex Richardson wrote: > +#if HAVE_STRUCT_STAT_ST_BIRTHTIME > else if (plan->flags & F_TIME_A) > ft = entry->fts_statp->st_atim; > else if (plan->flags & F_TIME_B) > ft = entry->fts_statp->st_birthtim; > +#endif Did you intend to put st_atim inside the #if? Eric ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328939 - head/contrib/gcc/config/mips
Author: jhb Date: Tue Feb 6 17:01:10 2018 New Revision: 328939 URL: https://svnweb.freebsd.org/changeset/base/328939 Log: Use a workaround to compile the crt init functions correctly with clang. The MIPS assembly parser treats forward-declared local symbols as global symbols. This results in CALL16 relocations being used against local (private) symbols which then fail to resolve when linking binaries. Add .local to force the init and fini functions to be treated as local as a workaround. Submitted by: sbruno Sponsored by: DARPA / AFRL Modified: head/contrib/gcc/config/mips/mips.h Modified: head/contrib/gcc/config/mips/mips.h == --- head/contrib/gcc/config/mips/mips.h Tue Feb 6 16:02:33 2018 (r328938) +++ head/contrib/gcc/config/mips/mips.h Tue Feb 6 17:01:10 2018 (r328939) @@ -2721,6 +2721,7 @@ while (0) nop\n\ 1: .cpload $31\n\ .set reorder\n\ + .local " USER_LABEL_PREFIX #FUNC "\n\ jal " USER_LABEL_PREFIX #FUNC "\n\ " TEXT_SECTION_ASM_OP); #endif /* Switch to #elif when we're no longer limited by K&R C. */ @@ -2733,6 +2734,7 @@ while (0) nop\n\ 1: .set reorder\n\ .cpsetup $31, $2, 1b\n\ + .local " USER_LABEL_PREFIX #FUNC "\n\ jal " USER_LABEL_PREFIX #FUNC "\n\ " TEXT_SECTION_ASM_OP); #endif ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328940 - head/sys/vm
Author: markj Date: Tue Feb 6 17:26:11 2018 New Revision: 328940 URL: https://svnweb.freebsd.org/changeset/base/328940 Log: Delete a declaration for a variable removed in r305362. Modified: head/sys/vm/vm_page.h Modified: head/sys/vm/vm_page.h == --- head/sys/vm/vm_page.h Tue Feb 6 17:01:10 2018(r328939) +++ head/sys/vm/vm_page.h Tue Feb 6 17:26:11 2018(r328940) @@ -379,8 +379,6 @@ extern struct mtx_padalign pa_lock[]; * */ -extern int vm_page_zero_count; - extern vm_page_t vm_page_array;/* First resident page in table */ extern long vm_page_array_size;/* number of vm_page_t's */ extern long first_page;/* first physical page number */ ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
On Tue, Feb 06, 2018 at 10:54:05AM +, Andrew Turner wrote: A> > Author: glebius A> > Date: Tue Feb 6 04:16:00 2018 A> > New Revision: 328916 A> > URL: https://svnweb.freebsd.org/changeset/base/328916 A> > A> > Log: A> > Followup on r302393 by cperciva, improving calculation of boot pages required A> > for UMA startup. A> > A> > o Introduce another stage of UMA startup, which is entered after A> >vm_page_startup() finishes. After this stage we don't yet enable buckets, A> >but we can ask VM for pages. Rename stages to meaningful names while here. A> >New list of stages: BOOT_COLD, BOOT_STRAPPED, BOOT_PAGEALLOC, BOOT_BUCKETS, A> >BOOT_RUNNING. A> >Enabling page alloc earlier allows us to dramatically reduce number of A> >boot pages required. What is more important number of zones becomes A> >consistent across different machines, as no MD allocations are done before A> >the BOOT_PAGEALLOC stage. Now only UMA internal zones actually need to use A> >startup_alloc(), however that may change, so vm_page_startup() provides A> >its need for early zones as argument. A> > o Introduce uma_startup_count() function, to avoid code duplication. The A> >functions calculates sizes of zones zone and kegs zone, and calculates how A> >many pages UMA will need to bootstrap. A> >It counts not only of zone structures, but also of kegs, slabs and hashes. A> > o Hide uma_startup_foo() declarations from public file. A> > o Provide several DIAGNOSTIC printfs on boot_pages usage. A> > o Bugfix: when calculating zone of zones size use (mp_maxid + 1) instead of A> >mp_ncpus. Use resulting number not only in the size argument to zone_ctor() A> >but also as args.size. A> A> With this I’m getting "panic: UMA: Increase vm.boot_pages” on an arm64 simulator with 4GB of memory. I can increase vm.boot_pages manually, however this isn’t useful in the long term. Sorry for problems. Can you please put DIAGNOSTIC into beginning of uma_core.c, and boot with vm.boot_pages bumped, and provide the dmesg to me? -- Gleb Smirnoff ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
Peter, On Tue, Feb 06, 2018 at 12:30:17PM +0100, Peter Holm wrote: P> i386 also doesn't boot: P> P> FreeBSD 12.0-CURRENT #1 r328924: Tue Feb 6 11:51:17 CET 2018 P> p...@x4.osted.lan:/usr/src/sys/i386/compile/PHO i386 P> FreeBSD clang version 6.0.0 (branches/release_60 321788) (based on LLVM 6.0.0) P> WARNING: WITNESS option enabled, expect reduced performance. P> WARNING: DIAGNOSTIC option enabled, expect reduced performance. P> Entering uma_startup with 4 boot pages configured P> startup_alloc from "UMA Kegs", 3 boot pages left P> startup_alloc from "UMA Zones", 2 boot pages left P> startup_alloc from "UMA Hash", 1 boot pages left P> Entering uma_startup1 with 0 boot pages left P> kernel trap 12 with interrupts disabled Sorry for problems, I'm working on this. How do you think, is it a good idea to quickly add a one line change to subversion that would bump boot_pages value to arbitrary +100, to fixup the head while I'm working on a proper fix? -- Gleb Smirnoff ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328945 - head/sys/sys
Author: bz Date: Tue Feb 6 19:14:15 2018 New Revision: 328945 URL: https://svnweb.freebsd.org/changeset/base/328945 Log: Remove a trailing whitspace. Modified: head/sys/sys/bus_dma.h Modified: head/sys/sys/bus_dma.h == --- head/sys/sys/bus_dma.h Tue Feb 6 19:13:44 2018(r328944) +++ head/sys/sys/bus_dma.h Tue Feb 6 19:14:15 2018(r328945) @@ -178,7 +178,7 @@ int bus_dma_tag_create(bus_dma_tag_t parent, bus_size_ /* * Set the memory domain to be used for allocations. - * + * * Automatic for PCI devices. Must be set prior to creating maps or * allocating memory. */ ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328934 - in head: . bin/sh
> Author: arichardson
> Date: Tue Feb 6 15:41:35 2018
> New Revision: 328934
> URL: https://svnweb.freebsd.org/changeset/base/328934
>
> Log:
> Don't hardcode /usr/bin as the path for mktemp in build tools
>
> It won't work e.g. when crossbuilding from Ubuntu Linux as mktemp is in
> /bin there.
>
> Reviewed By:bdrewery
> Approved By:jhb (mentor)
> Differential Revision: https://reviews.freebsd.org/D13937
Would it be better to create the variable MKTEMP to point at
either /bin/mktemp or /usr/bin/mktemp dependent on platform,
there are reasons we use full paths in Makefiles, mostly to
stop /usr/local/bin/foo contimaton, which I believe this
change now opens up, though very slight as I dont know of
a third party mktemp binary.
> Modified:
> head/Makefile.inc1
> head/bin/sh/mkbuiltins
> head/bin/sh/mktokens
>
> Modified: head/Makefile.inc1
> ==
> --- head/Makefile.inc1Tue Feb 6 15:41:26 2018(r328933)
> +++ head/Makefile.inc1Tue Feb 6 15:41:35 2018(r328934)
> @@ -463,7 +463,7 @@ TMPPATH= ${STRICTTMPPATH}:${PATH}
> # when in the middle of installing over this system.
> #
> .if make(distributeworld) || make(installworld) || make(stageworld)
> -INSTALLTMP!= /usr/bin/mktemp -d -u -t install
> +INSTALLTMP!= mktemp -d -u -t install
> .endif
>
> .if make(stagekernel) || make(distributekernel)
>
> Modified: head/bin/sh/mkbuiltins
> ==
> --- head/bin/sh/mkbuiltinsTue Feb 6 15:41:26 2018(r328933)
> +++ head/bin/sh/mkbuiltinsTue Feb 6 15:41:35 2018(r328934)
> @@ -34,7 +34,7 @@
> #@(#)mkbuiltins 8.2 (Berkeley) 5/4/95
> # $FreeBSD$
>
> -temp=`/usr/bin/mktemp -t ka`
> +temp=`mktemp -t ka`
> havehist=1
> if [ "X$1" = "X-h" ]; then
> havehist=0
>
> Modified: head/bin/sh/mktokens
> ==
> --- head/bin/sh/mktokens Tue Feb 6 15:41:26 2018(r328933)
> +++ head/bin/sh/mktokens Tue Feb 6 15:41:35 2018(r328934)
> @@ -38,7 +38,7 @@
> # token marks the end of a list. The third column is the name to print in
> # error messages.
>
> -temp=`/usr/bin/mktemp -t ka`
> +temp=`mktemp -t ka`
> cat > $temp <<\!
> TEOF 1 end of file
> TNL 0 newline
>
>
--
Rod Grimes rgri...@freebsd.org
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
On Tue, Feb 06, 2018 at 10:28:39AM -0800, Gleb Smirnoff wrote: > Peter, > > On Tue, Feb 06, 2018 at 12:30:17PM +0100, Peter Holm wrote: > P> i386 also doesn't boot: > P> > P> FreeBSD 12.0-CURRENT #1 r328924: Tue Feb 6 11:51:17 CET 2018 > P> p...@x4.osted.lan:/usr/src/sys/i386/compile/PHO i386 > P> FreeBSD clang version 6.0.0 (branches/release_60 321788) (based on LLVM > 6.0.0) > P> WARNING: WITNESS option enabled, expect reduced performance. > P> WARNING: DIAGNOSTIC option enabled, expect reduced performance. > P> Entering uma_startup with 4 boot pages configured > P> startup_alloc from "UMA Kegs", 3 boot pages left > P> startup_alloc from "UMA Zones", 2 boot pages left > P> startup_alloc from "UMA Hash", 1 boot pages left > P> Entering uma_startup1 with 0 boot pages left > P> kernel trap 12 with interrupts disabled > > Sorry for problems, I'm working on this. > > How do you think, is it a good idea to quickly add a one line change > to subversion that would bump boot_pages value to arbitrary +100, > to fixup the head while I'm working on a proper fix? > Oh, this is not a problem for me. Increasing boot_pages does however not seem to change anything: OK set vm.boot_pages=164 OK boot /boot/kernel/kernel text=0x146eead data=0xf0da0+0x2aa460 syms=[0x4+0xf0420+0x4+0x1865f5] /boot/entropy size=0x1000 /boot/kernel/amdtemp.ko text=0x2154 data=0x164+0x4 syms=[0x4+0x600+0x4+0x6d9] loading required module 'amdsmn' /boot/kernel/amdsmn.ko text=0xe1c data=0x124+0x4 syms=[0x4+0x4a0+0x4+0x4fd] Booting... GDB: no debug ports present KDB: debugger backends: ddb KDB: current backend: ddb Copyright (c) 1992-2018 The FreeBSD Project. Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 The Regents of the University of California. All rights reserved. FreeBSD is a registered trademark of The FreeBSD Foundation. FreeBSD 12.0-CURRENT #3 r328941: Tue Feb 6 20:20:12 CET 2018 p...@x4.osted.lan:/usr/src/sys/i386/compile/PHO i386 FreeBSD clang version 6.0.0 (branches/release_60 321788) (based on LLVM 6.0.0) WARNING: WITNESS option enabled, expect reduced performance. WARNING: DIAGNOSTIC option enabled, expect reduced performance. Entering uma_startup with 164 boot pages configured startup_alloc from "UMA Kegs", 163 boot pages left startup_alloc from "UMA Zones", 162 boot pages left startup_alloc from "UMA Hash", 161 boot pages left Entering uma_startup1 with 160 boot pages left kernel trap 12 with interrupts disabled Fatal trap 12: page fault while in kernel mode cpuid = 0; apic id = 00 fault virtual address = 0x19c fault code = supervisor read, page not present instruction pointer = 0x20:0xc0cdac38 stack pointer = 0x28:0xc2022bd0 frame pointer = 0x28:0xc2022c00 code segment= base 0x0, limit 0xf, type 0x1b = DPL 0, pres 1, def32 1, gran 1 processor eflags= resume, IOPL = 0 current process = 0 () [ thread pid 0 tid 0 ] Stopped at vmem_alloc+0x88:cmpl%esi,0x19c(%edi) db> bt Tracing pid 0 tid 0 td 0xc1be3a40 vmem_alloc(0,1000,2102,c2022c1c,c169f490,...) at vmem_alloc+0x88/frame 0xc2022c00 kmem_malloc_domain(0,1000,102) at kmem_malloc_domain+0x44/frame 0xc2022c2c page_alloc(c2318000,1000,0,c2022c77,102,...) at page_alloc+0x26/frame 0xc2022c44 keg_alloc_slab(0,2,c169f490,d19,c2315b70,...) at keg_alloc_slab+0xf6/frame 0xc2022c8c uma_prealloc(c2318000,4,0,0,0,...) at uma_prealloc+0x95/frame 0xc2022cbc vmem_startup(c2157000,c0cc523b,c12b048b,c2022d0c,c2022d0c,...) at vmem_startup+0xf5/frame 0xc2022ce4 vm_mem_init(0,0,0,0,c1869340,...) at vm_mem_init+0x24/frame 0xc2022d10 mi_startup() at mi_startup+0xf7/frame 0xc2022d38 begin() at begin+0x2f db> - Peter ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328934 - in head: . bin/sh
On Tue, 2018-02-06 at 11:25 -0800, Rodney W. Grimes wrote: > > > > Author: arichardson > > Date: Tue Feb 6 15:41:35 2018 > > New Revision: 328934 > > URL: https://svnweb.freebsd.org/changeset/base/328934 > > > > Log: > > Don't hardcode /usr/bin as the path for mktemp in build tools > > > > It won't work e.g. when crossbuilding from Ubuntu Linux as mktemp > > is in > > /bin there. > > > > Reviewed By: bdrewery > > Approved By: jhb (mentor) > > Differential Revision: https://reviews.freebsd.org/D13937 > Would it be better to create the variable MKTEMP to point at > either /bin/mktemp or /usr/bin/mktemp dependent on platform, > there are reasons we use full paths in Makefiles, mostly to > stop /usr/local/bin/foo contimaton, which I believe this > change now opens up, though very slight as I dont know of > a third party mktemp binary. > I don't understand this idea of /usr/local "polluting" a system. It seems to me exactly the opposite would be the case... if I have found some 3rd party version of mktemp that I like better, it would be installed in /usr/local. If I went out of my way to install that, then naturally I WANT it to be used. To me, it's insane that the /usr/local paths are not in front of the base system paths by default, and it's even more insane that the base system works so hard to NOT use the replacements I've installed (even if I've arranged PATH so that the right versions should be used) so that I have to track down why it's using the wrong thing and apply ad-hoc fixes. -- Ian ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328949 - head/etc
Author: feld (ports committer)
Date: Tue Feb 6 20:12:05 2018
New Revision: 328949
URL: https://svnweb.freebsd.org/changeset/base/328949
Log:
Fix firstboot fs mount logic
The firstboot logic has an error which causes the filesystem to be
mounted readonly even though root_rw_mount=YES. This fixes the error to
ensure that the root filesystem is mounted rw as expected after the run
of the firstboot scripts.
Reviewed by: imp
MFC after:3 days
Differential Revision:https://reviews.freebsd.org/D14226
Modified:
head/etc/rc
Modified: head/etc/rc
==
--- head/etc/rc Tue Feb 6 19:17:40 2018(r328948)
+++ head/etc/rc Tue Feb 6 20:12:05 2018(r328949)
@@ -141,10 +141,10 @@ if [ -e ${firstboot_sentinel} ]; then
if [ -e ${firstboot_sentinel}-reboot ]; then
chflags -R 0 ${firstboot_sentinel}-reboot
rm -rf ${firstboot_sentinel}-reboot
- checkyesno root_rw_mount && mount -ur /
+ checkyesno root_rw_mount || mount -ur /
kill -INT 1
fi
- checkyesno root_rw_mount && mount -ur /
+ checkyesno root_rw_mount || mount -ur /
fi
echo ''
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328950 - in head/sys/dev: mpr mps
Author: scottl
Date: Tue Feb 6 21:01:38 2018
New Revision: 328950
URL: https://svnweb.freebsd.org/changeset/base/328950
Log:
Cache the value of the request and reply frame size since it's used quite
a bit in the normal operation of the driver. Covert it to represent bytes
instead of 32bit words. Fix what I believe to be is a bug in this respect
with the Tri-mode cards.
Sponsored by: Netflix
Modified:
head/sys/dev/mpr/mpr.c
head/sys/dev/mpr/mpr_user.c
head/sys/dev/mpr/mprvar.h
head/sys/dev/mps/mps.c
head/sys/dev/mps/mps_sas.c
head/sys/dev/mps/mps_user.c
head/sys/dev/mps/mpsvar.h
Modified: head/sys/dev/mpr/mpr.c
==
--- head/sys/dev/mpr/mpr.c Tue Feb 6 20:12:05 2018(r328949)
+++ head/sys/dev/mpr/mpr.c Tue Feb 6 21:01:38 2018(r328950)
@@ -1133,6 +1133,14 @@ mpr_send_iocinit(struct mpr_softc *sc)
MPR_FUNCTRACE(sc);
mpr_dprint(sc, MPR_INIT, "%s entered\n", __func__);
+ /* Do a quick sanity check on proper initialization */
+ if ((sc->pqdepth == 0) || (sc->fqdepth == 0) || (sc->reqframesz == 0)
+ || (sc->replyframesz == 0)) {
+ mpr_dprint(sc, MPR_INIT|MPR_ERROR,
+ "Driver not fully initialized for IOCInit\n");
+ return (EINVAL);
+ }
+
req_sz = sizeof(MPI2_IOC_INIT_REQUEST);
reply_sz = sizeof(MPI2_IOC_INIT_REPLY);
bzero(&init, req_sz);
@@ -1147,7 +1155,7 @@ mpr_send_iocinit(struct mpr_softc *sc)
init.WhoInit = MPI2_WHOINIT_HOST_DRIVER;
init.MsgVersion = htole16(MPI2_VERSION);
init.HeaderVersion = htole16(MPI2_HEADER_VERSION);
- init.SystemRequestFrameSize = htole16(sc->facts->IOCRequestFrameSize);
+ init.SystemRequestFrameSize = htole16((uint16_t)(sc->reqframesz / 4));
init.ReplyDescriptorPostQueueDepth = htole16(sc->pqdepth);
init.ReplyFreeQueueDepth = htole16(sc->fqdepth);
init.SenseBufferAddressHigh = 0;
@@ -1303,6 +1311,9 @@ mpr_alloc_replies(struct mpr_softc *sc)
{
int rsize, num_replies;
+ /* Store the reply frame size in bytes rather than as 32bit words */
+ sc->replyframesz = sc->facts->ReplyFrameSize * 4;
+
/*
* sc->num_replies should be one less than sc->fqdepth. We need to
* allocate space for sc->fqdepth replies, but only sc->num_replies
@@ -1310,7 +1321,7 @@ mpr_alloc_replies(struct mpr_softc *sc)
*/
num_replies = max(sc->fqdepth, sc->num_replies);
- rsize = sc->facts->ReplyFrameSize * num_replies * 4;
+ rsize = sc->replyframesz * num_replies;
if (bus_dma_tag_create( sc->mpr_parent_dmat,/* parent */
4, 0, /* algnmnt, boundary */
BUS_SPACE_MAXADDR_32BIT,/* lowaddr */
@@ -1344,7 +1355,10 @@ mpr_alloc_requests(struct mpr_softc *sc)
struct mpr_chain *chain;
int i, rsize, nsegs;
- rsize = sc->facts->IOCRequestFrameSize * sc->num_reqs * 4;
+ /* Store the request frame size in bytes rather than as 32bit words */
+ sc->reqframesz = sc->facts->IOCRequestFrameSize * 4;
+
+ rsize = sc->reqframesz * sc->num_reqs;
if (bus_dma_tag_create( sc->mpr_parent_dmat,/* parent */
16, 0, /* algnmnt, boundary */
BUS_SPACE_MAXADDR_32BIT,/* lowaddr */
@@ -1387,7 +1401,7 @@ mpr_alloc_requests(struct mpr_softc *sc)
MPR_MAX_CHAIN_ELEMENT_SIZE;
}
} else {
- sc->chain_frame_size = sc->facts->IOCRequestFrameSize * 4;
+ sc->chain_frame_size = sc->reqframesz;
}
rsize = sc->chain_frame_size * sc->max_chains;
if (bus_dma_tag_create( sc->mpr_parent_dmat,/* parent */
@@ -1493,10 +1507,8 @@ mpr_alloc_requests(struct mpr_softc *sc)
}
for (i = 1; i < sc->num_reqs; i++) {
cm = &sc->commands[i];
- cm->cm_req = sc->req_frames +
- i * sc->facts->IOCRequestFrameSize * 4;
- cm->cm_req_busaddr = sc->req_busaddr +
- i * sc->facts->IOCRequestFrameSize * 4;
+ cm->cm_req = sc->req_frames + i * sc->reqframesz;
+ cm->cm_req_busaddr = sc->req_busaddr + i * sc->reqframesz;
cm->cm_sense = &sc->sense_frames[i];
cm->cm_sense_busaddr = sc->sense_busaddr + i * MPR_SENSE_LEN;
cm->cm_desc.Default.SMID = i;
@@ -1621,8 +1633,7 @@ mpr_init_queues(struct mpr_softc *sc)
* Initialize all of the free queue entries.
*/
for (i = 0; i < sc->fqdepth; i++) {
- sc->free_queue[i] = sc->reply_busaddr +
- (i * sc->facts->ReplyFrameSize * 4);
+ sc->free_queue[i] = sc->reply_busaddr + (i * sc->replyframesz);
}
svn commit: r328951 - head/etc/rc.d
Author: feld (ports committer)
Date: Tue Feb 6 21:35:41 2018
New Revision: 328951
URL: https://svnweb.freebsd.org/changeset/base/328951
Log:
Refactor cleanvar to remove shell expansion vulnerability
If any process creates a directory named "-P" in /var/run or
/var/spool/lock it will cause the purgedir function to start to rm -r /.
Simplify a lot of complicated shell logic by leveraging find(1).
Reviewed by: allanjude
MFC after:3 days
Differential Revision:https://reviews.freebsd.org/D13778
Modified:
head/etc/rc.d/cleanvar
Modified: head/etc/rc.d/cleanvar
==
--- head/etc/rc.d/cleanvar Tue Feb 6 21:01:38 2018(r328950)
+++ head/etc/rc.d/cleanvar Tue Feb 6 21:35:41 2018(r328951)
@@ -19,34 +19,6 @@ stop_cmd=":"
extra_commands="reload"
reload_cmd="${name}_start"
-purgedir()
-{
- local dir file
-
- if [ $# -eq 0 ]; then
- purgedir .
- else
- for dir
- do
- (
- cd "$dir" && for file in .* *
- do
- # Skip over logging sockets
- [ -S "$file" -a "$file" = "log" ] && continue
- [ -S "$file" -a "$file" = "logpriv" ] &&
continue
- [ ."$file" = .. -o ."$file" = ... ] && continue
- if [ -d "$file" -a ! -L "$file" ]
- then
- purgedir "$file"
- else
- rm -f -- "$file"
- fi
- done
- )
- done
- fi
-}
-
cleanvar_prestart()
{
# These files must be removed only the first time this script is run
@@ -58,14 +30,17 @@ cleanvar_prestart()
cleanvar_start()
{
if [ -d /var/run -a ! -f /var/run/clean_var ]; then
- purgedir /var/run
+ # Skip over logging sockets
+ find /var/run \( -type f -or -type s ! -name log -and ! -name
logpriv \) -delete
>/var/run/clean_var
fi
if [ -d /var/spool/lock -a ! -f /var/spool/lock/clean_var ]; then
- purgedir /var/spool/lock
+ find /var/spool/lock -type f -delete
>/var/spool/lock/clean_var
fi
- rm -rf /var/spool/uucp/.Temp/*
+ if [ -d /var/spool/uucp/.Temp ]; then
+ find /var/spool/uucp/.Temp -delete
+ fi
}
load_rc_config $name
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328952 - in head/sys: kern vm
Author: glebius
Date: Tue Feb 6 22:06:59 2018
New Revision: 328952
URL: https://svnweb.freebsd.org/changeset/base/328952
Log:
Fix boot_pages calculation for machines that don't have UMA_MD_SMALL_ALLOC.
o Call uma_startup1() after initializing kmem, vmem and domains.
o Include 8 eight VM startup pages into uma_startup_count() calculation.
o Account for vmem_startup() and vm_map_startup() preallocating pages.
o Account for extra two allocations done by kmem_init() and vmem_create().
o Hardcode the place of execution of vm_radix_reserve_kva(). Using SYSINIT
allowed several other SYSINITs to sneak in before it, thus bumping
requirement for amount of boot pages.
Modified:
head/sys/kern/subr_vmem.c
head/sys/vm/vm_init.c
head/sys/vm/vm_page.c
head/sys/vm/vm_radix.c
Modified: head/sys/kern/subr_vmem.c
==
--- head/sys/kern/subr_vmem.c Tue Feb 6 21:35:41 2018(r328951)
+++ head/sys/kern/subr_vmem.c Tue Feb 6 22:06:59 2018(r328952)
@@ -72,7 +72,10 @@ __FBSDID("$FreeBSD$");
#include
#include
#include
+#include
+intvmem_startup_count(void);
+
#defineVMEM_OPTORDER 5
#defineVMEM_OPTVALUE (1 << VMEM_OPTORDER)
#defineVMEM_MAXORDER \
@@ -652,6 +655,16 @@ vmem_bt_alloc(uma_zone_t zone, vm_size_t bytes, int do
pause("btalloc", 1);
return (NULL);
+}
+
+/*
+ * How many pages do we need to startup_alloc.
+ */
+int
+vmem_startup_count(void)
+{
+
+ return (howmany(BT_MAXALLOC, UMA_SLAB_SIZE / sizeof(struct vmem_btag)));
}
#endif
Modified: head/sys/vm/vm_init.c
==
--- head/sys/vm/vm_init.c Tue Feb 6 21:35:41 2018(r328951)
+++ head/sys/vm/vm_init.c Tue Feb 6 22:06:59 2018(r328952)
@@ -93,6 +93,8 @@ __FBSDID("$FreeBSD$");
#include
#include
+extern voiduma_startup1(void);
+extern voidvm_radix_reserve_kva(void);
#if VM_NRESERVLEVEL > 0
#defineKVA_QUANTUM (1 << (VM_LEVEL_0_ORDER + PAGE_SHIFT))
@@ -150,7 +152,11 @@ vm_mem_init(dummy)
*/
vm_set_page_size();
virtual_avail = vm_page_startup(virtual_avail);
-
+
+#ifdef UMA_MD_SMALL_ALLOC
+ /* Announce page availability to UMA. */
+ uma_startup1();
+#endif
/*
* Initialize other VM packages
*/
@@ -173,6 +179,12 @@ vm_mem_init(dummy)
KVA_QUANTUM);
}
+#ifndefUMA_MD_SMALL_ALLOC
+ /* Set up radix zone to use noobj_alloc. */
+ vm_radix_reserve_kva();
+ /* Announce page availability to UMA. */
+ uma_startup1();
+#endif
kmem_init_zero_region();
pmap_init();
vm_pager_init();
Modified: head/sys/vm/vm_page.c
==
--- head/sys/vm/vm_page.c Tue Feb 6 21:35:41 2018(r328951)
+++ head/sys/vm/vm_page.c Tue Feb 6 22:06:59 2018(r328952)
@@ -112,6 +112,7 @@ __FBSDID("$FreeBSD$");
#include
#include
#include
+#include
#include
#include
#include
@@ -127,7 +128,7 @@ __FBSDID("$FreeBSD$");
extern int uma_startup_count(int);
extern voiduma_startup(void *, int);
-extern voiduma_startup1(void);
+extern int vmem_startup_count(void);
/*
* Associated with page of user-allocatable memory is a
@@ -501,12 +502,33 @@ vm_page_startup(vm_offset_t vaddr)
/*
* Allocate memory for use when boot strapping the kernel memory
-* allocator.
-*
+* allocator. Tell UMA how many zones we are going to create
+* before going fully functional. UMA will add its zones.
+*/
+#ifdef UMA_MD_SMALL_ALLOC
+ boot_pages = uma_startup_count(0);
+#else
+ /*
+* VM startup zones: vmem, vmem_btag, VM OBJECT, RADIX NODE, MAP,
+* KMAP ENTRY, MAP ENTRY, VMSPACE.
+*/
+ boot_pages = uma_startup_count(8);
+
+ /* vmem_startup() calls uma_prealloc(). */
+ boot_pages += vmem_startup_count();
+ /* vm_map_startup() calls uma_prealloc(). */
+ boot_pages += howmany(MAX_KMAP, UMA_SLAB_SIZE / sizeof(struct vm_map));
+
+ /*
+* Before going fully functional kmem_init() does allocation
+* from "KMAP ENTRY" and vmem_create() does allocation from "vmem".
+*/
+ boot_pages += 2;
+#endif
+ /*
* CTFLAG_RDTUN doesn't work during the early boot process, so we must
* manually fetch the value.
*/
- boot_pages = uma_startup_count(0);
TUNABLE_INT_FETCH("vm.boot_pages", &boot_pages);
new_end = end - (boot_pages * UMA_SLAB_SIZE);
new_end = trunc_page(new_end);
@@ -739,9 +761,6 @@ vm_page_startup(vm_offset_t vaddr)
* can work.
*/
domainset_zero();
svn commit: r328953 - head/sys/vm
Author: glebius
Date: Tue Feb 6 22:08:43 2018
New Revision: 328953
URL: https://svnweb.freebsd.org/changeset/base/328953
Log:
Improve DIAGNOSTIC printf. Report using a boot page every time regardless
of booted status.
Modified:
head/sys/vm/uma_core.c
Modified: head/sys/vm/uma_core.c
==
--- head/sys/vm/uma_core.c Tue Feb 6 22:06:59 2018(r328952)
+++ head/sys/vm/uma_core.c Tue Feb 6 22:08:43 2018(r328953)
@@ -1088,12 +1088,11 @@ startup_alloc(uma_zone_t zone, vm_size_t bytes, int do
* Check our small startup cache to see if it has pages remaining.
*/
mtx_lock(&uma_boot_pages_mtx);
+ if (pages <= boot_pages) {
#ifdef DIAGNOSTIC
- if (booted < BOOT_PAGEALLOC)
printf("%s from \"%s\", %d boot pages left\n", __func__,
zone->uz_name, boot_pages);
#endif
- if (pages <= boot_pages) {
mem = bootmem;
boot_pages -= pages;
bootmem += pages * PAGE_SIZE;
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328954 - in head/sys: amd64/amd64 arm/arm cddl/compat/opensolaris/sys cddl/contrib/opensolaris/uts/common/fs/zfs compat/linprocfs fs/tmpfs i386/i386 kern mips/mips powerpc/booke powerp...
Author: jeff
Date: Tue Feb 6 22:10:07 2018
New Revision: 328954
URL: https://svnweb.freebsd.org/changeset/base/328954
Log:
Use per-domain locks for vm page queue free. Move paging control from
global to per-domain state. Protect reservations with the free lock
from the domain that they belong to. Refactor to make vm domains more
of a first class object.
Reviewed by:markj, kib, gallatin
Tested by: pho
Sponsored by: Netflix, Dell/EMC Isilon
Differential Revision: https://reviews.freebsd.org/D14000
Added:
head/sys/vm/vm_pagequeue.h (contents, props changed)
Modified:
head/sys/amd64/amd64/machdep.c
head/sys/arm/arm/machdep.c
head/sys/arm/arm/pmap-v4.c
head/sys/cddl/compat/opensolaris/sys/kmem.h
head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
head/sys/compat/linprocfs/linprocfs.c
head/sys/fs/tmpfs/tmpfs_subr.c
head/sys/i386/i386/machdep.c
head/sys/kern/init_main.c
head/sys/kern/subr_vmem.c
head/sys/kern/subr_witness.c
head/sys/mips/mips/machdep.c
head/sys/powerpc/booke/pmap.c
head/sys/powerpc/powerpc/machdep.c
head/sys/sparc64/sparc64/machdep.c
head/sys/sys/vmmeter.h
head/sys/vm/swap_pager.c
head/sys/vm/uma_core.c
head/sys/vm/vm_extern.h
head/sys/vm/vm_glue.c
head/sys/vm/vm_init.c
head/sys/vm/vm_kern.c
head/sys/vm/vm_map.c
head/sys/vm/vm_meter.c
head/sys/vm/vm_object.c
head/sys/vm/vm_object.h
head/sys/vm/vm_page.c
head/sys/vm/vm_page.h
head/sys/vm/vm_pageout.c
head/sys/vm/vm_pageout.h
head/sys/vm/vm_phys.c
head/sys/vm/vm_phys.h
head/sys/vm/vm_reserv.c
head/sys/vm/vm_reserv.h
head/sys/vm/vm_swapout.c
head/sys/vm/vnode_pager.c
Modified: head/sys/amd64/amd64/machdep.c
==
--- head/sys/amd64/amd64/machdep.c Tue Feb 6 22:08:43 2018
(r328953)
+++ head/sys/amd64/amd64/machdep.c Tue Feb 6 22:10:07 2018
(r328954)
@@ -282,7 +282,7 @@ cpu_startup(dummy)
memsize = (uintmax_t)strtoul(sysenv, (char **)NULL, 10) << 10;
freeenv(sysenv);
}
- if (memsize < ptoa((uintmax_t)vm_cnt.v_free_count))
+ if (memsize < ptoa((uintmax_t)vm_free_count()))
memsize = ptoa((uintmax_t)Maxmem);
printf("real memory = %ju (%ju MB)\n", memsize, memsize >> 20);
realmem = atop(memsize);
@@ -309,8 +309,8 @@ cpu_startup(dummy)
vm_ksubmap_init(&kmi);
printf("avail memory = %ju (%ju MB)\n",
- ptoa((uintmax_t)vm_cnt.v_free_count),
- ptoa((uintmax_t)vm_cnt.v_free_count) / 1048576);
+ ptoa((uintmax_t)vm_free_count()),
+ ptoa((uintmax_t)vm_free_count()) / 1048576);
/*
* Set up buffers, so they can be used to read disk labels.
Modified: head/sys/arm/arm/machdep.c
==
--- head/sys/arm/arm/machdep.c Tue Feb 6 22:08:43 2018(r328953)
+++ head/sys/arm/arm/machdep.c Tue Feb 6 22:10:07 2018(r328954)
@@ -228,8 +228,8 @@ cpu_startup(void *dummy)
(uintmax_t)arm32_ptob(realmem),
(uintmax_t)arm32_ptob(realmem) / mbyte);
printf("avail memory = %ju (%ju MB)\n",
- (uintmax_t)arm32_ptob(vm_cnt.v_free_count),
- (uintmax_t)arm32_ptob(vm_cnt.v_free_count) / mbyte);
+ (uintmax_t)arm32_ptob(vm_free_count()),
+ (uintmax_t)arm32_ptob(vm_free_count()) / mbyte);
if (bootverbose) {
arm_physmem_print_tables();
devmap_print_table();
Modified: head/sys/arm/arm/pmap-v4.c
==
--- head/sys/arm/arm/pmap-v4.c Tue Feb 6 22:08:43 2018(r328953)
+++ head/sys/arm/arm/pmap-v4.c Tue Feb 6 22:10:07 2018(r328954)
@@ -3817,7 +3817,7 @@ pmap_get_pv_entry(void)
pv_entry_count++;
if (pv_entry_count > pv_entry_high_water)
- pagedaemon_wakeup();
+ pagedaemon_wakeup(0); /* XXX ARM NUMA */
ret_value = uma_zalloc(pvzone, M_NOWAIT);
return ret_value;
}
Modified: head/sys/cddl/compat/opensolaris/sys/kmem.h
==
--- head/sys/cddl/compat/opensolaris/sys/kmem.h Tue Feb 6 22:08:43 2018
(r328953)
+++ head/sys/cddl/compat/opensolaris/sys/kmem.h Tue Feb 6 22:10:07 2018
(r328954)
@@ -78,7 +78,7 @@ void kmem_reap(void);
int kmem_debugging(void);
void *calloc(size_t n, size_t s);
-#definefreemem vm_cnt.v_free_count
+#definefreemem vm_free_count()
#defineminfree vm_cnt.v_free_min
#defineheap_arena kernel_arena
#definezio_arena NULL
Modified: head/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
=
svn commit: r328955 - head/sys/vm
Author: glebius Date: Tue Feb 6 22:13:40 2018 New Revision: 328955 URL: https://svnweb.freebsd.org/changeset/base/328955 Log: Use correct arithmetic to calculate how many pages we need for kegs and hashes. There is no functional change with current sizes. Modified: head/sys/vm/uma_core.c Modified: head/sys/vm/uma_core.c == --- head/sys/vm/uma_core.c Tue Feb 6 22:10:07 2018(r328954) +++ head/sys/vm/uma_core.c Tue Feb 6 22:13:40 2018(r328955) @@ -1810,14 +1810,14 @@ uma_startup_count(int zones) pages += howmany(zones, UMA_SLAB_SIZE / zsize); /* ... and their kegs. */ - pages += howmany(ksize * zones, UMA_SLAB_SIZE); + pages += howmany(zones, UMA_SLAB_SIZE / ksize); /* * Take conservative approach that every zone * is going to allocate hash. */ - pages += howmany(sizeof(struct slabhead *) * UMA_HASH_SIZE_INIT * - zones, UMA_SLAB_SIZE); + pages += howmany(zones, UMA_SLAB_SIZE / + (sizeof(struct slabhead *) * UMA_HASH_SIZE_INIT)); return (pages); } ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
Peter, can you please check post r328952 kernel? -- Gleb Smirnoff ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328956 - in head/sys: kern sys
Author: ian
Date: Tue Feb 6 22:17:01 2018
New Revision: 328956
URL: https://svnweb.freebsd.org/changeset/base/328956
Log:
Use const pointers for input data not modified by clock utility functions.
Modified:
head/sys/kern/subr_clock.c
head/sys/kern/subr_fattime.c
head/sys/sys/clock.h
Modified: head/sys/kern/subr_clock.c
==
--- head/sys/kern/subr_clock.c Tue Feb 6 22:13:40 2018(r328955)
+++ head/sys/kern/subr_clock.c Tue Feb 6 22:17:01 2018(r328956)
@@ -132,7 +132,7 @@ leapyear(int year)
}
static void
-print_ct(struct clocktime *ct)
+print_ct(const struct clocktime *ct)
{
printf("[%04d-%02d-%02d %02d:%02d:%02d]",
ct->year, ct->mon, ct->day,
@@ -140,7 +140,7 @@ print_ct(struct clocktime *ct)
}
int
-clock_ct_to_ts(struct clocktime *ct, struct timespec *ts)
+clock_ct_to_ts(const struct clocktime *ct, struct timespec *ts)
{
int i, year, days;
@@ -200,7 +200,7 @@ clock_ct_to_ts(struct clocktime *ct, struct timespec *
}
int
-clock_bcd_to_ts(struct bcd_clocktime *bct, struct timespec *ts, bool ampm)
+clock_bcd_to_ts(const struct bcd_clocktime *bct, struct timespec *ts, bool
ampm)
{
struct clocktime ct;
int bcent, byear;
@@ -249,7 +249,7 @@ clock_bcd_to_ts(struct bcd_clocktime *bct, struct time
}
void
-clock_ts_to_ct(struct timespec *ts, struct clocktime *ct)
+clock_ts_to_ct(const struct timespec *ts, struct clocktime *ct)
{
time_t i, year, days;
time_t rsec;/* remainder seconds */
@@ -310,7 +310,7 @@ clock_ts_to_ct(struct timespec *ts, struct clocktime *
}
void
-clock_ts_to_bcd(struct timespec *ts, struct bcd_clocktime *bct, bool ampm)
+clock_ts_to_bcd(const struct timespec *ts, struct bcd_clocktime *bct, bool
ampm)
{
struct clocktime ct;
Modified: head/sys/kern/subr_fattime.c
==
--- head/sys/kern/subr_fattime.cTue Feb 6 22:13:40 2018
(r328955)
+++ head/sys/kern/subr_fattime.cTue Feb 6 22:17:01 2018
(r328956)
@@ -137,7 +137,8 @@ static const struct {
void
-timespec2fattime(struct timespec *tsp, int utc, uint16_t *ddp, uint16_t *dtp,
uint8_t *dhp)
+timespec2fattime(const struct timespec *tsp, int utc, uint16_t *ddp,
+uint16_t *dtp, uint8_t *dhp)
{
time_t t1;
unsigned t2, l, m;
@@ -217,7 +218,8 @@ static const uint16_t daytab[64] = {
};
void
-fattime2timespec(unsigned dd, unsigned dt, unsigned dh, int utc, struct
timespec *tsp)
+fattime2timespec(unsigned dd, unsigned dt, unsigned dh, int utc,
+struct timespec *tsp)
{
unsigned day;
Modified: head/sys/sys/clock.h
==
--- head/sys/sys/clock.hTue Feb 6 22:13:40 2018(r328955)
+++ head/sys/sys/clock.hTue Feb 6 22:17:01 2018(r328956)
@@ -88,8 +88,8 @@ struct clocktime {
longnsec; /* nano seconds */
};
-int clock_ct_to_ts(struct clocktime *, struct timespec *);
-void clock_ts_to_ct(struct timespec *, struct clocktime *);
+int clock_ct_to_ts(const struct clocktime *, struct timespec *);
+void clock_ts_to_ct(const struct timespec *, struct clocktime *);
/*
* Structure to hold the values typically reported by time-of-day clocks,
@@ -125,8 +125,8 @@ struct bcd_clocktime {
bool ispm; /* true if hour represents pm time */
};
-int clock_bcd_to_ts(struct bcd_clocktime *, struct timespec *, bool ampm);
-void clock_ts_to_bcd(struct timespec *, struct bcd_clocktime *, bool ampm);
+int clock_bcd_to_ts(const struct bcd_clocktime *, struct timespec *, bool
ampm);
+void clock_ts_to_bcd(const struct timespec *, struct bcd_clocktime *, bool
ampm);
/*
* Time-of-day clock functions and flags. These functions might sleep.
@@ -177,8 +177,10 @@ void clock_unregister(device_t _clockdev);
/* Traditional POSIX base year */
#definePOSIX_BASE_YEAR 1970
-void timespec2fattime(struct timespec *tsp, int utc, u_int16_t *ddp, u_int16_t
*dtp, u_int8_t *dhp);
-void fattime2timespec(unsigned dd, unsigned dt, unsigned dh, int utc, struct
timespec *tsp);
+void timespec2fattime(const struct timespec *tsp, int utc, u_int16_t *ddp,
+u_int16_t *dtp, u_int8_t *dhp);
+void fattime2timespec(unsigned dd, unsigned dt, unsigned dh, int utc,
+struct timespec *tsp);
#endif /* _KERNEL */
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328957 - in head/sys: fs/ext2fs ufs/ufs
Author: pfg
Date: Tue Feb 6 22:38:19 2018
New Revision: 328957
URL: https://svnweb.freebsd.org/changeset/base/328957
Log:
{ext2|ufs}_readdir: Avoid setting negative ncookies.
ncookies cannot be negative or the allocator will fail. This should only
happen if a caller is very broken but we can still try to survive the
event.
We should probably also verify for uio_resid > MAXPHYS but in that case
it is not clear that just clipping the ncookies value is an adequate
response.
MFC after:2 weeks
Modified:
head/sys/fs/ext2fs/ext2_lookup.c
head/sys/ufs/ufs/ufs_vnops.c
Modified: head/sys/fs/ext2fs/ext2_lookup.c
==
--- head/sys/fs/ext2fs/ext2_lookup.cTue Feb 6 22:17:01 2018
(r328956)
+++ head/sys/fs/ext2fs/ext2_lookup.cTue Feb 6 22:38:19 2018
(r328957)
@@ -153,7 +153,10 @@ ext2_readdir(struct vop_readdir_args *ap)
return (EINVAL);
ip = VTOI(vp);
if (ap->a_ncookies != NULL) {
- ncookies = uio->uio_resid;
+ if (uio->uio_resid < 0)
+ ncookies = 0;
+ else
+ ncookies = uio->uio_resid;
if (uio->uio_offset >= ip->i_size)
ncookies = 0;
else if (ip->i_size - uio->uio_offset < ncookies)
Modified: head/sys/ufs/ufs/ufs_vnops.c
==
--- head/sys/ufs/ufs/ufs_vnops.cTue Feb 6 22:17:01 2018
(r328956)
+++ head/sys/ufs/ufs/ufs_vnops.cTue Feb 6 22:38:19 2018
(r328957)
@@ -2179,7 +2179,10 @@ ufs_readdir(ap)
if (ip->i_effnlink == 0)
return (0);
if (ap->a_ncookies != NULL) {
- ncookies = uio->uio_resid;
+ if (uio->uio_resid < 0)
+ ncookies = 0;
+ else
+ ncookies = uio->uio_resid;
if (uio->uio_offset >= ip->i_size)
ncookies = 0;
else if (ip->i_size - uio->uio_offset < ncookies)
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
On Tue, Feb 06, 2018 at 02:15:55PM -0800, Gleb Smirnoff wrote: > Peter, > > can you please check post r328952 kernel? > Sure. FreeBSD 12.0-CURRENT #0 r328956: Tue Feb 6 23:26:58 CET 2018 p...@x4.osted.lan:/usr/src/sys/i386/compile/PHO i386 FreeBSD clang version 6.0.0 (branches/release_60 321788) (based on LLVM 6.0.0) WARNING: WITNESS option enabled, expect reduced performance. WARNING: DIAGNOSTIC option enabled, expect reduced performance. Entering uma_startup with 9 boot pages configured startup_alloc from "UMA Kegs", 8 boot pages left startup_alloc from "UMA Zones", 7 boot pages left startup_alloc from "UMA Hash", 6 boot pages left startup_alloc from "UMA Zones", 5 boot pages left startup_alloc from "vmem btag", 4 boot pages left startup_alloc from "MAP", 3 boot pages left startup_alloc from "UMA Kegs", 2 boot pages left startup_alloc from "KMAP ENTRY", 1 boot pages left panic: UMA: Increase vm.boot_pages cpuid = 0 time = 1 KDB: stack backtrace: db_trace_self_wrapper(c1654440,c23af290,c1be5bc0,8,c1bbec20,...) at db_trace_self_wrapper+0x2a/frame 0xc2022a70 kdb_backtrace(c164e477,1,0,c2022b30,0,...) at kdb_backtrace+0x2d/frame 0xc2022ad8 vpanic(c16a125d,c2022b30,c2022b30,c2022b48,c0fb7933,...) at vpanic+0x133/frame 0xc2022b10 panic(c16a125d,0,c16a0cb0,44f,c23b0a80,...) at panic+0x1b/frame 0xc2022b24 startup_alloc(c23b1dc0,1000,0,c2022b7b,102,...) at startup_alloc+0x143/frame 0xc2022b48 keg_alloc_slab(0,2,c16a0cb0,983,c23b7000,...) at keg_alloc_slab+0xf6/frame 0xc2022b90 keg_fetch_slab(,2,c16a0cb0,9e5,c23b1dc0,...) at keg_fetch_slab+0x10e/frame 0xc2022be8 zone_fetch_slab(c23b1dc0,0,,2,c23b5000,...) at zone_fetch_slab+0x61/frame 0xc2022c0c zone_import(c23b1dc0,c2022c68,1,,2,...) at zone_import+0x3b/frame 0xc2022c44 zone_alloc_item(,2,c16a0cc7,c16357bd,c1b5d050,...) at zone_alloc_item+0x3d/frame 0xc2022c78 uma_zalloc_arg(c23b1dc0,0,2,c1b5d050,0,...) at uma_zalloc_arg+0x66c/frame 0xc2022cb4 vmem_create(c16a1c9a,0,0,1000,0,...) at vmem_create+0x2a/frame 0xc2022ce4 vm_mem_init(0,0,0,0,c186a9d4,...) at vm_mem_init+0xf4/frame 0xc2022d10 mi_startup() at mi_startup+0xf7/frame 0xc2022d38 begin() at begin+0x2f KDB: enter: panic [ thread pid 0 tid 0 ] Stopped at kdb_enter+0x3a: movl$0,kdb_why db> Setting vm.boot_pages in the loader to 128 fixes this. - Peter ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
On Tue, Feb 06, 2018 at 11:52:44PM +0100, Peter Holm wrote:
P> On Tue, Feb 06, 2018 at 02:15:55PM -0800, Gleb Smirnoff wrote:
P> > Peter,
P> >
P> > can you please check post r328952 kernel?
P> >
P>
P> Sure.
P>
P> FreeBSD 12.0-CURRENT #0 r328956: Tue Feb 6 23:26:58 CET 2018
P> p...@x4.osted.lan:/usr/src/sys/i386/compile/PHO i386
P> FreeBSD clang version 6.0.0 (branches/release_60 321788) (based on LLVM
6.0.0)
P> WARNING: WITNESS option enabled, expect reduced performance.
P> WARNING: DIAGNOSTIC option enabled, expect reduced performance.
P> Entering uma_startup with 9 boot pages configured
P> startup_alloc from "UMA Kegs", 8 boot pages left
P> startup_alloc from "UMA Zones", 7 boot pages left
P> startup_alloc from "UMA Hash", 6 boot pages left
P> startup_alloc from "UMA Zones", 5 boot pages left
P> startup_alloc from "vmem btag", 4 boot pages left
P> startup_alloc from "MAP", 3 boot pages left
P> startup_alloc from "UMA Kegs", 2 boot pages left
P> startup_alloc from "KMAP ENTRY", 1 boot pages left
P> panic: UMA: Increase vm.boot_pages
Off by one :( Can you please boot with this patch and show dmesg?
--
Gleb Smirnoff
Index: sys/vm/uma_core.c
===
--- sys/vm/uma_core.c (revision 328955)
+++ sys/vm/uma_core.c (working copy)
@@ -1800,6 +1800,7 @@ uma_startup_count(int zones)
/* Memory for the zone of zones and zone of kegs. */
pages = howmany(roundup(zsize, CACHE_LINE_SIZE) * 2 +
roundup(ksize, CACHE_LINE_SIZE), PAGE_SIZE);
+ printf("boot_pages master %d\n", pages);
zones += UMA_BOOT_ZONES;
@@ -1808,9 +1809,11 @@ uma_startup_count(int zones)
pages += zones * howmany(zsize, UMA_SLAB_SIZE);
else
pages += howmany(zones, UMA_SLAB_SIZE / zsize);
+ printf("boot_pages zones %d\n", pages);
/* ... and their kegs. */
pages += howmany(zones, UMA_SLAB_SIZE / ksize);
+ printf("boot_pages kegs %d\n", pages);
/*
* Take conservative approach that every zone
@@ -1818,6 +1821,7 @@ uma_startup_count(int zones)
*/
pages += howmany(zones, UMA_SLAB_SIZE /
(sizeof(struct slabhead *) * UMA_HASH_SIZE_INIT));
+ printf("boot_pages hash %d\n", pages);
return (pages);
}
Index: sys/vm/vm_page.c
===
--- sys/vm/vm_page.c (revision 328955)
+++ sys/vm/vm_page.c (working copy)
@@ -518,8 +518,10 @@ vm_page_startup(vm_offset_t vaddr)
/* vmem_startup() calls uma_prealloc(). */
boot_pages += vmem_startup_count();
+ printf("boot_pages vmem %s\n", boot_pages);
/* vm_map_startup() calls uma_prealloc(). */
boot_pages += howmany(MAX_KMAP, UMA_SLAB_SIZE / sizeof(struct vm_map));
+ printf("boot_pages kmap %s\n", boot_pages);
/*
* Before going fully functional kmem_init() does allocation
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328958 - head/share/zoneinfo
Author: imp
Date: Tue Feb 6 23:12:16 2018
New Revision: 328958
URL: https://svnweb.freebsd.org/changeset/base/328958
Log:
Avoid find -s, use find | sort instead.
find -s was introduced to make the metalog more
deterministic. However, find -s is not portable. find | sort is
portable and accomplishes the same goals, even if it isn't
pedantically the same. TZS is the same before / after the change so
any fussy differences between the two are moot and there won't be
METALOG churn across this change.
Differential Revision: https://reviews.freebsd.org/D14231
Modified:
head/share/zoneinfo/Makefile
Modified: head/share/zoneinfo/Makefile
==
--- head/share/zoneinfo/MakefileTue Feb 6 22:38:19 2018
(r328957)
+++ head/share/zoneinfo/MakefileTue Feb 6 23:12:16 2018
(r328958)
@@ -95,8 +95,13 @@ zoneinfo: yearistype ${TDATA}
zic -D -d ${TZBUILDDIR} -p ${POSIXRULES} -m ${NOBINMODE} \
${LEAPFILE} -y ${.OBJDIR}/yearistype ${TZFILES}
+#
+# Sort TZS to ensure they are the same every build. find -s might
+# be a shorter way to express this, but it's non-portable. Any
+# differences between the two don't matter for this purpose.
+#
.if make(*install*)
-TZS!= cd ${TZBUILDDIR} && find -s * -type f
+TZS!= cd ${TZBUILDDIR} && find * -type f | env LC_ALL=C sort
.endif
beforeinstall: install-zoneinfo
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328959 - head/lib/libc/sys
Author: cem Date: Tue Feb 6 23:12:47 2018 New Revision: 328959 URL: https://svnweb.freebsd.org/changeset/base/328959 Log: fsync.2: Cross-reference fsync(1) Reported by: rpokala Sponsored by: Dell EMC Isilon Modified: head/lib/libc/sys/fsync.2 Modified: head/lib/libc/sys/fsync.2 == --- head/lib/libc/sys/fsync.2 Tue Feb 6 23:12:16 2018(r328958) +++ head/lib/libc/sys/fsync.2 Tue Feb 6 23:12:47 2018(r328959) @@ -34,7 +34,7 @@ .\" @(#)fsync.28.1 (Berkeley) 6/4/93 .\" $FreeBSD$ .\" -.Dd May 24, 2017 +.Dd February 6, 2018 .Dt FSYNC 2 .Os .Sh NAME @@ -108,6 +108,7 @@ refers to a socket, not to a file. An I/O error occurred while reading from or writing to the file system. .El .Sh SEE ALSO +.Xr fsync 1 , .Xr sync 2 , .Xr syncer 4 , .Xr sync 8 ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328934 - in head: . bin/sh
Ian Lepore ian at freebsd.org wrote on Tue Feb 6 19:35:21 UTC 2018 : > I don't understand this idea of /usr/local "polluting" a system. It > seems to me exactly the opposite would be the case... if I have found > some 3rd party version of mktemp that I like better, it would be > installed in /usr/local. If I went out of my way to install that, then > naturally I WANT it to be used. To me, it's insane that the /usr/local > paths are not in front of the base system paths by default, and it's > even more insane that the base system works so hard to NOT use the > replacements I've installed (even if I've arranged PATH so that the > right versions should be used) so that I have to track down why it's > using the wrong thing and apply ad-hoc fixes. I've had mixed results over the years. I've had ports install files for other 'internal' purposes (not my overall goals) that in turned broke buildworld for powerpc(64) for some alternative compiler/tool chains that induced /usr/local/ header file usage if a file name accidentally matched. There was a period of time where I'd rename some specific header files as I switched activities in order to avoid systematic problems. (It seemed easier than than other alternatives that I considered at the time.) [But I've been doing odd "target powerpc64 and powerpc without gcc 4.2.1" experiments for a few years. Not exactly a typical context.] As the build system for buildworld has progressed, I've not had such problems for some time: it became easier to avoid /usr/local/include/ getting involved, at least for what I've been doing. Also, I do external toolchain activity less often these days. === Mark Millard marklmi at yahoo.com ( markmi at dsl-only.net is going away in 2018-Feb, late) ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328960 - head/sys/cam
Author: imp
Date: Tue Feb 6 23:21:08 2018
New Revision: 328960
URL: https://svnweb.freebsd.org/changeset/base/328960
Log:
Keep a counter for number of requests completed with an error.
Sponsored by: Netflix
Modified:
head/sys/cam/cam_iosched.c
Modified: head/sys/cam/cam_iosched.c
==
--- head/sys/cam/cam_iosched.c Tue Feb 6 23:12:47 2018(r328959)
+++ head/sys/cam/cam_iosched.c Tue Feb 6 23:21:08 2018(r328960)
@@ -223,6 +223,7 @@ struct iop_stats {
int total; /* Total for all time -- wraps */
int in; /* number queued all time -- wraps */
int out;/* number completed all time -- wraps */
+ int errs; /* Number of I/Os completed with error
-- wraps */
/*
* Statistics on different bits of the process.
@@ -781,6 +782,7 @@ cam_iosched_iop_stats_init(struct cam_iosched_softc *i
ios->max = ios->current = 30;
ios->min = 1;
ios->out = 0;
+ ios->errs = 0;
ios->pending = 0;
ios->queued = 0;
ios->total = 0;
@@ -971,7 +973,11 @@ cam_iosched_iop_stats_sysctl_init(struct cam_iosched_s
SYSCTL_ADD_INT(ctx, n,
OID_AUTO, "out", CTLFLAG_RD,
&ios->out, 0,
- "# of transactions completed");
+ "# of transactions completed (including with error)");
+ SYSCTL_ADD_INT(ctx, n,
+ OID_AUTO, "errs", CTLFLAG_RD,
+ &ios->errs, 0,
+ "# of transactions completed with an error");
SYSCTL_ADD_PROC(ctx, n,
OID_AUTO, "limiter", CTLTYPE_STRING | CTLFLAG_RW,
@@ -1463,13 +1469,19 @@ cam_iosched_bio_complete(struct cam_iosched_softc *isc
printf("done: %p %#x\n", bp, bp->bio_cmd);
if (bp->bio_cmd == BIO_WRITE) {
retval = cam_iosched_limiter_iodone(&isc->write_stats, bp);
+ if (!(bp->bio_flags & BIO_ERROR))
+ isc->write_stats.errs++;
isc->write_stats.out++;
isc->write_stats.pending--;
} else if (bp->bio_cmd == BIO_READ) {
retval = cam_iosched_limiter_iodone(&isc->read_stats, bp);
+ if (!(bp->bio_flags & BIO_ERROR))
+ isc->read_stats.errs++;
isc->read_stats.out++;
isc->read_stats.pending--;
} else if (bp->bio_cmd == BIO_DELETE) {
+ if (!(bp->bio_flags & BIO_ERROR))
+ isc->trim_stats.errs++;
isc->trim_stats.out++;
isc->trim_stats.pending--;
} else if (bp->bio_cmd != BIO_FLUSH) {
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
On Tue, Feb 06, 2018 at 03:06:35PM -0800, Gleb Smirnoff wrote: > On Tue, Feb 06, 2018 at 11:52:44PM +0100, Peter Holm wrote: > P> On Tue, Feb 06, 2018 at 02:15:55PM -0800, Gleb Smirnoff wrote: > P> > Peter, > P> > > P> > can you please check post r328952 kernel? > P> > > P> > P> Sure. > P> > P> FreeBSD 12.0-CURRENT #0 r328956: Tue Feb 6 23:26:58 CET 2018 > P> p...@x4.osted.lan:/usr/src/sys/i386/compile/PHO i386 > P> FreeBSD clang version 6.0.0 (branches/release_60 321788) (based on LLVM > 6.0.0) > P> WARNING: WITNESS option enabled, expect reduced performance. > P> WARNING: DIAGNOSTIC option enabled, expect reduced performance. > P> Entering uma_startup with 9 boot pages configured > P> startup_alloc from "UMA Kegs", 8 boot pages left > P> startup_alloc from "UMA Zones", 7 boot pages left > P> startup_alloc from "UMA Hash", 6 boot pages left > P> startup_alloc from "UMA Zones", 5 boot pages left > P> startup_alloc from "vmem btag", 4 boot pages left > P> startup_alloc from "MAP", 3 boot pages left > P> startup_alloc from "UMA Kegs", 2 boot pages left > P> startup_alloc from "KMAP ENTRY", 1 boot pages left > P> panic: UMA: Increase vm.boot_pages > > Off by one :( Can you please boot with this patch and show dmesg? > FreeBSD 12.0-CURRENT #2 r328956M: Wed Feb 7 00:20:47 CET 2018 p...@x4.osted.lan:/usr/src/sys/i386/compile/PHO i386 FreeBSD clang version 6.0.0 (branches/release_60 321788) (based on LLVM 6.0.0) WARNING: WITNESS option enabled, expect reduced performance. WARNING: DIAGNOSTIC option enabled, expect reduced performance. boot_pages master 1 boot_pages zones 3 boot_pages kegs 4 boot_pages hash 5 boot_pages vmem 6 boot_pages kmap 7 Entering uma_startup with 9 boot pages configured startup_alloc from "UMA Kegs", 8 boot pages left startup_alloc from "UMA Zones", 7 boot pages left startup_alloc from "UMA Hash", 6 boot pages left startup_alloc from "UMA Zones", 5 boot pages left startup_alloc from "vmem btag", 4 boot pages left startup_alloc from "MAP", 3 boot pages left startup_alloc from "UMA Kegs", 2 boot pages left startup_alloc from "KMAP ENTRY", 1 boot pages left panic: UMA: Increase vm.boot_pages cpuid = 0 time = 1 KDB: stack backtrace: db_trace_self_wrapper(c16544c0,c23af290,c1be5bc0,8,c1bbec20,...) at db_trace_self_wrapper+0x2a/frame 0xc2022a70 kdb_backtrace(c164e4f7,1,0,c2022b30,0,...) at kdb_backtrace+0x2d/frame 0xc2022ad8 vpanic(c16a1330,c2022b30,c2022b30,c2022b48,c0fb7983,...) at vpanic+0x133/frame 0xc2022b10 panic(c16a1330,0,c16a0d83,44f,c23b0a80,...) at panic+0x1b/frame 0xc2022b24 startup_alloc(c23b1dc0,1000,0,c2022b7b,102,...) at startup_alloc+0x143/frame 0xc2022b48 keg_alloc_slab(0,2,c16a0d83,987,c23b7000,...) at keg_alloc_slab+0xf6/frame 0xc2022b90 keg_fetch_slab(,2,c16a0d83,9e9,c23b1dc0,...) at keg_fetch_slab+0x10e/frame 0xc2022be8 zone_fetch_slab(c23b1dc0,0,,2,c23b5000,...) at zone_fetch_slab+0x61/frame 0xc2022c0c zone_import(c23b1dc0,c2022c68,1,,2,...) at zone_import+0x3b/frame 0xc2022c44 zone_alloc_item(,2,c16a0d9a,c163583d,c1b5d050,...) at zone_alloc_item+0x3d/frame 0xc2022c78 uma_zalloc_arg(c23b1dc0,0,2,c1b5d050,0,...) at uma_zalloc_arg+0x66c/frame 0xc2022cb4 vmem_create(c16a1d6e,0,0,1000,0,...) at vmem_create+0x2a/frame 0xc2022ce4 vm_mem_init(0,0,0,0,c186aab4,...) at vm_mem_init+0xf4/frame 0xc2022d10 mi_startup() at mi_startup+0xf7/frame 0xc2022d38 begin() at begin+0x2f KDB: enter: panic [ thread pid 0 tid 0 ] Stopped at kdb_enter+0x3a: movl$0,kdb_why db> - Peter ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328951 - head/etc/rc.d
> Author: feld (ports committer)
> Date: Tue Feb 6 21:35:41 2018
> New Revision: 328951
> URL: https://svnweb.freebsd.org/changeset/base/328951
>
> Log:
> Refactor cleanvar to remove shell expansion vulnerability
>
> If any process creates a directory named "-P" in /var/run or
> /var/spool/lock it will cause the purgedir function to start to rm -r /.
>
> Simplify a lot of complicated shell logic by leveraging find(1).
>
> Reviewed by:allanjude
> MFC after: 3 days
> Differential Revision: https://reviews.freebsd.org/D13778
Please be careful about use of binaries from /usr/bin in
/etc/rc.d, though in this case it is probably ok as /usr
has been mounted by the time cleanvar runs, that is not
always the case.
> Modified:
> head/etc/rc.d/cleanvar
>
> Modified: head/etc/rc.d/cleanvar
> ==
> --- head/etc/rc.d/cleanvarTue Feb 6 21:01:38 2018(r328950)
> +++ head/etc/rc.d/cleanvarTue Feb 6 21:35:41 2018(r328951)
> @@ -19,34 +19,6 @@ stop_cmd=":"
> extra_commands="reload"
> reload_cmd="${name}_start"
>
> -purgedir()
> -{
> - local dir file
> -
> - if [ $# -eq 0 ]; then
> - purgedir .
> - else
> - for dir
> - do
> - (
> - cd "$dir" && for file in .* *
> - do
> - # Skip over logging sockets
> - [ -S "$file" -a "$file" = "log" ] && continue
> - [ -S "$file" -a "$file" = "logpriv" ] &&
> continue
> - [ ."$file" = .. -o ."$file" = ... ] && continue
> - if [ -d "$file" -a ! -L "$file" ]
> - then
> - purgedir "$file"
> - else
> - rm -f -- "$file"
> - fi
> - done
> - )
> - done
> - fi
> -}
> -
> cleanvar_prestart()
> {
> # These files must be removed only the first time this script is run
> @@ -58,14 +30,17 @@ cleanvar_prestart()
> cleanvar_start()
> {
> if [ -d /var/run -a ! -f /var/run/clean_var ]; then
> - purgedir /var/run
> + # Skip over logging sockets
> + find /var/run \( -type f -or -type s ! -name log -and ! -name
> logpriv \) -delete
> >/var/run/clean_var
> fi
> if [ -d /var/spool/lock -a ! -f /var/spool/lock/clean_var ]; then
> - purgedir /var/spool/lock
> + find /var/spool/lock -type f -delete
> >/var/spool/lock/clean_var
> fi
> - rm -rf /var/spool/uucp/.Temp/*
> + if [ -d /var/spool/uucp/.Temp ]; then
> + find /var/spool/uucp/.Temp -delete
> + fi
> }
>
> load_rc_config $name
>
>
--
Rod Grimes rgri...@freebsd.org
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328934 - in head: . bin/sh
[ Charset ISO-8859-1 unsupported, converting... ] > On Tue, 2018-02-06 at 11:25 -0800, Rodney W. Grimes wrote: > > > > > > Author: arichardson > > > Date: Tue Feb??6 15:41:35 2018 > > > New Revision: 328934 > > > URL: https://svnweb.freebsd.org/changeset/base/328934 > > > > > > Log: > > > ? Don't hardcode /usr/bin as the path for mktemp in build tools > > > ?? > > > ? It won't work e.g. when crossbuilding from Ubuntu Linux as mktemp > > > is in > > > ? /bin there. > > > ?? > > > ? Reviewed By:bdrewery > > > ? Approved By:jhb (mentor) > > > ? Differential Revision: https://reviews.freebsd.org/D13937 > > Would it be better to create the variable MKTEMP to point at > > either /bin/mktemp or /usr/bin/mktemp dependent on platform, > > there are reasons we use full paths in Makefiles, mostly to > > stop /usr/local/bin/foo contimaton, which I believe this > > change now opens up, though very slight as I dont know of > > a third party mktemp binary. > > > > I don't understand this idea of /usr/local "polluting" a system. ?It > seems to me exactly the opposite would be the case... if I have found > some 3rd party version of mktemp that I like better, it would be > installed in /usr/local. ?If I went out of my way to install that, then > naturally I WANT it to be used. ?To me, it's insane that the /usr/local > paths are not in front of the base system paths by default, and it's > even more insane that the base system works so hard to NOT use the > replacements I've installed (even if I've arranged PATH so that the > right versions should be used) so that I have to track down why it's > using the wrong thing and apply ad-hoc fixes. Your welcome to wonder down that twisty maze of passages, I however shall not be following you. If you start to accept what ever the users path is for building the freebsd /usr/src tree the rate of failed builds and complaints, and hours spent chasing after oh crap it picked up foo from /usr/local is not a game I care to par take in. You personally, could probably get away with this and be fine, our user base would probably not be very lucky.w Oh, and reproducibility goes totally out the window if you fool with that. > -- Ian -- Rod Grimes rgri...@freebsd.org ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
Hi Peter,
can you please try this patch? In either case success
or not, please provide me with dmesg. Thanks a lot!
--
Gleb Smirnoff
Index: uma_core.c
===
--- uma_core.c (revision 328955)
+++ uma_core.c (working copy)
@@ -96,6 +96,7 @@ __FBSDID("$FreeBSD$");
#ifdef DEBUG_MEMGUARD
#include
#endif
+#define DIAGNOSTIC
/*
* This is the zone and keg from which all zones are spawned.
@@ -1800,6 +1801,7 @@ uma_startup_count(int zones)
/* Memory for the zone of zones and zone of kegs. */
pages = howmany(roundup(zsize, CACHE_LINE_SIZE) * 2 +
roundup(ksize, CACHE_LINE_SIZE), PAGE_SIZE);
+ printf("boot_pages master %d\n", pages);
zones += UMA_BOOT_ZONES;
@@ -1807,17 +1809,20 @@ uma_startup_count(int zones)
if (zsize > UMA_SLAB_SIZE)
pages += zones * howmany(zsize, UMA_SLAB_SIZE);
else
- pages += howmany(zones, UMA_SLAB_SIZE / zsize);
+ pages += howmany(zones, UMA_SLAB_SPACE / zsize);
+ printf("boot_pages zones %d\n", pages);
/* ... and their kegs. */
- pages += howmany(zones, UMA_SLAB_SIZE / ksize);
+ pages += howmany(zones, UMA_SLAB_SPACE / ksize);
+ printf("boot_pages kegs %d\n", pages);
/*
* Take conservative approach that every zone
* is going to allocate hash.
*/
- pages += howmany(zones, UMA_SLAB_SIZE /
+ pages += howmany(zones, UMA_SLAB_SPACE /
(sizeof(struct slabhead *) * UMA_HASH_SIZE_INIT));
+ printf("boot_pages hash %d\n", pages);
return (pages);
}
Index: uma_int.h
===
--- uma_int.h (revision 328955)
+++ uma_int.h (working copy)
@@ -138,6 +138,11 @@
#define UMA_MAX_WASTE 10
/*
+ * Size of memory in a not offpage slab available for actual items.
+ */
+#define UMA_SLAB_SPACE (UMA_SLAB_SIZE - sizeof(struct uma_slab))
+
+/*
* I doubt there will be many cases where this is exceeded. This is the initial
* size of the hash table for uma_slabs that are managed off page. This hash
* does expand by powers of two. Currently it doesn't get smaller.
Index: vm_page.c
===
--- vm_page.c (revision 328955)
+++ vm_page.c (working copy)
@@ -518,8 +518,11 @@ vm_page_startup(vm_offset_t vaddr)
/* vmem_startup() calls uma_prealloc(). */
boot_pages += vmem_startup_count();
+ printf("boot_pages vmem %d\n", boot_pages);
/* vm_map_startup() calls uma_prealloc(). */
- boot_pages += howmany(MAX_KMAP, UMA_SLAB_SIZE / sizeof(struct vm_map));
+ boot_pages += howmany(MAX_KMAP,
+ UMA_SLAB_SPACE / sizeof(struct vm_map));
+ printf("boot_pages kmap %d\n", boot_pages);
/*
* Before going fully functional kmem_init() does allocation
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
On Tue, Feb 06, 2018 at 04:42:13PM -0800, Gleb Smirnoff wrote:
T> Hi Peter,
T>
T> can you please try this patch? In either case success
T> or not, please provide me with dmesg. Thanks a lot!
Sorry, patch was missing one file. 99.9% this is a no-op,
but better use full patch.
--
Gleb Smirnoff
Index: sys/kern/subr_vmem.c
===
--- sys/kern/subr_vmem.c (revision 328955)
+++ sys/kern/subr_vmem.c (working copy)
@@ -667,7 +667,8 @@ int
vmem_startup_count(void)
{
- return (howmany(BT_MAXALLOC, UMA_SLAB_SIZE / sizeof(struct vmem_btag)));
+ return (howmany(BT_MAXALLOC,
+ UMA_SLAB_SPACE / sizeof(struct vmem_btag)));
}
#endif
Index: sys/vm/uma_core.c
===
--- sys/vm/uma_core.c (revision 328955)
+++ sys/vm/uma_core.c (working copy)
@@ -96,6 +96,7 @@ __FBSDID("$FreeBSD$");
#ifdef DEBUG_MEMGUARD
#include
#endif
+#define DIAGNOSTIC
/*
* This is the zone and keg from which all zones are spawned.
@@ -1800,6 +1801,7 @@ uma_startup_count(int zones)
/* Memory for the zone of zones and zone of kegs. */
pages = howmany(roundup(zsize, CACHE_LINE_SIZE) * 2 +
roundup(ksize, CACHE_LINE_SIZE), PAGE_SIZE);
+ printf("boot_pages master %d\n", pages);
zones += UMA_BOOT_ZONES;
@@ -1807,17 +1809,20 @@ uma_startup_count(int zones)
if (zsize > UMA_SLAB_SIZE)
pages += zones * howmany(zsize, UMA_SLAB_SIZE);
else
- pages += howmany(zones, UMA_SLAB_SIZE / zsize);
+ pages += howmany(zones, UMA_SLAB_SPACE / zsize);
+ printf("boot_pages zones %d\n", pages);
/* ... and their kegs. */
- pages += howmany(zones, UMA_SLAB_SIZE / ksize);
+ pages += howmany(zones, UMA_SLAB_SPACE / ksize);
+ printf("boot_pages kegs %d\n", pages);
/*
* Take conservative approach that every zone
* is going to allocate hash.
*/
- pages += howmany(zones, UMA_SLAB_SIZE /
+ pages += howmany(zones, UMA_SLAB_SPACE /
(sizeof(struct slabhead *) * UMA_HASH_SIZE_INIT));
+ printf("boot_pages hash %d\n", pages);
return (pages);
}
Index: sys/vm/uma_int.h
===
--- sys/vm/uma_int.h (revision 328955)
+++ sys/vm/uma_int.h (working copy)
@@ -138,6 +138,11 @@
#define UMA_MAX_WASTE 10
/*
+ * Size of memory in a not offpage slab available for actual items.
+ */
+#define UMA_SLAB_SPACE (UMA_SLAB_SIZE - sizeof(struct uma_slab))
+
+/*
* I doubt there will be many cases where this is exceeded. This is the initial
* size of the hash table for uma_slabs that are managed off page. This hash
* does expand by powers of two. Currently it doesn't get smaller.
Index: sys/vm/vm_page.c
===
--- sys/vm/vm_page.c (revision 328955)
+++ sys/vm/vm_page.c (working copy)
@@ -518,8 +518,11 @@ vm_page_startup(vm_offset_t vaddr)
/* vmem_startup() calls uma_prealloc(). */
boot_pages += vmem_startup_count();
+ printf("boot_pages vmem %d\n", boot_pages);
/* vm_map_startup() calls uma_prealloc(). */
- boot_pages += howmany(MAX_KMAP, UMA_SLAB_SIZE / sizeof(struct vm_map));
+ boot_pages += howmany(MAX_KMAP,
+ UMA_SLAB_SPACE / sizeof(struct vm_map));
+ printf("boot_pages kmap %d\n", boot_pages);
/*
* Before going fully functional kmem_init() does allocation
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r328964 - head/sys/arm/allwinner
Author: kevans
Date: Wed Feb 7 01:54:13 2018
New Revision: 328964
URL: https://svnweb.freebsd.org/changeset/base/328964
Log:
if_awg: Skip emac reset if configured for internal PHY
On the OrangePi One at least, emac reset when an ethernet cable is not
plugged in seems to break ethernet. Soft reset will fail, even with
increasing the delay and retries to wait for up to 20 seconds. This can be
reproduced across at least two different OrangePi One's by simply leaving
ethernet cable unplugged when awg attaches. Whether it's plugged in or not
through u-boot process makes no difference.
Skipping the reset in this configuration doesn't seem to cause any problems,
tried across many many reboots with and without ethernet cable plugged in.
Tested on:OrangePi One
Tested on:Other boards (manu)
Reviewed by: manu
Differential Revision:https://reviews.freebsd.org/D13974
Modified:
head/sys/arm/allwinner/if_awg.c
Modified: head/sys/arm/allwinner/if_awg.c
==
--- head/sys/arm/allwinner/if_awg.c Wed Feb 7 01:24:49 2018
(r328963)
+++ head/sys/arm/allwinner/if_awg.c Wed Feb 7 01:54:13 2018
(r328964)
@@ -1834,9 +1834,11 @@ awg_attach(device_t dev)
awg_get_eaddr(dev, eaddr);
/* Soft reset EMAC core */
- error = awg_reset(dev);
- if (error != 0)
- return (error);
+ if (!awg_has_internal_phy(dev)) {
+ error = awg_reset(dev);
+ if (error != 0)
+ return (error);
+ }
/* Setup DMA descriptors */
error = awg_setup_dma(dev);
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
On Tue, Feb 06, 2018 at 04:45:49PM -0800, Gleb Smirnoff wrote: > On Tue, Feb 06, 2018 at 04:42:13PM -0800, Gleb Smirnoff wrote: > T> Hi Peter, > T> > T> can you please try this patch? In either case success > T> or not, please provide me with dmesg. Thanks a lot! > > Sorry, patch was missing one file. 99.9% this is a no-op, > but better use full patch. > GDB: no debug ports present KDB: debugger backends: ddb KDB: current backend: ddb Copyright (c) 1992-2018 The FreeBSD Project. Copyright (c) 1979, 1980, 1983, 1986, 1988, 1989, 1991, 1992, 1993, 1994 The Regents of the University of California. All rights reserved. FreeBSD is a registered trademark of The FreeBSD Foundation. FreeBSD 12.0-CURRENT #2 r328956M: Wed Feb 7 00:20:47 CET 2018 p...@x4.osted.lan:/usr/src/sys/i386/compile/PHO i386 FreeBSD clang version 6.0.0 (branches/release_60 321788) (based on LLVM 6.0.0) WARNING: WITNESS option enabled, expect reduced performance. WARNING: DIAGNOSTIC option enabled, expect reduced performance. boot_pages master 1 boot_pages zones 3 boot_pages kegs 4 boot_pages hash 5 boot_pages vmem 6 boot_pages kmap 7 Entering uma_startup with 9 boot pages configured startup_alloc from "UMA Kegs", 8 boot pages left startup_alloc from "UMA Zones", 7 boot pages left startup_alloc from "UMA Hash", 6 boot pages left startup_alloc from "UMA Zones", 5 boot pages left startup_alloc from "vmem btag", 4 boot pages left startup_alloc from "MAP", 3 boot pages left startup_alloc from "UMA Kegs", 2 boot pages left startup_alloc from "KMAP ENTRY", 1 boot pages left panic: UMA: Increase vm.boot_pages cpuid = 0 time = 1 KDB: stack backtrace: db_trace_self_wrapper(c16544c0,c23af290,c1be5bc0,8,c1bbec20,...) at db_trace_self_wrapper+0x2a/frame 0xc2022a70 kdb_backtrace(c164e4f7,1,0,c2022b30,0,...) at kdb_backtrace+0x2d/frame 0xc2022ad8 vpanic(c16a1330,c2022b30,c2022b30,c2022b48,c0fb7983,...) at vpanic+0x133/frame 0xc2022b10 panic(c16a1330,0,c16a0d83,44f,c23b0a80,...) at panic+0x1b/frame 0xc2022b24 startup_alloc(c23b1dc0,1000,0,c2022b7b,102,...) at startup_alloc+0x143/frame 0xc2022b48 keg_alloc_slab(0,2,c16a0d83,987,c23b7000,...) at keg_alloc_slab+0xf6/frame 0xc2022b90 keg_fetch_slab(,2,c16a0d83,9e9,c23b1dc0,...) at keg_fetch_slab+0x10e/frame 0xc2022be8 zone_fetch_slab(c23b1dc0,0,,2,c23b5000,...) at zone_fetch_slab+0x61/frame 0xc2022c0c zone_import(c23b1dc0,c2022c68,1,,2,...) at zone_import+0x3b/frame 0xc2022c44 zone_alloc_item(,2,c16a0d9a,c163583d,c1b5d050,...) at zone_alloc_item+0x3d/frame 0xc2022c78 uma_zalloc_arg(c23b1dc0,0,2,c1b5d050,0,...) at uma_zalloc_arg+0x66c/frame 0xc2022cb4 vmem_create(c16a1d6e,0,0,1000,0,...) at vmem_create+0x2a/frame 0xc2022ce4 vm_mem_init(0,0,0,0,c186aab4,...) at vm_mem_init+0xf4/frame 0xc2022d10 mi_startup() at mi_startup+0xf7/frame 0xc2022d38 begin() at begin+0x2f KDB: enter: panic [ thread pid 0 tid 0 ] Stopped at kdb_enter+0x3a: movl$0,kdb_why db> - Peter ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r328916 - in head/sys: kern vm
On Wed, Feb 07, 2018 at 07:46:17AM +0100, Peter Holm wrote: > On Tue, Feb 06, 2018 at 04:45:49PM -0800, Gleb Smirnoff wrote: > > On Tue, Feb 06, 2018 at 04:42:13PM -0800, Gleb Smirnoff wrote: > > T> Hi Peter, > > T> > > T> can you please try this patch? In either case success > > T> or not, please provide me with dmesg. Thanks a lot! > > > > Sorry, patch was missing one file. 99.9% this is a no-op, > > but better use full patch. > > > Here the output with vm.boot_pages=128 https://people.freebsd.org/~pho/gleb011.txt - Peter ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
