Commit Graph

180620 Commits

Author SHA1 Message Date
Steven Hartland
e88aa3fd2c Refactored scsi_xpt use of device_has_vpd to generic scsi_vpd_supported_page
so its available for use in generic scsi code.

This is a pre-requirement for using VPD queries to determine available SCSI
delete methods within scsi_da.

Reviewed by:	mav
Approved by:	pjd (mentor)
MFC after:	2 weeks
2013-04-26 16:11:03 +00:00
Steven Hartland
c213c55153 Updated TRIM calculations in cam/ata to be based off ATA_DSM_* defines
Reviewed by:	mav
Approved by:	pjd (mentor)
MFC after:	2 weeks
2013-04-26 15:59:19 +00:00
Steven Hartland
b1da0a9868 Added the ability to send ATA identify and Data Set Management (DSM) TRIM
commands to an ATA device attached via a SCSI control.

sys/cam/scsi/scsi_all.c:
        - Added scsi_ata_identify, scsi_ata_trim
          Which use ATA Pass-Through to send commands to the attached disk.

sys/cam/scsi/scsi_all.h:
        - Added defines for all missing ATA Pass-Through commands values.

        - Added scsi_ata_identify, scsi_ata_trim methods used in ATA TRIM
          support.

        - Added scsi_vpd_logical_block_prov structure used when querying for
          the supported sizes UNMAP commands.

        - Added scsi_vpd_block_limits structure used when querying for the
          supported sizes of the UNMAP command.

Reviewed by:	mav
Approved by:	pjd (mentor)
MFC after:	2 weeks
2013-04-26 15:53:22 +00:00
Steven Hartland
88e8691037 Added Dataset Management defines to be used by TRIM in cam ata and scsi to
calculate the size of blocks.

Reviewed by:	mav
Approved by:	pjd (mentor)
MFC after:	2 weeks
2013-04-26 15:46:09 +00:00
Steven Hartland
6f926c0b82 Added a sysctl (kern.geom.dev.delete_max_sectors) to control the maximum
size of a delete request sent to the providing device performed by g_dev_ioctl.

This allows the kernel and apps via ioctl e.g. newfs -E to request large LBA
deletes which siginificantly improves performance.

Previously this was hard coded to 65536 sectors, the new default is 262144
which doubles the throughput of deletes on commonly available SSD's.

In tests on a Intel 520 120GB FW: 400i disk it improved the delete throughput
from 1.6GB/s to over 2.6GB/s on a full disk delete such as that done via
newfs -E

For some SSD's where delete time is pretty much constant, no matter what
the request, setting this to 0 will provide significantly better throughput
e.g. Samsung 840 240GB FW DXT07B0Q @ 262144 = 79G/s, @ 0 = 2259G/s

Reviewed by:	mav
Approved by:	pjd (mentor)
MFC after:	2 weeks
2013-04-26 15:43:24 +00:00
Steven Hartland
6225bf48ae Removed unneeded tests in dadeletemethodset changing it to return void
Reviewed by:	mav
Approved by:	pjd (mentor)
MFC after:	2 weeks
2013-04-26 15:31:52 +00:00
Gleb Smirnoff
a63fedfda7 Document constantness of dst argument to if_output.
Submitted by:	pluknet
2013-04-26 13:18:07 +00:00
Gleb Smirnoff
45b395cd2f Add usie to LINT. 2013-04-26 13:03:22 +00:00
Gleb Smirnoff
2d875db321 On those machines that use critical(9) for counter(9) we need to
include systm.h.
2013-04-26 13:02:11 +00:00
Gleb Smirnoff
47e8d432d5 Add const qualifier to the dst parameter of the ifnet if_output method. 2013-04-26 12:50:32 +00:00
Ulrich Spörlein
e83643f4f6 Don't appease clang static analyzer after all and roll back
the free(3) of mntbuf ... again. There's no point in doing
useless extra work when we're about to exit.

See also r240565.

Not reading file history:	uqs
2013-04-26 12:27:30 +00:00
Steven Hartland
562a9d583b Changed ZFS TRIM sysctl from vfs.zfs.trim_disable -> vfs.zfs.trim.enabled
Enabled ZFS TRIM by default

Reviewed by:	pjd (mentor)
Approved by:	pjd (mentor)
MFC after:	2 weeks
2013-04-26 11:24:20 +00:00
Warner Losh
8e015fd886 Octeon 2 (6xxx) and newer CPUs don't use the clock CPU speed for its
I/O clock. Thankfully, the simple executive provies a way to querry
the proper clock that works on all models. Move to asking for the SCLK
via this interface.

This gets the serial console working after we start init and open the
console and set the divisor (which turned the output from good to
bad). I can login on the console now.
2013-04-26 05:42:35 +00:00
Justin Hibbits
70c29e3930 Remove a comment that shouldn't have gone in.
X-MFC-with:	r249864
2013-04-26 05:18:18 +00:00
Peter Grehan
199fee4ea3 Use a thread for the processing of virtio tx descriptors rather
than blocking the vCPU thread. This improves bulk data performance
by ~30-40% and doesn't harm req/resp time for stock netperf runs.

Future work will use a thread pool rather than a thread per tx queue.

Submitted by:	Dinakar Medavaram
Reviewed by:	neel, grehan
Obtained from:	NetApp
2013-04-26 05:13:48 +00:00
Neel Natu
b05c77ff84 Gripe if some <slot,function> tuple is specified more than once instead of
silently overwriting the previous assignment.

Gripe if the emulation is not recognized instead of silently ignoring the
emulated device.

If an error is detected by pci_parse_slot() then exit from the command line
parsing loop in main().

Submitted by (initial version):	Chris Torek (chris.torek@gmail.com)
2013-04-26 02:24:50 +00:00
Sean Bruno
78d58a11fc In the case where the controller supports an sg_list LESS than our predefined
and tuned value, we would advertise the unsupported value to CAM and it would
merrily destroy the controller with way too many IO operations.

This manifests itself in a Zero Memory RAID configuration for a P410 and
possibly other controllers.

Obtained from:	Yahoo! Inc.
MFC after:	2 weeks
2013-04-25 23:10:34 +00:00
Brooks Davis
20e0cc0abb Use the system MAKEOBJDIRPREFIX when running make targets in
mergemaster.  This allows bootstrap verions of tools to be used.

Add a note to UPDATING about this change.

Discussed with:	jhb
Sponsored by:	DARPA, AFRL
MFC after:	5 days
2013-04-25 21:19:50 +00:00
Neel Natu
ade4af668a Remove deprecated APIs to get the total and free memory available to vmm.ko.
These APIs were relevant when memory for virtual machine allocation was
hard partitioned away from the rest of the system but that is no longer
the case. The sysctls that provided this information were garbage collected
a while back.

Obtained from:	NetApp
2013-04-25 20:42:21 +00:00
Joel Dahl
452025cfc6 mdoc: remove superfluous paragraph macro. 2013-04-25 20:23:22 +00:00
Gleb Smirnoff
414676ba31 Fix couple of mbuf leaks in incoming ARP processing. 2013-04-25 17:38:04 +00:00
Warner Losh
0acf27956c Minor whitespace nit 2013-04-25 17:27:13 +00:00
Warner Losh
a16c54dec9 Use the offsets from pcb.h rather than regnum.h to store the registers
in the pcb. setjmp/longjmp in the kernel also used these values, so
continue to use them although their use isn't technically the pcb
register array (matching is all that's important for setjmp/longjmp in
the kernel). Finally, eliminate the old register names from regnum.h.

This is a lexical change only. The non-debug .o files have the same md5.
2013-04-25 17:23:54 +00:00
Gleb Smirnoff
ab210b8057 Don't free memory that is going to be used as error string.
PR:		bin/178121
Submitted by:	Garrett Cooper <yaneurabeya gmail.com>
2013-04-25 16:34:04 +00:00
Gleb Smirnoff
8cc5fb3d91 Restore the ipNetToMedia MIB, that was broken with new ARP commit
in the r186119.

Submitted by:	Konstantin Kukushkin <dark rambler-co.ru>
2013-04-25 16:23:22 +00:00
Steven Hartland
9e68761ce0 Adds Host Protected Area (HPA) support for ATA disks to camcontrol
Reviewed by:	mav
Approved by:	pjd (mentor)
MFC after:	2 weeks
2013-04-25 14:11:38 +00:00
Gleb Smirnoff
4c7a605968 Introduce a pointer to const variable gw, which points either at the
same place as dst, or to the sockaddr in the routing table.

The const constraint of gw makes us safe from modifing routing table
accidentially. And "onstantness" of dst allows us to remove several
bandaids, when we switched it back at &ro->ro_dst, now it always
points there.

Reviewed by:	rrs
2013-04-25 12:42:09 +00:00
Bryan Drewery
55d1493ab3 Fix installworld with DB_FROM_SRC after r249807
Approved by:	bapt
Reported by:	Tom Everett
2013-04-25 12:05:17 +00:00
Andriy Gapon
5d33cbbf9d revert r248644 because of the regression for usdt probes
USDT probes are advertised to kernel by initialization code with
atrribute((constructor))).  It seems that on Solaris the .init-ish code
of the main object is executed before RD_PREINIT point is hit.  On
FreeBSD that is not the case.  And because on FreeBSD there is no other
well-defined point between RD_PREINIT and main() we have to parse a
DTrace script when main is hit, for time being.

A footnote: currently we actually post RD_POSTINIT event, but that's a
bug because the event is triggered by hitting r_debug_state which
happens before any init code is executed.

Reported by:	markj
2013-04-25 07:04:56 +00:00
Martin Matuska
7f168e5ddb Respect the enoent_ok flag if reporting error for holding an non-existing
snapshot.

Related illumos ZFS issue:
  3699 zfs hold or release of a non-existent snapshot does not output error

Reported by:	Steven Hartland <smh@FreeBSD.org>
MFC after:	3 days
2013-04-25 06:55:57 +00:00
Warner Losh
5b1402e5e8 Make it possible to include this file in assembler .S sources. 2013-04-25 06:29:23 +00:00
Warner Losh
369149e740 Use the defines from pcb.h over the ones from regnum.h for this 'C'
code. In theory, the ones from regnum.h should be used only for
assembler code.
2013-04-25 06:28:19 +00:00
Rui Paulo
db62a45d65 Remove the Fortran directories. 2013-04-25 06:05:43 +00:00
Peter Grehan
d3c11f40a5 Add RIP-relative addressing to the instruction decoder.
Rework the guest register fetch code to allow the RIP to
be extracted from the VMCS while the kernel decoder is
functioning.

Hit by the OpenBSD local-apic code.

Submitted by:	neel
Reviewed by:	grehan
Obtained from:	NetApp
2013-04-25 04:56:43 +00:00
Warner Losh
3d8e9c1283 Fix mips64 and mipsn32 bilds by using proper register names. 2013-04-25 04:53:01 +00:00
Justin Hibbits
afd9cb6c6d Introduce kernel coredumps to ppc32 AIM. Leeched from the booke code.
MFC after:	2 weeks
2013-04-25 00:39:43 +00:00
Alan Somers
eccc722c93 Add myself as a src committer
Approved by:	ken (mentor)
2013-04-24 23:38:04 +00:00
Kenneth D. Merry
c4a4d2eb59 Add myself to the committers list. 2013-04-24 23:32:47 +00:00
Eitan Adler
a523543043 Fix the Dvorak Programmer Keymap
PR:		conf/177572
Submitted by:	Sean DuBois <sean@siobud.com>
2013-04-24 21:41:24 +00:00
Jilles Tjoelker
e160aec9a5 getdtablesize(2): Describe what this function actually does.
getdtablesize() returns the limit on new file descriptors; this says nothing
about existing descriptors.

MFC after:	1 week
2013-04-24 21:24:35 +00:00
Martin Matuska
95b6497f5e MFV r249857:
Merge vendor bugfix for a possible deadlock related to async destroy
and improve write performance by introducing a new lock protecting
tx_open_txg.

Illumos ZFS issues:
  3642 dsl_scan_active() should not issue I/O to determine if async
       destroying is active
  3643 txg_delay should not hold the tc_lock

MFC after:	1 week
2013-04-24 21:21:03 +00:00
Martin Matuska
ad0a233d8c Update vendor-sys/illumos/dist to illumos-gate 14019:624fa8f61951
Illumos ZFS issues:
   3642	dsl_scan_active() should not issue I/O to determine if async
	destroying is active
   3643	txg_delay should not hold the tc_lock
2013-04-24 21:00:52 +00:00
Alexander Motin
c33c978915 Move hptmv and mpt drivers shutdown a bit later to the SHUTDOWN_PRI_LAST
stage of shutdown_post_sync.  That should allow CAM to do final cache flush
at the SHUTDOWN_PRI_DEFAULT without using polling magic.

MFC after:	3 days
2013-04-24 19:00:45 +00:00
Randall Stewart
0be23a54cf This fixes the issue with the "randomly changing" default
route. What it was is there are two places in ip_output.c
where we do a goto again. One place was fine, it
copies out the new address and then resets dst = ro->rt_dst;
But the other place does *not* do that, which means earlier
when we found the gateway, we have dst pointing there
aka dst = ro->rt_gateway is done.. then we do a
goto again.. bam now we clobber the default route.

The fix is just to move the again so we are always
doing dst = &ro->rt_dst; in the again loop.

PR:	 174749,157796
MFC after:	1 week
2013-04-24 18:30:32 +00:00
Warner Losh
3a74eaf28f Fix N32/N64 ABIs to use proper registers after recent changes.
Pointy Hat to: imp
2013-04-24 18:00:28 +00:00
Dimitry Andric
dbfc29017c When rebooting (exiting) from the BTX loader, make sure to restore the
GDT from the correct segment, otherwise a triple fault would be caused.
In some virtual environments (VMware, VirtualBox, etc) this could lead
to a unhandled error or hang in the guest emulation software.

Thanks to avg and jhb for a few hints in the right direction.

Noticed by:	Jeremy Chadwick <jdc@koitsu.org> (and many others)
MFC after:	1 week
2013-04-24 17:20:45 +00:00
Hans Petter Selasky
ffae621edd Fix for duplicate sample rate detection after recent patches. 2013-04-24 16:52:03 +00:00
Hans Petter Selasky
bf02c706c7 Fix the USB audio feedback endpoint algorithm. There should not
be any need to bias the returned value.

Reported by:	Craig Leres <leres@ee.lbl.gov>
2013-04-24 16:22:53 +00:00
Andre Oppermann
2ebcc8ac4a Base the calculation of maxmbufmem in part on kmem_map size
instead of kernel_map size to prevent kernel memory exhaustion
by mbufs and a subsequent panic on physical page allocation
failure.

On architectures without a direct map all mbuf memory (except
for jumbo mbufs larger than PAGE_SIZE) comes from kmem_map.
It is the limiting factor hence.

For architectures with a direct map using the size of kmem_map
is a good proxy of available kernel memory as well.  If it is
much smaller the mbuf limit may be sub-optimal but remains
reasonable, while avoiding panics under exhaustion.

The overall mbuf memory limit calculation may be reconsidered
again later, however due to the many different mbuf sizes and
different backing KVM maps it is a tricky subject.

Found by:	pho's new network stress test
Pointed out by:	alc (kmem_map instead of kernel_map)
Tested by:	pho
2013-04-24 13:54:55 +00:00
Dag-Erling Smørgrav
733706130a Merge updated "no such identity file" patch.
PR:		bin/178060
2013-04-24 12:36:37 +00:00