summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* Improve transmit sending offload, TSO, algorithm in general.hselasky2014-09-2216-43/+264
| | | | | | | | | | | | | | | | | The current TSO limitation feature only takes the total number of bytes in an mbuf chain into account and does not limit by the number of mbufs in a chain. Some kinds of hardware is limited by two factors. One is the fragment length and the second is the fragment count. Both of these limits need to be taken into account when doing TSO. Else some kinds of hardware might have to drop completely valid mbuf chains because they cannot loaded into the given hardware's DMA engine. The new way of doing TSO limitation has been made backwards compatible as input from other FreeBSD developers and will use defaults for values not set. Reviewed by: adrian, rmacklem Sponsored by: Mellanox Technologies MFC after: 1 week
* Simplify legacy reservation handling. Drop it on I_T nexus loss.mav2014-09-222-27/+19
|
* svn revisions r269964 and r269963 seemed to have impaired small memorysbruno2014-09-221-0/+2
| | | | | | | | | | | | | footprint systems(32M/64M) and didn't leave enough free memory to load modules when it was setting up page tables that for sizes that are never used on these smallish boards. Set kmem_zmax to PAGE_SIZE on these smaller systems (< 128M) to keep this from happening. Verified on mips32 h/w. PR: 193465 Submitted by: delphij Reviewed by: adrian
* Don't report unsupported FUA_NV bit set in READ/WRITE commands as error.mav2014-09-221-6/+6
| | | | While this bit is obsolete in SBC-3, SBC-2 allowed to silently ignore it.
* Report proper errors codes for unsupported SERVICE ACTION values.mav2014-09-221-12/+22
|
* Polish INQUIRY command fields validation.mav2014-09-221-31/+9
|
* Ignore a test program which doesn't compile at the moment. It will bemarkj2014-09-211-1/+2
| | | | addresed properly when integrating the DTrace tests with Kyua.
* Add some ATF tests for libproc.markj2014-09-2110-201/+423
| | | | | Differential Revision: D710 Reviewed by: jmmv, ngie, rpaulo
* sh(1): Clarify that assignments before commands do not affect expansions.jilles2014-09-211-2/+3
| | | | | PR: 193759 MFC after: 1 week
* Output boot code warning when zpool upgrade -a is used to add features.smh2014-09-211-3/+8
| | | | | | | | In the case where new features where enabled by a zpool upgrade -a the boot code warning wasn't output. Submitted by: Jan Kokemueller MFC after: 3 days
* Add a few missing llvm/clang patches, update the other ones to be abledim2014-09-217-61/+261
| | | | | | | to apply with the same patch options onto a fresh upstream llvm/clang 3.4.1 checkout, and use approximately the same header tempate for them. MFC after: 3 days
* Allow SUBPAGE CODE field in MODE SENSE commands.mav2014-09-211-2/+2
|
* Turns out -1 is a perfectly valid error number, ERESTART. Remove uselesstrasz2014-09-211-11/+1
| | | | | | code written under assumption that it wasn't. Sponsored by: The FreeBSD Foundation
* Mention read_ahead tunables/sysctls.mav2014-09-211-3/+5
| | | | MFC after: 1 month
* Update and clarify comments. Remove the useless counter for impossible, butkib2014-09-212-12/+13
| | | | | | | seen in wild situation (on buggy hypervisors). In collaboration with: bde MFC after: 1 week
* Fix build.hrs2014-09-211-0/+2
|
* Fix a typo.hrs2014-09-211-1/+1
|
* Add a change missing in r271916.hrs2014-09-211-2/+2
|
* Fix a bug which could make routed(8) daemon exit by sending a special RIPhrs2014-09-215-7/+35
| | | | | | query from a remote machine, and disable accepting it by default. This requests a routed(8) daemon to dump routing information base for debugging purpose. An -i flag to enable it has been added.
* - Virtualize interface cloner for gre(4). This fixes a panic when destroyinghrs2014-09-213-40/+50
| | | | | | a vnet jail which has a gre(4) interface. - Make net.link.gre.max_nesting vnet-local.
* Virtualize interface cloner for gif(4). This fixes a panic when destroyinghrs2014-09-211-23/+37
| | | | a vnet jail which has a gif(4) interface.
* Make net.add_addr_allfibs vnet-local.hrs2014-09-212-6/+8
|
* Hide LRO code under #ifdef INET/INET6 to allow NOIP kernel to compile.bz2014-09-211-1/+12
| | | | | | | While the code comments indicate that LRO is currently only supported for IPv4/TCP, we'll still compile it in for IPv6. MFC after: 3 days
* Add unit tests for mkimg(1):marcel2014-09-20116-0/+3759
|
* Fix a problem that reply packets are not received when -i T option is sethrs2014-09-202-232/+166
| | | | | | | | | | | | and (T < RTT). - Use select(2) for timeout instead of interval timer. Remove poll(2) support. - Use sigaction(2) instead of signal(3). - Exit in SIGINT handler when two signals are received and doing reverse DNS lookup as ping(8) does. - Remove redundant variables used for getaddrinfo(3). PR: 151023
* Revert changes in r269180. It could cause -c N option to enter anhrs2014-09-201-8/+2
| | | | | | infinite loop if no reply packet is received. PR: 151023
* Use callout(9) instead of timeout(9).jhb2014-09-204-17/+12
| | | | Reviewed by: emax
* Add a man page for the cgem(4) driver.ian2014-09-202-2/+301
| | | | | Submitted by: Thomas Skibo <ThomasSkibo@sbcglobal.net> Reviewed by: wkoszek@, Yonghyeon PYUN <pyunyh@gmail.com>
* Make the ARM MPCore Timer driver work with published standard FDT bindings.ian2014-09-201-106/+183
| | | | | | | | | | | | | | | | | | | | | | | We've always considered the mpcore timers to be a single monolithic device and we defined our own fdt binding for it with our own compat string. The published bindings treat the timers as two separate devices, a global timer and a "timer-watchdog" device for the per-cpu private timers. Thus our binding has two tuples in the regs property, one set of registers for the global timer and one for the private timers. The published bindings have two separate devices, each with a single set of registers. (Note that we don't use the optional watchdog feature of the hardware.) These changes add the compat strings for the published bindings. If our own compat string appears, we expect to get two sets of memory resources. For the published bindings, there's only one set of memory resources, and only the private timers have an associated interrupt. The other major change is that there can no longer be a single global var for the softc pointer because now there may be multiple devices at runtime. Since the global timer is used only as a timecounter and the private timers only as eventtimers, and there will only be one of each, those are now the pointers which are global, and the priv fields of those structures backlink to the device softc.
* Fix typo in comment.ganbold2014-09-201-1/+2
| | | | | | Disable AHB clock gate for ehci0. Approved by: stas (mentor)
* Properly hide calls to ARP under #ifdef INET to allow IPv6-onlybz2014-09-202-0/+4
| | | | | | kernels to compile. MFC atfer: 3 days
* Make gcc happy by initialising the variable only set in a couple ofbz2014-09-201-0/+1
| | | | | | | case statements without a default. MFC after: 2 days X-MFC with: r271871
* Don't install /etc/rc.d/ftp-proxy unless MK_PF == yesngie2014-09-202-1/+5
| | | | | MFC after: 2 weeks Sponsored by: EMC / Isilon Storage Division
* Don't install /etc/rc.d/rwho unless MK_RCMDS == yesngie2014-09-202-1/+5
| | | | | MFC after: 2 weeks Sponsored by: EMC / Isilon Storage Division
* Sort the optional rc.d scripts by their knobsngie2014-09-201-11/+11
| | | | | MFC after: 1 week Sponsored by: EMC / Isilon Storage Division
* Add some more KTR events to help debugging.neel2014-09-202-1/+8
|
* MSR_KGSBASE is no longer saved and restored from the guest MSR save area. Thisneel2014-09-201-7/+0
| | | | | | | | behavior was changed in r271888 so update the comment block to reflect this. MSR_KGSBASE is accessible from the guest without triggering a VM-exit. The permission bitmap for MSR_KGSBASE is modified by vmx_msr_guest_init() so get rid of redundant code in vmx_vminit().
* Populate the device info string with _PXM (proximity domain) information.adrian2014-09-202-8/+22
| | | | | | | This is primarily useful for debugging right now - it'll show up in devinfo. Reviewed by: jhb
* Restructure the MSR handling so it is entirely handled by processor-specificneel2014-09-2015-383/+273
| | | | | | | | | | | | | | | | | | | code. There are only a handful of MSRs common between the two so there isn't too much duplicate functionality. The VT-x code has the following types of MSRs: - MSRs that are unconditionally saved/restored on every guest/host context switch (e.g., MSR_GSBASE). - MSRs that are restored to guest values on entry to vmx_run() and saved before returning. This is an optimization for MSRs that are not used in host kernel context (e.g., MSR_KGSBASE). - MSRs that are emulated and every access by the guest causes a trap into the hypervisor (e.g., MSR_IA32_MISC_ENABLE). Reviewed by: grehan
* Fix up the EDMA RX setup path to correctly initialise and reset the RX FIFO.adrian2014-09-203-83/+133
| | | | | | | | | | | | | | | | | | | | | | | | | | | The original code was .. well, slightly more than incorrect. It showed up as stalled RX queues if the NIC needed to be frequently reinitialised (eg during scans.) This is inspired by work done by Matt Dillon over at the DragonflyBSD project. So: * track when EDMA RX has been stopped and when the MAC has been reset; * re-initialise the ring only after a reset; * track whether RX has been stopped/started - just for debugging now; * don't bother with the RX EOL stuff for EDMA - we don't need the interrupt at all. We also don't need to disable/enable the interrupt or start DMA - once new frames are pushed into the ring via the normal RX path, it'll just restart RX DMA on its own. Tested: * AR9380, STA mode * AR9380, AP mode * AR9485, STA mode * AR9462, STA mode
* Quieten the AR9300 HAL attach a little - the setupXTxDesc methodadrian2014-09-201-0/+2
| | | | | | is only called during attach now. The EDMA TX path never calls this - only the legacy TX path does.
* Quieten the AR9300 HAL probe a bit.adrian2014-09-201-1/+1
|
* Don't double-call the ar9300 stub function method.adrian2014-09-201-3/+0
|
* Explain how to load the malo(4) firmware on boot.gavin2014-09-201-2/+11
| | | | | | PR: 193653 Submitted by: Juan Ramón Molina Menor MFC after: 3 days
* Style (9) fixes, spaces to tabs.smh2014-09-191-50/+50
| | | | | | | | Fix style issues noticed when I went to MFC r257820 MFC after: 2 days X-MFC-With: r257820 Sponsored by: Multiplay
* Fix partition alignment and image rounding when any of -P (block size),marcel2014-09-1910-54/+54
| | | | | | | | | | | | | | | | | | | | | | | | | | | | -T (track size) or -H (number of heads) is given: o scheme_metadata() always rounded to the block size. This is not always valid (e.g. vtoc8 that must have partitions start at cylinder boundaries). o The bsd and vtoc8 schemes "resized" the image to make it match the geometry, but since the geometry is an approximation and the size of the image computed from cylinders * heads * sectors is always smaller than the original image size, the partition information ran out of bounds. The fix is to have scheme_metadata() simply pass it's arguments to the per-scheme metadata callback, so that schemes not only know where the metadata is to go, but also what the current block address is. It's now up to the per-scheme callback to reserve room for metadata and to make sure alignment and rounding is applied. The BSD scheme now has the most elaborate alignment and rounding. Just to make the point: partitions are aligned on block boundaries, but the image is rounded to the next cyclinder boundary. vtoc8 now properly has all partitions aligned (and rounded) to the cyclinder boundary. Obtained from: Juniper Networks, Inc. MFC after: 3 days
* awk: Use random(3) instead of rand(3)pfg2014-09-193-5/+7
| | | | | | | | | | | | | While none of them is considered even near to cryptographic level, random(3) is a better random generator than rand(3). Use random(3) for awk as is done in other systems. Thanks to Chenguang Li for discussing this in the lists and submitting the patch upstream. PR: 193147 MFC after: 5 weeks
* unistd: drop argument names from setpgrp(3) prototype.pfg2014-09-191-1/+1
| | | | | | | They are useless and don't match the style of the header. While here adjust the comment with tabs. Suggested by: kevinlo
* When populating the dvd/packages/ directory, creategjb2014-09-191-0/+6
| | | | | | | | | | | | a symlink to All/pkg-*.txz in the Latest/ directory. This allows 'pkg bootstrap' to work out-of-box if the REPOS_DIR environment is properly set. Tested on: stable/10@r271848 MFC after: 3 days X-MFC-10.1: yes Sponsored by: The FreeBSD Foundation
* Reference the test case "packs" to fix warningsrodrigc2014-09-192-0/+6
| | | | | | | | Put the various atf_tc_pack_t variables generated by ATF_TC to use in the macros_h_test.c file so that we prevent some build warnings (and thus some build errors) with clang and -Wunused. Obtained from: atf (git 70126d013db25c72539803f19f7e00baddacdf51)
OpenPOWER on IntegriCloud