| created | 2019-04-30T03:43:25Z |
|---|---|
| begin | 2019-03-01T00:00:00Z |
| end | 2019-03-02T00:00:00Z |
| path | src/sys |
| commits | 11 |
| date | 2019-03-01T01:46:18Z | |||
|---|---|---|---|---|
| author | cheloha | |||
| files | src/sys/sys/mman.h | log | diff | annotate |
| src/sys/uvm/uvm.h | log | diff | annotate | |
| src/sys/uvm/uvm_extern.h | log | diff | annotate | |
| src/sys/uvm/uvm_map.c | log | diff | annotate | |
| src/sys/uvm/uvm_mmap.c | log | diff | annotate | |
| src/sys/uvm/uvm_unix.c | log | diff | annotate | |
| message |
New mmap(2) flag: MAP_CONCEAL. MAP_CONCEAL'd memory is not written to disk in the event of a core dump. It may grow other qualities in the future. Wanted by libressl, probably useful elsewhere, too. Prompted by deraadt@, concept from deraadt@/kettenis@. With input from deraadt@, cjeker@, kettenis@, otto@, bcook@, matthew@, guenther@, djm@, and tedu@. ok otto@ deraadt@ |
|||
| date | 2019-03-01T01:58:17Z | |||
|---|---|---|---|---|
| author | jmatthew | |||
| files | src/sys/dev/pci/mfii.c | log | diff | annotate |
| message |
Background init progress and consistency check progress are reported in different fields. Use the background init progress when that's the active task. lack of progress data noticed by deraadt@ tested on SAS2208 and SAS3508 ok dlg@ |
|||
| date | 2019-03-01T04:47:32Z | |||
|---|---|---|---|---|
| author | dlg | |||
| files | src/sys/net/if.c | log | diff | annotate |
| message |
rework how ifiq_input decides the stack is busy and whether it should drop previously ifiq_input uses the traditional backpressure or defense mechanism and counts packets to decide when to shed load by dropping. currently it ends up waiting for 10240 packets to get queued on the stack before it would decide to drop packets. this may be ok for some machines, but for a lot this was too much. this diff reworks how ifiqs measure how busy the stack is by introducing an ifiq_pressure counter that is incremented when ifiq_input is called, and cleared when ifiq_process calls the network stack to process the queue. if ifiq_input is called multiple times before ifiq_process in a net taskq runs, ifiq_pressure goes up, and ifiq_input uses a high value to decide the stack is busy and it should drop. i was hoping there would be no performance impact from this change, but hrvoje popovski notes a slight bump in forwarding performance. my own testing shows that the ifiq input list length grows to a fraction of the 10240 it used to get to, which means the maximum burst of packets through the stack is smoothed out a bit. instead of big lists of packets followed by big periods of drops, we get relatively small bursts of packets with smaller gaps where we drop. the follow-on from this is to make drivers implementing rx ring moderation to use the return value of ifiq_input to scale the ring allocation down, allowing the hardware to drop packets so software doesnt have to. |
|||
| date | 2019-03-01T04:47:33Z | |||
|---|---|---|---|---|
| author | dlg | |||
| files | src/sys/net/ifq.c | log | diff | annotate |
| src/sys/net/ifq.h | log | diff | annotate | |
| message |
rework how ifiq_input decides the stack is busy and whether it should drop previously ifiq_input uses the traditional backpressure or defense mechanism and counts packets to decide when to shed load by dropping. currently it ends up waiting for 10240 packets to get queued on the stack before it would decide to drop packets. this may be ok for some machines, but for a lot this was too much. this diff reworks how ifiqs measure how busy the stack is by introducing an ifiq_pressure counter that is incremented when ifiq_input is called, and cleared when ifiq_process calls the network stack to process the queue. if ifiq_input is called multiple times before ifiq_process in a net taskq runs, ifiq_pressure goes up, and ifiq_input uses a high value to decide the stack is busy and it should drop. i was hoping there would be no performance impact from this change, but hrvoje popovski notes a slight bump in forwarding performance. my own testing shows that the ifiq input list length grows to a fraction of the 10240 it used to get to, which means the maximum burst of packets through the stack is smoothed out a bit. instead of big lists of packets followed by big periods of drops, we get relatively small bursts of packets with smaller gaps where we drop. the follow-on from this is to make drivers implementing rx ring moderation to use the return value of ifiq_input to scale the ring allocation down, allowing the hardware to drop packets so software doesnt have to. |
|||
| date | 2019-03-01T06:15:59Z | |||
|---|---|---|---|---|
| author | dlg | |||
| files | src/sys/dev/pci/if_ix.c | log | diff | annotate |
| message | bus_dmamap_sync when we fill the rx ring, not for ever slot on the ring. | |||
| date | 2019-03-01T07:39:56Z | |||
|---|---|---|---|---|
| author | stsp | |||
| files | src/sys/dev/ic/athn.c | log | diff | annotate |
| message |
Make athn(4) configure timing of control frames explicitly. We now update the chip with values for SIFS interval and ACK/RTS/CTS timeout when configuration changes occur, like Linux ath9k does. Tested by myself, juanfra@, jmc@ on AR9280 and AR9271. |
|||
| date | 2019-03-01T08:09:00Z | |||
|---|---|---|---|---|
| author | stsp | |||
| files | src/sys/net80211/ieee80211_input.c | log | diff | annotate |
| message |
Pick up the AP's 11g ERP protection setting properly in 11n client mode. ok phessler@ |
|||
| date | 2019-03-01T08:13:11Z | |||
|---|---|---|---|---|
| author | stsp | |||
| files | src/sys/net80211/ieee80211_node.c | log | diff | annotate |
| src/sys/net80211/ieee80211_node.h | log | diff | annotate | |
| src/sys/net80211/ieee80211_var.h | log | diff | annotate | |
| message |
Improve handling of HT protection for 'mode 11n' hostap. Consider the mere presence of HTCAP IEs as indicator of 11n support, rather than checking advertised Rx MCS. Fixes some devices being misclassified as 11a/g, causing HT protection to be enabled even though it is not needed. Use CTS-to-self frames for HT protection instead of RTS/CTS. These changes make my athn(4) AR9280 hostap perform much better. ok phessler@ |
|||
| date | 2019-03-01T09:45:40Z | |||
|---|---|---|---|---|
| author | dlg | |||
| files | src/sys/dev/pci/if_ixl.c | log | diff | annotate |
| message |
use ifiq_input instead of if_input call if_rxr_livelocked if ifiq_input says to slow down |
|||
| date | 2019-03-01T10:02:44Z | |||
|---|---|---|---|---|
| author | dlg | |||
| files | src/sys/dev/pci/if_em.c | log | diff | annotate |
| src/sys/dev/pci/if_em.h | log | diff | annotate | |
| message |
use a timeout to refill the rx ring when it's empty. em had rxr, but didn't use a timeout cos it claimed to generate an RX overflow interrupt when packets fell off slots in the ring. turns out that's a lie on at least one chip, so add the timeout like other drivers. this was hit by mlarkin@, who had nfs and bufs steal all the packets and memory for packets from em, which didn't recover after the memory had been released back to the system. |
|||
| date | 2019-03-01T19:15:59Z | |||
|---|---|---|---|---|
| author | patrick | |||
| files | src/sys/dev/usb/xhci.c | log | diff | annotate |
| message |
Since ring->index points to the next free slot, once we reach index zero we know that the last non-link TRB has been written and we can look at the flags to set the chain bit in the link TRB. Since we will now toggle the cycle bit on the first TRB of a ring, set it on the ring reset. Tested by jcs@, jsg@ and visa@ "commit it" jcs@ |
|||