the controller to workaround silicon bug of i82557. Each reset will
re-establish link which in turn triggers MII status change
callback. The callback will try to reconfigure controller if the
controller is not i82557 to enable flow-control. This caused
endless link UP/DOWN when the workaround was enabled on non-i82557
controller.
To fix the issue, apply RX lockup workaround only for i82557.
Previously it blindly checked undocumented EEPROM location such
that it sometimes enabled the workaround for other controllers. At
this time, only i82557 is known to have the silicon bug.
This fixes a regression introduced in r215906 which enabled flow
control support for all controllers except i82557.
Reported by: Karl Denninger (karl <> denninger dot net)
Tested by: Karl Denninger (karl <> denninger dot net)
MFC after: 3 days
the dev.fxp.%d.noflow tunable as the same effect can now be achieved with
ifconfig(8) by setting the flowcontrol media option as desired (besides
the tunable never having a chance to actually enable flow control support
so far).
In joint forces with: yongari
VLAN hardware tagging to make TSO work over VLAN. So if VLAN
hardware tagging is disabled explicitly clear TSO over VLAN. While
I'm here allow disabling VLAN TX checksum offloading.
Tested by: Liudas < liudasb <> centras dot lt >
MFC after: 10 days
the NIC drivers as well as the PHY drivers to take advantage of the
mii_attach() introduced in r213878 to get rid of certain hacks. For
the most part these were:
- Artificially limiting miibus_{read,write}reg methods to certain PHY
addresses; we now let mii_attach() only probe the PHY at the desired
address(es) instead.
- PHY drivers setting MIIF_* flags based on the NIC driver they hang
off from, partly even based on grabbing and using the softc of the
parent; we now pass these flags down from the NIC to the PHY drivers
via mii_attach(). This got us rid of all such hacks except those of
brgphy() in combination with bce(4) and bge(4), which is way beyond
what can be expressed with simple flags.
While at it, I took the opportunity to change the NIC drivers to pass
up the error returned by mii_attach() (previously by mii_phy_probe())
and unify the error message used in this case where and as appropriate
as mii_attach() actually can fail for a number of reasons, not just
because of no PHY(s) being present at the expected address(es).
Reviewed by: jhb, yongari
header parser uses m_pullup(9) to get access to mbuf chain.
m_pullup(9) can allocate new mbuf chain and free old one if the
space left in the mbuf chain is not enough to hold requested
contiguous bytes. Previously drivers can use stale ip/tcp header
pointer if m_pullup(9) returned new mbuf chain.
Reported by: Andrew Boyer (aboyer <> averesystems dot com)
MFC after: 10 days
RFA. Also drop frames that have either CRC error or alignment
error. Normally bad frames are not received at all. But controllers
running in promiscuous mode will receive bad frames. 82557 will
also receive bad frames to receive VLAN oversized frames.
While I'm here mark RNR condition if driver happen to see RNR in
RFA status and restart RU to receive frames again. Because driver
checks all received frames in RX loop, RNR condition could be set
in the middle of RX processing. Just relying on RNR interrupt was
not enough.
This change fixes "Memory modified after free" issue when fxp(4)
is running as a member of if_bridge(4).
Tested by: Larry Baird <lab <> gta dot com>
MFC after: 5 days
fxp(4) already used to extract most hardware MAC statistics but it
didn't show them. With this change, all MAC statistics counters
are exported. Because there are a couple of new counters for 82558
and 82559, enable extended MAC statistics functionality to get
these counters. Accoring to public data sheet, 82559 MAC statistics
return 24 DWORD counters(3 counters are unknown at this moment) so
increase MAC counter structure to meet the MAC statistics block size.
The completion of MAC counter dump is now checked against
FXP_STATS_DR_COMPLETE status code which is appended at the end of
status block. Previously fxp(4) ignored the status of the
FXP_SCB_COMMAND_CU_DUMPRESET command. fxp(4) does not wait for the
completion of pending command before issuing
FXP_SCB_COMMAND_CU_DUMPRESET. Instead it skips the command and try
it next time. This scheme may show better performance but there is
chance to loose updated counters after stopping controller. So make
sure to update MAC statistics in fxp_stop().
While I'm here move sysctl node creation to fxp_sysctl_node().
Tested by: Larry Baird < lab <> gta dot com >
table. The default size of the configuration table was 22 bytes. To
use extended feature of 82550/82551 the configuration table size
was expanded to 32 bytes. The added configuration for 82550/82551
specifies VLAN hardware tagging and IPSec configuration as well as
TCO.
To make configuration easier fxp(4) used a configuration template
and the template was copied to configuration table. After that,
some parameters of the configuration table was changed depending on
controller type and operation mode. However the size of template
was 22 bytes so some configuration parameters were not properly
initialized on 82550/82551.
Fix this by increasing the template size. For 82557, 82558 and
82559 the size of the configuration is still 22 bytes.
unicast and the other for multicast. To receive multicast frames
that host didn't join in promiscuous mode, driver have to set
promiscuous mode for multicast frames as well.
The Open Source Software Developer Manual for i8255x was not clear
how to handle promiscuous mode.
PR: kern/145905
MFC after: 5 days
offloading. Datasheet said nothing about the limitation of 82559ER
except WOL. Explicitly disable Rx checksum offloading for
controllers that is known to lack the capability.
PR: kern/138135
Tested by: Gooderum, Mark < mgooderum <> websense dot com >
IF_ADDR_UNLOCK() across network device drivers when accessing the
per-interface multicast address list, if_multiaddrs. This will
allow us to change the locking strategy without affecting our driver
programming interface or binary interface.
For two wireless drivers, remove unnecessary locking, since they
don't actually access the multicast address list.
Approved by: re (kib)
MFC after: 6 weeks
To detect which controller is ICH based one, add a new member
variable ich to struct fxp_ident and move the struct to
if_fxpvar.h. Since I've faked controller revision, don't allow
microcode loading for ICH based controllers.
With this change all ICH based controllers will have WOL and Rx
checksum offload capability.
PR: kern/135451
Tested by: Alexey Shuvaev ( shuvaev <> physik dot uni-wuerzburg dot de ),
pluknet ( pluknet <> gmail dot com ),
Gary Jennejohn ( gary.jennejohn <> freenet dot de )
not allow multicast filter programming when controller is busy to
send/receive frames. So it used to mark need_mcsetup bit and defer
multicast filter programming until controller becomes idle state.
To detect when the controller is idle fxp(4) relied on Tx
completion interrupt with NOP command and fxp_start_body and
fxp_intr_body had to see whether pending multicast filter
programming was requested. This resulted in very complex logic and
sometimes it did not work as expected.
Since the controller should be in idle state before any multicast
filter modifications I changed it to reinitialize the controller
whenever multicast filter programming is required. This is the same
way what OpenBSD and NetBSD does. Also I added IFF_DRV_RUNNING
check in ioctl handler so controller would be reinitialized only if
it is absolutely needed.
With this change I guess we can remove fxp(4) DELAY hack in ifioctl
for IPv6 case.
upper stack in fxp_start_body().
fxp(4) drops driver lock in Rx path so check the fxp(4) is still
running after reacquiring driver lock in Rx path. Also don't
invoke fxp_intr_body if fxp(4) is not running. With this change
there is no need to set suspend bit in device attach phase.
hardware reset in attach phase. Selective reset does not clear
configured parameters so I think full hardware reset is required.
To prevent PCI bus lock-up, do selective reset first which will get
off the controller from PCI bus and request software reset after
selective reset. Software reset will unmask interrupts so disable
it after the reset.
common mbuf dma tag for both Tx and Rx path but Rx buffer should
have single DMA segment and maximum buffer size of the segment
should be less than MCLBYTES.
fxp(4) also have to check Tx completion status which was updated by
DMA so we need BUS_DMASYNC_PREREAD and BUS_DMASYNC_POSTWRITE
synchronization in Tx path. Fix all misuse of bus_dmamap_sync(9) in
fxp(4). I guess this change shall fix occasional driver breakage in
PAE environments.
While I'm here add error messages of dma tag/buffer creation and
correct messages.
driver should read updated status back after issuing a SCB command.
To send a command to controller and read updated status back,
driver should synchronize both memory read and write operations
with device. Fix bus_dmamap_sync operation specifier used in
fxp_dma_wait() by adding both memory read and memory write
operations.
CPU for too long period than necessary. Additively, interfaces are kept
polled (in the tick) even if no more packets are available.
In order to avoid such situations a new generic mechanism can be
implemented in proactive way, keeping track of the time spent on any
packet and fragmenting the time for any tick, stopping the processing
as soon as possible.
In order to implement such mechanism, the polling handler needs to
change, returning the number of packets processed.
While the intended logic is not part of this patch, the polling KPI is
broken by this commit, adding an int return value and the new flag
IFCAP_POLLING_NOCOUNT (which will signal that the return value is
meaningless for the installed handler and checking should be skipped).
Bump __FreeBSD_version in order to signal such situation.
Reviewed by: emaste
Sponsored by: Sandvine Incorporated
Remote host can advertise smaller MSS than that of sender so upper
stack might have adjusted the MSS which in turn generates IP
packets that are less size than that of interface MTU.
Reported by: Bjoern Koenig ( bkoenig <> alpha-tierchen dot de )
Tested by: Bjoern Koenig ( bkoenig <> alpha-tierchen dot de )
MFC after: 3 days
payload length in TSO case. Leaving unused TBD also seem to cause
SCB timeouts under certain conditions when TSO/non-TSO traffics
are active at the same time.
offload for VLAN frames are also supported. The VLAN hardware
assistance is available only on 82550/82551 based controllers.
While I'm here change the confusing name of bit1 in byte 22 of
configuration block to vlan_drop_en. The bit controls whether
hardware strips VLAN tagged frame or not. Special thanks to wpaul
who sent valuable VLAN related information to me.
Tested on: i386, sparc64
events. Just reading PMDR register was not enough to have fxp(4)
immuninize against received magic packets during system boot.
Tested by: Alexey Shuvaev < shuvaev <> physik DOT uni-wuerzburg DOT de >
controllers. ICH based controllers are treated as 82559. 82557,
earlier revision of 82558 and 82559ER have no WOL capability.
o WOL support requires help of a firmware so add check whether
hardware is capable of handling magic frames by reading EEPROM.
o Enable accepting WOL frames only when hardware is about to
suspend or shutdown. Previously fxp(4) used to allow receipt of
magic frame under normal operation mode which could cause
hardware hang if magic frame is received by hardware. Datasheet
clearly states driver should not allow WOL frames under normal
operation mode.
o Disable WOL frame reception in device attach so have fxp(4)
immunize against system hang which can be triggered by magic
packets when the hardware is not in fully initialized state.
o Don't reset all hardware configuration data in fxp_stop()
otherwise important configuration data is lost and this would
reset WOL configuration to default state which in turn cause
hardware hang on receipt of magic frames. To fix the issue,
preserve hardware configuration data by issuing a selective
reset.
o Explicitly disable interrupts after issuing selective reset as
reset may unmask interrupts.
Tested by: Alexey Shuvaev < shuvaev <> physik DOT uni-wuerzburg DOT de >
o Configure controller to use dynamic TBD as TSO requires that
operation mode.
o Add a dummy TBD to tx_cb_u as TSO can access one more TBD in TSO
operation.
o Increase a DMA segment size to 4096 to hold a full IP segment
with link layer header.
o Unlike other TSO capable controllers, 82550/82551 does not
modify the first IP packet in TSO operation so driver should
create an IP packet with proper header. Subsequent IP packets
are generated from the header information in the first IP packet
header. Likewise pseudo checksum also should be computed by
driver for the first packet.
o TSO requires one more TBD to hold total TCP payload. To make
code simple for TSO/non-TSO case, increase the index of the
first available TBD array.
o Remove KASSERT that checks the size of a DMA segment should be
less than or equal to MCLBYTES as it's no longer valid in TSO.
o Tx threshold and number of TBDs field is used to store MSS in
TSO. So don't set the Tx threshold in TSO case.
82559 or later controllers added simple checksum calculation logic
in RU. For backward compatibility the computed checksum is appended
at the end of the data posted to Rx buffer. This type of simple
checksum calculation support had been used on several vendors such
as Sun HME/GEM, SysKonnect GENESIS and Marvell Yukon controllers.
Because this type of checksum offload support requires parsing of
received frame and pseudo checksum calculation with software
routine it still consumes more CPU cycles than that of full-fledged
checksum offload controller. But it's still better than software
checksum calculation.
Rx buffer and loads DMA map. Also add a function
fxp_discard_rfabuf that handles reusing Rx buffer/DMA map. With
this change fxp_add_rfabuf just handles appending a new RFA to
existing chain.
o Initialize mbuf length in fxp_new_rfabuf.
o Don't reset rnr and have fxp(4) handle received frames even if
it couldn't allocate new Rx buffer. This will make fxp(4) reload
updated RFA under rnr case. The rnr would still be reset to 0 if
polling is active and fxp(4) processed number of allowed Rx
events.
o Update if_iqdrops if fxp(4) couldn't allocate Rx buffer.
Previously fxp(4) used to try to reuse Rx buffer when new buffer
allocation is failed. But fxp(4) didn't take into account loaded
DMA map such that the same DMA map was loaded again without
unloading the map. There is no reason to unload the loaded map and
reload the same map again, just reusing the map is enough. I
believe the spare DMA map in softc was introduced to implement this
behaviour. Also fxp(4) used to stop Rx processing if once Rx buffer
allocation or DMA map load fails which in turn resulted in losing
incoming frames under heavy network load. With this change fxp(4)
should survive from resource shortage condition.
IFF_DRV_OACTIVE to note resource shortage to upper stack.
- Don't count number of mbuf chains. Default 32 DMA segments for a
frame is enough for most cases. If bus_dmamap_mbuf_sg fails use
m_collapse(9) to collapse the mbuf chain instead of relying on
expensive m_defrag(9).
- Move bpf handling to fxp_start_body() which is supposed to be
more appropriate place.
- Always arm watchdog timer whenever a new Tx request is made.
Previously fxp(4) used to arm watchdog timer only when
FXP_CXINT_THRESH-th Tx request is made. Because fxp(4) does not
rely on Tx interrupt to reclaim transmitted mbufs it's better to
arm watchdog timer to detect potential lockups.
- Add more aggresive Tx buffer reclaiming in fxp_start_body to make
room for new Tx requests. Since fxp(4) does not request Tx
completion interrupt for every frames it's necessary to clean
TXCBs in advance to saturate link.
- Make fxp(4) try to start more packets transmitting regardless of
interrupt type in fxp_intr_body.
check to fxp_txeof(). While I'm here unarm watchdog timer only if
there are no pending queued Tx requests.
Previously the watchdog timer was unarmed whenever Tx interrupt is
raised. This could be resulted in hiding root cause of watchdog
timeouts.
checksum offload configuration. Now checksum offload can be
controlled by ifconfig(8).
While I'm here add an additional check for interface capabilities
before applying user's request.
exhaustion is encountered. There was a fix made previously for this
problem but the solution (breaking out of the receive loop) does not
seem to work. mbuf reuse strategy is already adopted by other drivers
such as if_bge. The problem was recreated and the patch is also
verified in the same test environment.
sparc64 GENERIC and the sound device drivers known working on sparc64
to use bus_get_dma_tag() to obtain the parent DMA tag so we can get rid
of the sparc64_root_dma_tag kludge eventually. Except for ath(4), sk(4),
stge(4) and ti(4) these changes are runtime tested (unless I booted up
the wrong kernels again...).
