re: Sync with Realtek's re-196.04; and fix re_start/re_tx_collect.

- Add several new chips' support, noticably RTL8125 support.
- Reorder chip related initializaton on the attach path a bit;
mainl

re: Sync with Realtek's re-196.04; and fix re_start/re_tx_collect.

- Add several new chips' support, noticably RTL8125 support.
- Reorder chip related initializaton on the attach path a bit;
mainly to make sure ifnet is initialized.
- In re_start(), solely rely on the re_tx_free for the TX descriptor
setup loop. This avoids the fixup on the end of the loop. Add
assertion that the TX descriptor's corresponding mbuf is not setup
yet.
- On RTL8125, the EOF in the TX descriptor is no longer left uncleared
upon TX done, so simply rely on whether the TX descriptor's
corresponding mbuf is set or not to unload the busdma(9) and free the
TX mbuf. We actually should do this instead of rely on the EOF.

Tested-by: me (8101, 8111, 8168, 8125)
Romick <yellowrabbit2010 gmail.com> (8111)
Daniel Fojt <df neosystem.org> (8111)

show more ...

re: MACFG50 requires defragmentation in software, but no padding for runt.

Reported-by: YONETANI Tomokazu
Tested-by: YONETANI Tomokazu

re: Sync with Realtek's re-1.94

Add two more chips' support.

Tested-by: zrj, Daniel Bilik <ddb neosystem org>

re: Leverage Realtek driver's chip/PHY initialization/reset.

It is mainly used to solve all kinds of weirdness, e.g. hanging after
watchdog timeout, and PHY read failure, we have encountered so far

re: Leverage Realtek driver's chip/PHY initialization/reset.

It is mainly used to solve all kinds of weirdness, e.g. hanging after
watchdog timeout, and PHY read failure, we have encountered so far on
these chips.

The DEVMETHODs, ifnet setup, and data path are still ours. miibus(4)
is no longer used, ifmedia(4) is used directly.

NOTE: This change drops the 8139C+ and 8169 w/ external PHY support,
which are deemed to be too old.

Realtek's driver is taken from their website: version 1.92 (2016/10/25).

Tested-by: dillon, zrj

show more ...

re: Reduce GMII PHY read/write timeout

Obtained-from: FreeBSD

re: Update Realtek PHY power management related bits

Obtained-from: FreeBSD

kernel: Use the new auto-created sysctl ctx/tree in various drivers.

Everywhere where we formerly rolled a hw.fooX.* per-device ctx/tree
ourselves.

This commit switches it to dev.foo.X.*

re: Remove the macver, which only serves proper EEPROM address handling

re: Disable IP header checksum offloading on 8168C/8168CP

These chips will generate wrong IP header checksum if IP options exist.

Obtained-from: FreeBSD

re: Use memory mapped IO for PCIe chips

re: Swicth from device_polling to ifpoll

re(4): Use MSI if device supports it.

RT8168E supports MSI.

re(4): Preliminary 8168F support

re(4): Add 8168DP support

re(4): Correct 8168E settings

re(4): Utilize bus_dmamap_load_mbuf_{segment,defrag}()

Increase RE_TXDESC_SPARE to cooperate 9K jumbo frame.

Split RX/TX bus dma tag. RX buffers should to be aligned on 8 bytes.

Correct busdma(9) parameters

Don't set max read request size on 10/100 only PCIe NICs.
Setting max read request size to non-default value seems to cause these
kinds of NICs' DMA engine malfunction.

Reported-by: josepht@
Tested-

Don't set max read request size on 10/100 only PCIe NICs.
Setting max read request size to non-default value seems to cause these
kinds of NICs' DMA engine malfunction.

Reported-by: josepht@
Tested-by: josepht@
Based-on-patch-from: josepht@

show more ...

Remove jumbo frame support for 8169; it never works (except slow ping) and
pullute the TX path a lot.

- Move RX filter configuration from re_init() into re_setmulti()
- IFF_BROADCAST will never be set/cleared
- Fix SIOCSIFFLAGS support; it was really annoying that each time when I ran
tcpdump, the

- Move RX filter configuration from re_init() into re_setmulti()
- IFF_BROADCAST will never be set/cleared
- Fix SIOCSIFFLAGS support; it was really annoying that each time when I ran
tcpdump, the NIC reinitialized itself.

show more ...

- In re_stop(), call re_reset(), which is supposed to stop TX/RX engines.
- In re_reset(), don't touch 0x82 (a magic CSR), which seems to be 8110/8169
specific. Write 1 to it on attach path.
Acc

- In re_stop(), call re_reset(), which is supposed to stop TX/RX engines.
- In re_reset(), don't touch 0x82 (a magic CSR), which seems to be 8110/8169
specific. Write 1 to it on attach path.
According-to: RealTek r8169-6.007.00
- For certain chips (looks like all MAC2 chips), RE_CMD_RESET will not stop
TX/RX engines, a seperate command (RE_CMD_STOPREQ) must be issued before
RE_CMD_RESET.
According-to: RealTek r8168-8.008.00

show more ...

- Pack boolean fields into re_softc.re_flags
- Nuke some unused fields in re_softc

# This file should have been committed along with if_re.c rev1.93

- Don't claim 7422 MTU size is support by various 8111/8169 chips (PCI devices);
6144 MTU size works reliably.
Set MTU above 6144 (6 * 1024) on these chips and do following test:
netperf -H

- Don't claim 7422 MTU size is support by various 8111/8169 chips (PCI devices);
6144 MTU size works reliably.
Set MTU above 6144 (6 * 1024) on these chips and do following test:
netperf -H host -l 30 -t UDP_STREAM -- -m (mtu-28)
All kinds of wiredness will pop up on the test box.
- Set max supported MTU to 9216 for 8168D.
Obtained-from: Realtek r8168-8.008.00
- Set max supported MTU to 6144 for non-8168D GigE chips.
- Cleanup jumbo frame/MTU size related macros.

# As usual, 8169(with 88E1000 PHY) does not seem to work well with any jumbo
# frame size

show more ...