Board can define its own AXI burst length to improve DWMAC DMA performance.
v2-changes:
- Avoid write burst len register when the Macro is not defined.
v3-changes:
- Add axi_bus register member to struct eth_dma_regs.
Signed-off-by: Sonic Zhang <sonic.zhang@analog.com>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
If for some reason DMA module fails to reset user oserves only this:
--->---
# dhcp
Trying dwmac.e0018000
FAIL
--->---
This message makes not much sense.
With proposed change error message will be more helpful:
--->---
# dhcp
Trying dwmac.e0018000
DMA reset timeout
FAIL
--->---
For example user may do power toggle to recover board functionality.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Chin Liang See <clsee@altera.com>
Cc: Dinh Nguyen <dinguyen@altera.com>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@ti.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Pavel Machek <pavel@denx.de>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Cc: Ian Campbell <ijc@hellion.org.uk>
Cc: Marek Vasut <marex@denx.de>
Reviewed-by: Tom Rini <trini@ti.com>
Acked-by: Pavel Machek <pavel@denx.de>
Acked-by: Joe Hershberger <joe.hershberger@ni.com>
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Add a few new variables to make the cache handling less cryptic.
Add a variable for DMA and DATA descriptor start and end, so the
correctness of the code is easier to inspect.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Chin Liang See <clsee@altera.com>
Cc: Dinh Nguyen <dinguyen@altera.com>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@ti.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Pavel Machek <pavel@denx.de>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Acked-by: Pavel Machek <pavel@denx.de>
Acked-by: Chin Liang See <clsee@altera.com>
Fix remaining cache alignment issues in the DWC Ethernet driver.
Please note that the cache handling in the driver is making the
code hideous and thus the next patch cleans that up. In order to
make this change reviewable though, the cleanup is split from it.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Chin Liang See <clsee@altera.com>
Cc: Dinh Nguyen <dinguyen@altera.com>
Cc: Albert Aribaud <albert.u.boot@aribaud.net>
Cc: Tom Rini <trini@ti.com>
Cc: Wolfgang Denk <wd@denx.de>
Cc: Pavel Machek <pavel@denx.de>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Acked-by: Pavel Machek <pavel@denx.de>
The correct value for this setting can vary across SoCs and boards, so make it
configurable.
Also reduce the default value to 8, which is the same default as used in the
Linux driver.
Signed-off-by: Ian Campbell <ijc@hellion.org.uk>
Cc: Alexey Brodkin <abrodkin@synopsys.com>
This is required at least on ARM.
When sending instead of simply invalidating the entire descriptor, flush
as little as possible while still respecting ARCH_DMA_MINALIGN, as
requested by Alexey.
Signed-off-by: Ian Campbell <ijc@hellion.org.uk>
Cc: Alexey Brodkin <abrodkin@synopsys.com>
struct dw_eth_dev contains fields which are accessed via DMA, so make sure it
is aligned to a dma boundary. Without this I see:
ERROR: v7_dcache_inval_range - start address is not aligned - 0x7fb677e0
Signed-off-by: Ian Campbell <ijc@hellion.org.uk>
Reviewed-by: Alexey Brodkin <abrodkin@synopsys.com>
Acked-by: Marek Vasut <marex@denx.de>
This sets up the linkage from the phydev back to the ethernet device. This
symptom of not doing this which I noticed was:
<NULL> Waiting for PHY auto negotiation to complete....
rather than:
dwmac.1c50000 Waiting for PHY auto negotiation to complete....
Signed-off-by: Ian Campbell <ijc@hellion.org.uk>
Cc: Alexey Brodkin <Alexey.Brodkin@synopsys.com>
Up until now this driver only worked with data cache disabled.
To make it work with enabled data cache following changes were required:
* Flush Tx/Rx buffer descriptors their modification
* Invalidate Tx/Rx buffer descriptors before reading its values
* Flush cache for data passed from CPU to GMAC
* Invalidate cache for data passed from GMAC to CPU
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Vipin Kumar <vipin.kumar@st.com>
Cc: Stefan Roese <sr@denx.de>
Cc: Mischa Jonker <mjonker@synopsys.com>
Cc: Shiraz Hashim <shiraz.hashim@st.com>
Cc: Albert ARIBAUD <albert.u.boot@aribaud.net>
Cc: Amit Virdi <amit.virdi@st.com>
Cc: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
With this change driver will benefit from existing phylib and thus
custom phy functionality implemented in the driver will go away:
* Instantiation of the driver is now much shorter - 2 parameters
instead of 4.
* Simplified phy management/functoinality in driver is replaced with
rich functionality of phylib.
* Support of custom phy initialization is now done with existing
"board_phy_config".
Note that after this change some previously used config options
(driver-specific PHY configuration) will be obsolete and they are simply
substituted with similar options of phylib.
For example:
* CONFIG_DW_AUTONEG - no need in this one. Autonegotiation is enabled
by default.
* CONFIG_DW_SEARCH_PHY - if one wants to specify attached phy
explicitly CONFIG_PHY_ADDR board config option has to be used, otherwise
automatically the first discovered on MDIO bus phy will be used
I believe there's no need now in "doc/README.designware_eth" because
user only needs to instantiate the driver with "designware_initialize"
whose prototype exists in "include/netdev.h".
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Vipin Kumar <vipin.kumar@st.com>
Cc: Stefan Roese <sr@denx.de>
Cc: Mischa Jonker <mjonker@synopsys.com>
Cc: Shiraz Hashim <shiraz.hashim@st.com>
Cc: Albert ARIBAUD <albert.u.boot@aribaud.net>
Cc: Amit Virdi <amit.virdi@st.com>
Cc: Sonic Zhang <sonic.zhang@analog.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Driver "init" function might be called multiple times.
On every "init" Tx/Rx buffer descriptors are initialized: "descs_init"
-> "{tx|rx}_descs_init".
In its turn those init functions set MAC's "{tx|rx}desclistaddr" to
point on the first buffer descriptor in the list.
So CPU to start operation from the first buffer descriptor as well after
every "init" we have to reset "{tx|rx}_currdescnum".
Cc: Joe Hershberger <joe.hershberger@ni.com>
Cc: Vipin Kumar <vipin.kumar@st.com>
Cc: Stefan Roese <sr@denx.de>
Cc: Mischa Jonker <mjonker@synopsys.com>
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
"bus mode" register contains lots of fields and some of them don't
expect to be written with 0 (zero). So since we're only interested in
resetting MAC (which is done with setting the least significant bit of
this register with "0") I believe it's better to modify only 1 bit of
the register.
Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Acked-by: Vipin Kumar <vipin.kumar@st.com>
Patch: 277864
This commit is an add-on to f6c4191f. There are a few registers where
consecutive writes to the same location should be avoided or have a delay.
According to Synopsys, here is a list of the registers and bit(s) where
consecutive writes should be avoided or a delay is required:
DMA Registers:
Register 0 Bit 7
Register 6 All bits except for 24, 16-13, 2-1.
GMAC Registers:
Registers 0-3 All bits
Registers 6-7 All bits
Register 10 All bits
Register 11 All bits except for 5-6.
Registers 16-47 All bits
Register 48 All bits except for 18-16, 14.
Register 448 Bit 4.
Register 459 Bits 0-3.
Signed-off-by: Dinh Nguyen <dinguyen@altera.com>
Reviewed-by: Matthew Gerlach <mgerlach@altera.com>
Acked-by: Amit Virdi <amit.virdi@st.com>
This patch adds the following changes to designware ethernet driver
found on the ST SPEAr SoC:
- Don't init MAC & PHY upon startup. This causes a delay, waiting for
the auto negotiation to complete. And we don't want this delay to
always happen. Especially not on platforms where ethernet is not
used at all (e.g. booting via flash).
Instead postpone the MAC / PHY configuration to the stage, where
ethernet is first used.
- Add possibility for board specific PHY init code. This is needed
for example on the X600 board, where the Vitesse PHY needs to be
configured for GMII mode.
This board specific PHY init is done via the function
designware_board_phy_init(). And this driver now adds a weak default
which can be overridden by board code.
- Use common functions miiphy_speed() & miiphy_duplex() to read
link status from PHY.
- Print status and progress of auto negotiation.
- Print link status (speed, dupex) upon first usage.
Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Amit Virdi <amit.virdi@st.com>
Cc: Vipin Kumar <vipin.kumar@st.com>
Cc: Joe Hershberger <joe.hershberger@gmail.com>
Acked-by: Joe Hershberger <joe.hershberger@gmail.com>
Few Designware peripheral registers need to be modified based on the
ethernet interface selected by the board. This patch supports interface
information in ethernet driver
Signed-off-by: Vipin Kumar <vipin.kumar@st.com>
Signed-off-by: Amit Virdi <amit.virdi@st.com>
Signed-off-by: Stefan Roese <sr@denx.de>
These are all the files which use the API incorrectly but did not get
built using MAKEALL -a powerpc|arm. I have no compiler for them, but
the remaining issues should be far less than without this patch.
Any outstanding issues are left to the maintainers of boards that use
these drivers.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
The new implementation changes the timeout loop implementation to avoid 1 ms
delay in each failing test. It also configures the delay to 10usec.
Signed-off-by: Amit Virdi <amit.virdi@st.com>
This patch forces the advertised capabilities during auto
negotiation to always be 10/100 Mbps and half/full as duplexing.
Signed-off-by: Armando Visconti <armando.visconti@st.com>
Signed-off-by: Amit Virdi <amit.virdi@st.com>
If AN(auto-negotiation) is ON, speed bit of control register are not
applicable. Also phy registers were not getting programmed as per the
result of AN. This patch sets only AN bit & restart AN bit for AN ON
selection & programs PHY registers as per AN result.
Signed-off-by: Vikas Manocha <vikas.manocha@st.com>
Signed-off-by: Amit Virdi <amit.virdi@st.com>
Phy autonegotiation works only when the ethernet cable is plugged in.
Since the phy was configured only at the init time, a plugged in cable
was necessary to initialize the phy properly.
This patch keeps a flag to check if the phy initialization has
succeeded, and calls configure_phy routine at every init if this flag
reports otherwise.
Signed-off-by: Vipin Kumar <vipin.kumar@st.com>
Signed-off-by: Amit Virdi <amit.virdi@st.com>
This patch solves a TX/RX problem which happens at 10Mbps, due to the
fact that we are not respecting 4 cyles of the phy_clk (2.5MHz) between
two consecutive writes on the same register.
Signed-off-by: Armando Visconti <armando.visconti@st.com>
Signed-off-by: Amit Virdi <amit.virdi@st.com>
The code assumes the phy address to be > 0, which is not true, the phy address
can be in the range 0-31.
Signed-off-by: Vipin Kumar <vipin.kumar@st.com>
Signed-off-by: Amit Virdi <amit.virdi@st.com>
The network controller mac resets hardware address stored in MAC_HI and MAC_LO
registers if mac is resetted. So, hw mac address needs to be restored in case
mac is explicitly resetted from driver.
Signed-off-by: Vipin Kumar <vipin.kumar@st.com>
Signed-off-by: Amit Virdi <amit.virdi@st.com>
Built-in Ethernet adapters support setting the mac address by means of a
ethaddr environment variable for each interface (ethaddr, eth1addr, eth2addr).
This adds similar support to the USB network side, using the names
usbethaddr, usbeth1addr, etc. They are kept separate since we don't want
a USB device taking the MAC address of a built-in device or vice versa.
Signed-off-by: Simon Glass <sjg@chromium.org>
Tested-by: Eric Bénard <eric@eukrea.com>
When CONFIG_DW_SEARCH_PHY is disabled, the local phy_addr variable
never gets initialized which causes random behavior at runtime and a
gcc warning. So set it by default to the stored phy address.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Reviewed-by: Vipin Kumar <vipin.kumar@st.com>
Fix commit message.
Signed-off-by: Wolfgang Denk <wd@denx.de>
The ctrl variable is only used when autoneg support is disabled, so only
declare it under those conditions to avoid an unused variable warning.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Reviewed-by: Vipin Kumar <vipin.kumar@st.com>
The include/miiphy.h header duplicates a lot of things from linux/mii.h.
So punt all the things that overlap to keep the API simple and to make
merging between U-Boot and Linux simpler.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
The driver name does not need to be writable, so constify it.
Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
Designware network driver support added.
This is a Synopsys ethernet controller
Signed-off-by: Vipin Kumar <vipin.kumar@st.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
