- fdt_support: Add fdt_for_each_node_by_compatible() helper macro
(Marek)
- turris_omnia: Fixup SATA or PCIe nodes at runtime in DT blob (Pali)
- pci_mvebu: Add support for Kirkwood PCIe controllers (Pali)
- SPL: More verifications for kwbimage in SPL (Pali)
- mvebu: Remove comphy_update_map() (Pali)
- Minor misc stuff
This weak function is not used anymore, so completely remove it.
Private struct comphy_map is not used by any board code anymore, so move it
into private driver header file comphy_core.h.
Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Placing Unicode control codes <U+0080><U+0093> in the middle of a comment
does not make much sense. Let's get rid of all Unicode in
drivers/ram/octeon/octeon3_lmc.c.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Stefan Roese <sr@denx.de>
Add parameter spl_boot_device to spl_parse_board_header(), which allows
the implementations to see from which device we are booting and do
boot-device-specific checks of the image header.
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Kirkwood uses macros KW_DEFADR_PCI_MEM and KW_DEFADR_PCI_IO for base
address of PCIe mappings. Size of PCIe windows is not defined in any macro
yet, so export them in new KW_DEFADR_PCI_MEM_SIZE and KW_DEFADR_PCI_IO_SIZE
macros.
Kirkwood arch code already maps mbus windows for io and mem, so avoid
calling mvebu_mbus_add_window_by_id() function which would try to do
duplicate window mapping.
Kirkwood PCIe controllers already use "marvell,kirkwood-pcie" DT compatible
string, so mark pci_mvebu.c driver as compatible for it.
Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Add macro fdt_for_each_node_by_compatible() to allow iterating over
fdt nodes by compatible string.
Convert various usages of
off = fdt_node_offset_by_compatible(fdt, start, compat);
while (off > 0) {
code();
off = fdt_node_offset_by_compatible(fdt, off, compat);
}
and similar, to
fdt_for_each_node_by_compatible(off, fdt, start, compat)
code();
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
gpio:
- Add modepin driver
net:
- Save random mac addresses to eth variable
zynqmp gem:
- Add support for mdio bus DT description
- Add support for reset and SGMII phy configuration
- Reduce timeout for MDIO accesses
zynqmp clk:
- Fix clock handling for gem and usb
phy:
- Add zynqmp phy/serdes driver
serial:
- Add one missing compatible string
microblaze:
- Symbol alignement
- SPL fixups
- Code cleanups
zynqmp:
- Various dt changes, DP pre-reloc, gem resets, gem clocks
- Switch SOM to shared psu configuration
- Move dcache handling to firmware driver
- Workaround gmii2rgmii DT description issue
- Enable broadcasts again
- Change firmware enablement logic
- Small adjustement in firmware driver
versal:
- Support new mmc@ DT nodes
- Fix run time variable handling
- Add missing I2C_PMC ID for power domain
-----BEGIN PGP SIGNATURE-----
iF0EABECAB0WIQQbPNTMvXmYlBPRwx7KSWXLKUoMIQUCYeg7sQAKCRDKSWXLKUoM
IVhJAKCAiNx/joEeFBJ0XgThtJzFhCjdMwCfYKY9Ewz4L0n2I56lDgR3UJroct0=
=HtB+
-----END PGP SIGNATURE-----
Merge tag 'xilinx-for-v2022.04-rc1' of https://source.denx.de/u-boot/custodians/u-boot-microblaze
Xilinx changes for v2022.04-rc1
gpio:
- Add modepin driver
net:
- Save random mac addresses to eth variable
zynqmp gem:
- Add support for mdio bus DT description
- Add support for reset and SGMII phy configuration
- Reduce timeout for MDIO accesses
zynqmp clk:
- Fix clock handling for gem and usb
phy:
- Add zynqmp phy/serdes driver
serial:
- Add one missing compatible string
microblaze:
- Symbol alignement
- SPL fixups
- Code cleanups
zynqmp:
- Various dt changes, DP pre-reloc, gem resets, gem clocks
- Switch SOM to shared psu configuration
- Move dcache handling to firmware driver
- Workaround gmii2rgmii DT description issue
- Enable broadcasts again
- Change firmware enablement logic
- Small adjustement in firmware driver
versal:
- Support new mmc@ DT nodes
- Fix run time variable handling
- Add missing I2C_PMC ID for power domain
Power domain driver sends PM fragment to PMUFW. It is sent for every node
which is listed in DT. But some nodes could be already enabled but driver
is not capable to find it out. That's why it blinly sents request for every
listed IP. When PMUFW response by XST_PM_ALREADY_CONFIGURED error code
there is no need to show any error message because node is already enabled.
That's why cover this case with message when DEBUG is enabled.
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Link: https://lore.kernel.org/r/8c15ef0b68cf191f693d3d010f70ac24cfd8171f.1642163135.git.michal.simek@xilinx.com
The clock delay of the RMII/RGMII interface is controlled by SCU340~35C.
These values are obtained by measurement and experiments so we simply
use macro to define them.
Signed-off-by: Dylan Hung <dylan_hung@aspeedtech.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Add support of the MAC controller of Aspeed AST2600 SOC. The MAC
controller is the same with AST2500, except it has stand-alone MDIO
hardware block.
Signed-off-by: Dylan Hung <dylan_hung@aspeedtech.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Add support for DM_MDIO to connect to PHY. For the systems that have a
stand-alone MDIO hardware block, enable CONFIG_DM_MDIO to use driver
model for MDIO devices.
Signed-off-by: Dylan Hung <dylan_hung@aspeedtech.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
This commit adds SD/MMC clocks, and provides .set/get_rate callbacks
for SD/MMC device present on Actions OWL S700 SoCs.
Signed-off-by: Amit Singh Tomar <amittomer25@gmail.com>
This commit introduces get/set_rate callbacks, these are dummy at
the moment, and can be used to get/set clock for various devices
based on the clk id.
Signed-off-by: Amit Singh Tomar <amittomer25@gmail.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
Changes to the am33xx device (33e9021a) trees have been merged in from
the upstream linux kernel which now means the device tree uses the new
pins format (as of 5.10) where the confinguration can be stores as a
separate configuration value and pin mux mode which are then OR'd
together.
This patch adds support for the new format to u-boot so that
pinctrl-cells is now respected when reading in pinctrl-single,pins
Signed-off-by: Anthony Bagwell <anthony.bagwell@hivehome.com>
Fix fastboot flash bug.
If the downloading file size is equal to the partition size, "fastboot
flash" can't work, at least in sunxi platform, because used an
uninitalized point: ep->desc.
This patch also fixed 'data abort' bug in am335x platform.
Reproduce: fastboot flash loader1 spl/sunxi-spl.bin.
Signed-off-by: qianfan Zhao <qianfanguijin@163.com>
Reviewed-by: Sean Anderson <seanga2@gmail.com>
Currently the pxa3xx driver does not set the udevice in the mtd_info
struct and this prevents the mtd from parsing the partitions via DTS
like for SPI-NOR.
So simply set the mtd->dev to the driver udevice.
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
Reviewed-by: Stefan Roese <sr@denx.de>
- Let am335x_evm use the CPSW or PRUSS ethernet.
- Implement timer_get_boot_us in the omap timer driver
- gpmc bitflip, QSPI clock calculation on am437x, da8xx_gpio bugfixes
- Assorted K3 updates
-----BEGIN PGP SIGNATURE-----
iQFQBAABCgA6FiEEqxhEmNJ6d7ZdeFLIHrMeAg6sL8gFAmHlNOccHGV1Z2VuLmhy
aXN0ZXZAbWljcm9jaGlwLmNvbQAKCRAesx4CDqwvyL0JB/4+sKNWPMgOYvP2uLCC
euRGomhxJ5WmxJVDtJeBIRg0YKSkwnE8JwIN2qqGu15xJnNX9Ck5qnIMPbiZ/io9
ocW5Z/4C4rfJX/OzD6FtmbMNsvG/umVmZkjRU3ryRXKGyNVpVjLQ2EEECaJfk1WZ
Nd/J6/Jo8nafb1x4OeY+F6+xwZAuj7Mrun9SsDSyrIvmEqSnaQNioj+lStkauL51
NwQgBxZBQmbdCgF7Ooy6rYlXZBR6r/rgWB9n3RP8emDCwfCMul9HOMT432oKNbhW
bGrcMu9AWv/pgzXogxSzYbEdTAnoVEODtU1suCa8LV6KJPEG693gARnRm/KjvPHa
FfsK
=39IY
-----END PGP SIGNATURE-----
Merge tag 'u-boot-at91-2022.04-b' of https://source.denx.de/u-boot/custodians/u-boot-at91
Second set of u-boot-at91 features for the 2022.04 cycle:
This small feature set includes few changes for sama7g5 and sama7g5ek:
turn blue led on at boot, changes required for the Rev4 of the board,
better sync with the Linux DT with regards to the new DT nodes.
For configurations with gmii2rgmii and external phy the DT nodes link
should be gem->gmii2rgmii->phy. But due to limitation in Linux driver
the DT is mentioned as gem->phy and gmii2rgmii->phy as shown in below DT.
ethernet@ff0c0000 {
compatible = "cdns,zynqmp-gem\0cdns,gem";
status = "okay";
interrupt-parent = <0x04>;
interrupts = <0x00 0x3b 0x04 0x00 0x3b 0x04>;
reg = <0x00 0xff0c0000 0x00 0x1000>;
clock-names = "pclk\0hclk\0tx_clk\0rx_clk\0tsu_clk";
#address-cells = <0x01>;
#size-cells = <0x00>;
#stream-id-cells = <0x01>;
iommus = <0x0d 0x875>;
power-domains = <0x0c 0x1e>;
clocks = <0x03 0x1f 0x03 0x69 0x03 0x2e 0x03 0x32 0x03 0x2c>;
phy-handle = <0x0e>;
phy-mode = "gmii";
xlnx,ptp-enet-clock = <0x00>;
local-mac-address = [ff ff ff ff ff ff];
phandle = <0x4d>;
mdio {
#address-cells = <0x01>;
#size-cells = <0x00>;
phandle = <0x4e>;
ethernet-phy@1 {
reg = <0x01>;
rxc-skew-ps = <0x708>;
txc-skew-ps = <0x708>;
phandle = <0x0e>;
};
gmii_to_rgmii_0@8 {
compatible = "xlnx,gmii-to-rgmii-1.0";
phy-handle = <0x0e>;
reg = <0x08>;
phandle = <0x4f>;
};
};
};
Since same DT is used in Linux and U-Boot we need to workaround this
issue by using the gmii2rgmii node which points to phy and we should
ignore the gem pointing to phy directly.
Do this workaround by updating priv->phydev->node value with
priv->phy_of_node only if it is not valid node.
Signed-off-by: Ashok Reddy Soma <ashok.reddy.soma@xilinx.com>
Signed-off-by: Michal Simek <michal.simek@xilinx.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
Link: https://lore.kernel.org/r/641eb13425ffe80e0743f60cf90d0f940577b9e9.1642162085.git.michal.simek@xilinx.com
This driver supports power domains for the power management
controller found on Apple SoCs.
Signed-off-by: Mark Kettenis <kettenis@openbsd.org>
Reviewed-by: Jaehoon Chung <jh80.chung@samsung.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
In the case of an erased (sub)page both the data and ECC are all 0xFF
bytes. This fails the normal ECC verification, as the computed ECC of
all-0xFF is not also 0xFF. The GPMC NAND driver attempted to detect
erased pages by checking that the ECC bytes are all-0xFF, but this had
two problems:
1) bitflips in the data were not corrected, so the data looked not-erased
2) bitflips in the ECC bytes were reported as uncorrectable ECC errors
The equivalent Linux driver [1] correctly handles this by counting the
number of 0-bits in the combination of data and ECC bytes. If the number
of 0-bits is less than the amount of bits correctable by the selected
ECC algorithm, then it is treated as an erased page with correctable
bitflips.
Implement similar, though simplified, logic in omap_correct_data_bch().
[1] see omap_elm_correct_data() in omap2.c
Signed-off-by: David Rivshin <drivshin@allworx.com>
On AM4372 the SPI_GCLK input gets its clock from the PRCM module which
divides the PER_CLKOUTM2 frequency (192MHz) by a fixed factor of 4.
See AM437x Reference Manual in section 27 QSPI >> 27.2 Integration.
The QSPI_FCLK therefore needs to take this factor into account and
becomes (192000000 / 4).
Signed-off-by: Stefan Mätje <stefan.maetje@esd.eu>
To make the OMAP DM timer driver useful for the timing of
bootstages, we need to implement timer_get_boot_us(..).
Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
The GPIO bank numbers do not appear in the device tree,
so make the gpio name based on the address
(ie gpio@42110000_25 vs 25)
Signed-off-by: chao zeng <chao.zeng@siemens.com>
The fastboot protocol uses per default the UDP port 5554. In some cases
it might be needed to change the used port. The fastboot utility provides
a way to specifiy an other port number to use already.
fastboot -s udp:192.168.1.76:1234 boot fastboot.img
Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Heiko Schocher <hs@denx.de>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
NXP's mEMAC reference manual, Chapter 6.5.5 "MDIO Ethernet Management
Interface usage", specifies to poll the BSY (0) bit in the CFG/STAT
register to wait until a transaction has finished, not bit 31 in the
data register.
In the Linux kernel, this has already been fixed in commit 26eee0210ad7
("net/fsl: fix a bug in xgmac_mdio").
This patch changes the register in the fman_mdio and fsl_ls_mdio
drivers.
As the MDIO_DATA_BSY define is no longer in use, this patch also removes
its definition from the fsl_memac header.
Signed-off-by: Markus Koch <markus@notsyncing.net>
Reviewed-by: Camelia Groza <camelia.groza@nxp.com>
Reduce the missing phy-handle log message to debug message. It is
possible for ethernet DT node to have no phy-handle e.g. in case
of a fixed-link connection. Furthermore, drop the FEC: prefix,
which is a copy-paste error and rather print the ethernet device
name.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Ramon Fried <rfried.dev@gmail.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
This driver is based on an older downstream TI kernel, with
changes and cleanups to work with mainline device-tree bindings.
Signed-off-by: Dominic Rath <rath@ibv-augsburg.net>
Signed-off-by: Christian Gmeiner <christian.gmeiner@gmail.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
The sja1105_check_device_id() function contains logic to work without
changing the device tree on reworked boards, one of which I have (the
NXP LS1021A-TSN normally has a SJA1105T, but I have a version with a
resoldered SJA1105Q which is pin compatible). This logic is taken from
the Linux driver.
However this logic gets shortcircuited in U-Boot by an earlier check for
the exact device ID specified in the device tree. So the reworked board
does not probe the SJA1105Q switch. Remove this duplicated logic and let
the automatic device ID detection do its job.
Fixes: f24b666b22 ("net: dsa: add driver for NXP SJA1105 L2 switch")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Ramon Fried <rfried.dev@gmail.com>
On certain TI SoC, like AM64x there is a CPSW3G which supports 2
external independent MAC ports for single CPSW instance.
It is not possible for Ethernet driver to register more than one port
for given instance.
This patch modifies top level CPSW NUSS as UCLASS_MISC and binds
UCLASS_ETH to individual ports so as to support bring up more than one
Ethernet interface in U-Boot.
Note that there is no isolation in the since, CPSW NUSS is in promisc
mode and forwards all packets to host.
Since top level driver is now UCLASS_MISC, board files would need to
instantiate this driver explicitly.
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
Driver has a bug in that it uses rflow_in_use bitmap when setting up free rflow range
from TISCI but use rflow_map for reservation in __udma_reserve_rflow()
Fix this by dropping rflow_in_use bitmap array and use rflow_map for
PKTDMA. BCDMA does not need rflow_in_use either.
This fixes CPSW3g not able to get DMA channels at R5 SPL on AM64x
Signed-off-by: Vignesh Raghavendra <vigneshr@ti.com>
On boards using the RK3399 SoC, the USB OHCI and EHCI controllers share
the same PHY device instance. While these controllers are being stopped
they both attempt to power-off and deinitialize it, but trying to
power-off the deinitialized PHY device results in a hang. This usually
happens just before booting an OS, and can be explicitly triggered by
running "usb start; usb stop" in the U-Boot shell.
Implement a uclass-wide counting mechanism for PHY initialization and
power state change requests, so that we don't power-off/deinitialize a
PHY instance until all of its users want it done. The Allwinner A10 USB
PHY driver does this counting in-driver, remove those parts in favour of
this in-uclass implementation.
The sandbox PHY operations test needs some changes since the uclass will
no longer call into the drivers for actions matching its tracked state
(e.g. powering-off a powered-off PHY). Update that test, and add a new
one which simulates multiple users of a single PHY.
The major complication here is that PHY handles aren't deduplicated per
instance, so the obvious idea of putting the counts in the PHY handles
don't immediately work. It seems possible to bind a child udevice per
PHY instance to the PHY provider and deduplicate the handles in each
child's uclass-private areas, like in the CLK framework. An alternative
approach could be to use those bound child udevices themselves as the
PHY handles. Instead, to avoid the architectural changes those would
require, this patch solves things by dynamically allocating a list of
structs (one per instance) in the provider's uclass-private area.
Signed-off-by: Alper Nebi Yasak <alpernebiyasak@gmail.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Tested-by: Peter Robinson <pbrobinson@gmail.com> - Rock960
Previously, a subpage with an uncorrectable error followed by a subpage
with a correctable error would return an erroneous correctable status.
Signed-off-by: Joel Peshkin <joel.peshkin@broadcom.com>
Cc: Simon Glass <sjg@chromium.org>
Reviewed-by: Simon Glass <sjg@chromium.org>
Calling 'nvme scan' followed by 'nvme detail' crashes U-Boot on Turris
Omnia with the following error:
undefined instruction
pc : [<0a000000>] lr : [<7ff80bfc>]
reloc pc : [<8a8c0000>] lr : [<00840bfc>]
sp : 7fb2b908 ip : 0000002a fp : 02000000
r10: 04000000 r9 : 7fb2fed0 r8 : e1000000
r7 : 0c000000 r6 : 03000000 r5 : 06000000 r4 : 01000000
r3 : 7fb30928 r2 : 7fb30928 r1 : 00000000 r0 : 00000000
Flags: nZCv IRQs off FIQs off Mode SVC_32
Code: 0f0fb4f0 0f0fb4f0 0f0fb4f0 0f0fb4f0 (f0f04b0f)
Resetting CPU ...
This happens when nvme_print_info() tries to return to the caller. It
looks like this error is caused by trying to allocate 8 KiB of memory
on the stack by the two uses of ALLOC_CACHE_ALIGN_BUFFER().
Use malloc_cache_aligned() to allocate this memory dynamically instead.
This fixes 'nvme detail' on Turris Omnia.
Note that similar change was applied to file drivers/nvme/nvme.c in past by
commit 2f83481dff ("nvme: use page-aligned buffer for identify command").
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Attempt to handle cases with a downstream port of a PCIe switch where
link training never completes and the link continues switching between
speeds indefinitely with the data link layer never reaching the active
state.
It has been observed with a downstream port of the ASMedia ASM2824 Gen 3
switch wired to the upstream port of the Pericom PI7C9X2G304 Gen 2
switch, using a Delock Riser Card PCI Express x1 > 2 x PCIe x1 device,
P/N 41433, wired to a SiFive HiFive Unmatched board. In this setup the
switches are supposed to negotiate the link speed of preferably 5.0GT/s,
falling back to 2.5GT/s.
However the link continues oscillating between the two speeds, at the
rate of 34-35 times per second, with link training reported repeatedly
active ~84% of the time, e.g.:
02:03.0 PCI bridge [0604]: ASMedia Technology Inc. ASM2824 PCIe Gen3 Packet Switch [1b21:2824] (rev 01) (prog-if 00 [Normal decode])
[...]
Bus: primary=02, secondary=05, subordinate=05, sec-latency=0
[...]
Capabilities: [80] Express (v2) Downstream Port (Slot+), MSI 00
[...]
LnkSta: Speed 5GT/s (downgraded), Width x1 (ok)
TrErr- Train+ SlotClk+ DLActive- BWMgmt+ ABWMgmt-
[...]
LnkCtl2: Target Link Speed: 8GT/s, EnterCompliance- SpeedDis+, Selectable De-emphasis: -3.5dB
Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-
Compliance De-emphasis: -6dB
[...]
Forcibly limiting the target link speed to 2.5GT/s with the upstream
ASM2824 device makes the two switches communicate correctly however:
02:03.0 PCI bridge [0604]: ASMedia Technology Inc. ASM2824 PCIe Gen3 Packet Switch [1b21:2824] (rev 01) (prog-if 00 [Normal decode])
[...]
Bus: primary=02, secondary=05, subordinate=09, sec-latency=0
[...]
Capabilities: [80] Express (v2) Downstream Port (Slot+), MSI 00
[...]
LnkSta: Speed 2.5GT/s (downgraded), Width x1 (ok)
TrErr- Train- SlotClk+ DLActive+ BWMgmt- ABWMgmt-
[...]
LnkCtl2: Target Link Speed: 2.5GT/s, EnterCompliance- SpeedDis+, Selectable De-emphasis: -3.5dB
Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-
Compliance De-emphasis: -6dB
[...]
and then:
05:00.0 PCI bridge [0604]: Pericom Semiconductor PI7C9X2G304 EL/SL PCIe2 3-Port/4-Lane Packet Switch [12d8:2304] (rev 05) (prog-if 00 [Normal decode])
[...]
Bus: primary=05, secondary=06, subordinate=09, sec-latency=0
[...]
Capabilities: [c0] Express (v2) Upstream Port, MSI 00
[...]
LnkSta: Speed 2.5GT/s (downgraded), Width x1 (downgraded)
TrErr- Train- SlotClk+ DLActive- BWMgmt- ABWMgmt-
[...]
LnkCtl2: Target Link Speed: 5GT/s, EnterCompliance- SpeedDis-
Transmit Margin: Normal Operating Range, EnterModifiedCompliance- ComplianceSOS-
Compliance De-emphasis: -6dB
[...]
Make use of this observation then and attempt to detect the inability to
negotiate the link speed automatically, and then handle it by hand. Use
the Data Link Layer Link Active status flag as the primary indicator of
successful link speed negotiation, but given that the flag is optional
by hardware to implement (the ASM2824 does have it though), resort to
checking for the mandatory Link Bandwidth Management Status flag showing
that the link speed or width has been changed in an attempt to correct
unreliable link operation (the ASM2824 does set it too).
If these checks indicate that link may not operate correctly, then poll
the Data Link Layer Link Active status flag along with the Link Training
flag for the duration of 200ms to see if the link has stabilised, that
is either that the Data Link Layer Link Active status flag has been set
or that Link Training has been inactive during at least the second half
of the interval.
If that has indicated failure, restrict the target speed to 2.5GT/s,
request a link retrain and check again if the link has stabilised. If
that does not work either, then restore the original speed setting and
claim defeat, otherwise we are done.
NB interestingly enough with the ASM2824 vs PI7C9X2G304 configuration
referred above asking the ASM2824 to retrain with a higher target link
speed once the 2.5GT/s speed has been negotiated makes the two devices
successfully negotiate 5.0GT/s. Lifting the 2.5GT/s speed restriction
would however prevent our workaround from working with an OS that issues
a reset and that is unaware of the problem. This is because the devices
would then try to negotiate a higher link speed from scratch and fail,
while the sticky property of the Target Link Speed setting will keep the
2.5GT/s speed restriction across a reset.
Keep the 2.5GT/s speed restriction then, conservatively, if functional
once applied.
Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk>
Reviewed-by: Stefan Roese <sr@denx.de>
This is a cleaned up and fixed version of a patch
mv_ddr: a380: fix SPLIT_OUT_MIX state decision
in each pattern cycle the bus state can be changed
in order to avoide it, need to back to the same bus state on each
pattern cycle
by
Moti Boskula <motib@marvell.com>
The original patch is not in Marvell's mv-ddr-marvell repository. It was
gives to us by Marvell to fix an issues with DDR training on some
boards, but it cannot be applied as is to mv-ddr-marvell, because it is
a very dirty draft patch that would certainly break other things, mainly
DDR4 training code in mv-ddr-marvell, since it changes common functions.
I have cleaned up the patch and removed stuff that seemed unnecessary
(when removed, it still fixed things). Note that I don't understand
completely what the code does exactly, since I haven't studied the DDR
training code extensively (and I suspect that no one besides some few
people in Marvell understand the code completely).
Anyway after the cleanup the patch still fixes isssues with DDR training
on the failing boards.
There was also a problem with the original patch on some of the Allied
Telesis' x530 boards, reported by Chris Packham. I have asked Chris to
send me some logs, and managed to fix it:
- if you look at the change, you'll notice that it introduces
subtraction of cur_start_win[] and cur_end_win[] members, depending on
a bit set in the current_byte_status variable
- the original patch subtracted cur_start_win[] if either
BYTE_SPLIT_OUT_MIX or BYTE_HOMOGENEOUS_SPLIT_OUT bits were set, but
subtracted cur_end_win[] only if the first one (BYTE_SPLIT_OUT_MIX)
was set
- from Chris Packham logs I discovered that the x530 board where the
original patch introduced DDR training failure, only the
BYTE_HOMOGENEOUS_SPLIT_OUT bit was set, and on our boards where the
patch is needed only the BYTE_SPLIT_OUT_MIX is set in the
current_byte_status variable
- this led me to the hypothesis that both cur_start_win[] and
cur_end_win[] should be subtracted only if BYTE_SPLIT_OUT_MIX bit is
set, the BYTE_HOMOGENEOUS_SPLIT_OUT bit shouldn't be considered at all
- this hypothesis also gains credibility when considering the commit
title ("fix SPLIT_OUT_MIX state decision")
Hopefully this will fix things without breaking anything else.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Reviewed-by: Stefan Roese <sr@denx.de>
Tested-by: Chris Packham <judge.packham@gmail.com>
Before commit 4c289425752f ("mv_ddr: a38x: add support for ddr async
mode"), Asynchornous Mode was only used when the CPU Subsystem Clock
Options[4:0] field in the SAR1 register was set to value 0x13: CPU at
2 GHz and DDR at 933 MHz.
Then commit 4c289425752f ("mv_ddr: a38x: add support for ddr async
mode") added support for Asynchornous Modes with frequencies other than
933 MHz (but at least 467 MHz), but the code it added to check for
whether Asynchornous Mode should be used is wrong: it checks whether the
frequency setting in board DDR topology map is set to value other than
MV_DDR_FREQ_SAR.
Thus boards which define a specific value, greater than 400 MHz, for DDR
frequency in their board topology (e.g. Turris Omnia defines
MV_DDR_FREQ_800), are incorrectly put into Asynchornous Mode after that
commit.
The A38x Functional Specification, section 10.12 DRAM Clocking, says:
In Synchornous mode, the DRAM and CPU clocks are edge aligned and run
in 1:2 or 1:3 CPU to DRAM frequency ratios.
Change the check for whether Asynchornous Mode should be used according
to this explanation in Functional Specification.
Signed-off-by: Marek Behún <marek.behun@nic.cz>
Tested-by: Chris Packham <judge.packham@gmail.com>
Reviewed-by: Stefan Roese <sr@denx.de>
The DT properties for the "enable-arbiter" and "keep-config" config
knobs were previously named inconsistently:
- The u-boot driver used "nand-enable-arbiter" and "nand-keep-config"
names, without Marvell prefixes.
- The Linux driver uses "marvell,nand-keep-config" ("enable-arbiter"
does not exist anymore in recent kernels, but it also used to be
"marvell,nand-enable-arbiter").
- The device trees almost all use "marvell," prefixed names, except for
one single instance of "nand-enable-arbiter" without vendor prefix.
This commit standardizes on the vendor prefixed version, making the
u-boot driver read from DT props "marvell,nand-enable-arbiter" and
"marvell,nand-keep-config". The one device tree using the unprefixed
version is also changed to use the new naming.
This has the side effect of making the previously no-op "marvell,"
config knobs already present in some DTs actually do something. This was
likely the original intention of the DT authors, but note that this
commit was not tested on every single impacted board.
Signed-off-by: Pierre Bourdon <delroth@gmail.com>
As explained in commit 3bedbcc3aa ("arm: mvebu: a38x: serdes: Don't
overwrite read-only SAR PCIe registers") it is required to set Maximum Link
Width bits of PCIe Root Port Link Capabilities Register depending of number
of used serdes lanes. As this register is part of PCIe address space and
not serdes address space, move it into pci_mvebu.c driver.
Read number of PCIe lanes from DT property "num-lanes" which is used also
by other PCIe controller drivers in Linux kernel. If this property is
absent then it defaults to 1. This property needs to be set to 4 for every
mvebu board which use PEX_ROOT_COMPLEX_X4 or PEX_BUS_MODE_X4.
Enabling of PCIe port needs to be done afer all registers in PCIe address
space are properly configure. For this purpose use new mvebu-reset driver
(part of system-controller) and remove this code from serdes code.
Because some PCIe ports cannot be enabled individually, it is required to
first setup all PCIe ports and then enable them.
This change contains also all required "num-lanes" and "resets" DTS
properties, to make pci_mvebu.c driver work correctly.
Signed-off-by: Pali Rohár <pali@kernel.org>
After function mvebu_pcie_probe() returns U-Boot DM expects that PCIe link
is already up. In followup patches link initialization will be moved from
SPL to proper and therefore explicitly link up delay is required.
Delay mvebu_pcie_probe() for 100ms to ensure that PCIe link is up after
function finish. In the case when no card is connected to the PCIe slot,
this will delay probe time by 100ms, which should not be problematic.
This change fixes detection and initialization of some QCA98xx cards on
the first serdes when configured in x1 mode. Default configuration of
the first serdes on A385 is x4 mode, so it looks as if some delay is
required when x4 is changed to x1 and card correctly links with A385.
Other PCIe serdes ports on A385 are x1-only, and so they don't have this
problem.
Signed-off-by: Pali Rohár <pali@kernel.org>
In first phase just parse DT properties and fill struct mvebu_pcie. In
second phase setup all PCIe links (without enabling them). And in the last
third phase enable all PCIe links and create UCLASS_PCI device for each
one.
Because parsing of DT is done before UCLASS_PCI is created, we cannot use
DM for this action anymore. So remove .of_to_plat callback and replace it
by ad-hoc function for parsing DT properties and filling struct mvebu_pcie.
Signed-off-by: Pali Rohár <pali@kernel.org>
SoC specific macro SOC_REGS_PHY_BASE is used for two things:
* calculation of base PCIe port address
* filling PCIe register with address of internal registers
For calculating base PCIe port address use function
ofnode_translate_address() which translates DT "assigned-addresses" to
final PCIe port address.
And for calculating address of internal registers use untranslated and
translated DT "assigned-addresses".
Basically this change reads SOC_REGS_PHY_BASE address indirectly from DT.
Signed-off-by: Pali Rohár <pali@kernel.org>
Function mvebu_pcie_port_parse_dt() is called only from
mvebu_pcie_of_to_plat() function. Both these function parse DT properties
required to setup mvebu pcie. So inline mvebu_pcie_port_parse_dt() function
into mvebu_pcie_of_to_plat() to have all code related to parsing DT
properties at one place.
Signed-off-by: Pali Rohár <pali@kernel.org>
