Move the dram helper functions to a separate C file, rather then having them
as inline helpers in dram.h. This saves 144 bytes in the .text segment for
sun6i builds.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
After reboot, reset or even short power off, DRAM typically retains
the old stale data for some period of time (for this type of memory,
the bits of data are stored in slowly discharging capacitors).
The current sun6i/sun8i DRAM size detection logic, which is
inherited from the Allwinner code, relies on using a large magic
signature with the hope that it is unique enough and unlikely to
ever accidentally match this leftover garbage data in RAM. But
this approach is inherently unsafe, as can be demonstrated using
the following test program:
/***** A testcase for reproducing the problem ******/
void main(int argc, char *argv[])
{
size_t size, i;
uint32_t *buf;
/* Allocate the buffer */
if (argc < 2 || !(size = (size_t)atoi(argv[1]) * 1048576) ||
!(buf = malloc(size))) {
printf("Need buffer size in MiB as a cmdline argument\n");
exit(1);
}
/* Fill it with the Allwinner DRAM "magic" values */
for (i = 0; i < size / 4; i++)
buf[i] = 0xaa55aa55 + ((uintptr_t)&buf[i] / 4) % 64;
/* Try to reboot */
system("reboot");
/* And wait */
for (;;) {}
}
/***************************************************/
If this test program is run on the device (giving it a large
chunk of memory), then the DRAM size detection logic in u-boot
gets confused after reboot and fails to initialize DRAM properly.
A better approach is not to rely on luck and abstain from making
any assumptions about the properties of the leftover garbage
data in RAM. Instead just use a more reliable code for testing
whether two different addresses refer to the same memory location.
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Based on the register / dram_para headers from the Allwinner u-boot / linux
sources + the init sequences from boot0.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
The sun8i boot0 code fills the DRAM with a "random" pattern before comparing
it at different offsets to do columns, etc. detection. The sun6i boot0 code
does not do it, instead relying on the memory contents being random enough
to begin with for the memcmp to properly detect the wrap-around address, iow
it is working purely by chance. Since our sun6i dram code was modelled after
the boot0 code it contained the same issue.
This commit fixes this by filling the memory with a unique, distinct pattern.
The new mctl_mem_fill function this introduces is added as an inline helper
in dram.h, so that it can be shared with the sun8i dram code.
While at it move mctl_mem_matches to dram.h for re-use in sun8i too.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
The await_completion helper is already copy pasted between the sun4i and sun6i
dram code, and we need it for sun8i too, so lets make it an inline helper in
dram.h, rather then adding yet another copy.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Add full support for dram initialization, using a fixed clock and autodetection
of the memory organization (numbers of channels, bus-width, etc.).
This is based on dram_sun6i.c and dram.h from u-boot in the Allwinner A31 SDK,
extended with extra initialization sequences and the autodetect algorithm
from boot0.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
The hardware DQS gate training is a bit unreliable and does not
always find the best delay settings.
So we introduce a 32-bit 'dqs_gating_delay' variable, where each
byte encodes the DQS gating delay for each byte lane. The delay
granularity is 1/4 cycle.
Also we allow to enable the active DQS gating window mode, which
works better than the passive mode in practice. The DDR3 spec
says that there is a 0.9 cycles preamble and 0.3 cycle postamble.
The DQS window has to be opened during preamble and closed during
postamble. In the passive window mode, the gating window is opened
and closed by just using the gating delay settings. And because
of the 1/4 cycle delay granularity, accurately hitting the 0.3
cycle long postamble is a bit tough. In the active window mode,
the gating window is auto-closing with the help of monitoring
the DQS line, which relaxes the gating delay accuracy requirements.
But the hardware DQS gate training is still performed in the passive
window mode. It is a more strict test, which is reducing the results
variance compared to the training with active window mode.
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
It is going to be useful in more than one place.
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
The stale error status should be cleared for all sun4i/sun5i/sun7i
hardware and not just for sun7i. Also there are two types of DQS
gate training errors ("found no result" and "found more than one
possible result"). Both are handled now.
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
The sun5i hardware (Allwinner A13) introduced configurable MBUS clock
speed. Allwinner A13 uses only 16-bit data bus width to connect the
external DRAM, which is halved compared to the 32-bit data bus of sun4i
(Allwinner A10), so it does not make much sense to clock a wider
internal bus at a very high speed. The Allwinner A13 manual specifies
300 MHz MBUS clock speed limit and 533 MHz DRAM clock speed limit. Newer
sun7i hardware (Allwinner A20) has a full width 32-bit external memory
interface again, but still keeps the MBUS clock speed configurable.
Clocking MBUS too low inhibits memory performance and one has to find
the optimal MBUS/DRAM clock speed ratio, which may depend on many
factors:
http://linux-sunxi.org/A10_DRAM_Controller_Performance
This patch introduces a new 'mbus_clock' parameter for the 'dram_para'
struct and uses it as a desired MBUS clock speed target. If 'mbus_clock'
is not set, 300 MHz is used by default to match the older hardcoded
settings.
PLL5P and PLL6 are both evaluated as possible clock sources. Preferring
the one, which can provide higher clock frequency that is lower or
equal to the 'mbus_clock' target. In the case of a tie, PLL5P has
higher priority.
Attempting to set the MBUS clock speed has no effect on sun4i, but does
no harm either.
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
The DRAM controller allows to configure impedance either by using the
calibration against an external high precision 240 ohm resistor, or
by skipping the calibration and loading pre-defined data. The DRAM
controller register guide is available here:
http://linux-sunxi.org/A10_DRAM_Controller_Register_Guide#SDR_ZQCR0
The new code supports both of the impedance configuration modes:
- If the higher bits of the 'zq' parameter in the 'dram_para' struct
are zero, then the lowest 8 bits are used as the ZPROG value, where
two divisors encoded in lower and higher 4 bits. One divisor is
used for calibrating the termination impedance, and another is used
for the output impedance.
- If bits 27:8 in the 'zq' parameters are non-zero, then they are
used as the pre-defined ZDATA value instead of performing the ZQ
calibration.
Two lowest bits in the 'odt_en' parameter enable ODT for the DQ and DQS
lines individually. Enabling ODT for both DQ and DQS means that the
'odt_en' parameter needs to be set to 3.
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Acked-by: Ian Campbell <ijc@hellion.org.uk>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
This patch adds DRAM initialisation support for the Allwinner A20 (sun7i)
processor. This code will not been compiled until the build is hooked up in a
later patch. It has been split out to keep the patches manageable.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Signed-off-by: Emilio López <emilio@elopez.com.ar>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Henrik Nordstrom <henrik@henriknordstrom.net>
Signed-off-by: Jens Kuske <jenskuske@gmail.com>
Signed-off-by: Luke Leighton <lkcl@lkcl.net>
Signed-off-by: Oliver Schinagl <oliver@schinagl.nl>
Signed-off-by: Siarhei Siamashka <siarhei.siamashka@gmail.com>
Signed-off-by: Stefan Roese <sr@denx.de>
Signed-off-by: Ian Campbell <ijc@hellion.org.uk>
Reviewed-by: Marek Vasut <marex@denx.de>
Cc: Tom Cubie <Mr.hipboi@gmail.com>
Reviewed-by: Tom Rini <trini@ti.com>
