The CPU can arbitrarily reorder calls to rdtsc, significantly
reducing the precision of timing using the CPUs time stamp counter.
Unfortunately the method of synchronizing rdtsc is different
on AMD and Intel CPUs. There is a generic method, using the cpuid
instruction, but that uses up a lot of registers, and is very slow.
Hence, use the correct lfence/mfence instructions (for CPUs that
we know support it)
Change-Id: I17ecb48d283f38f23148c13159aceda704c64ea5
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/1422
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
Adding ranges directly into coreboot memory table raised issues
as those methods bypassed the MTRR setup. Such regions are now
added as resources, so declare the functions again as static.
Change-Id: If78613da40eabc5c99c49dbe2d6047cb22a71b69
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1415
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
The includes removed here were previously required for
struct lb_memory and lb_add_memory_range().
Change-Id: Ie6c0d4ef55c2225aa709cf3fbad30ff1080e3610
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1391
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
Dropping mainboard's chip.h broke execution of the mainboard's enable
function and the addition of mainboard specific smbios tables.
The former was fixed by Kyosti in http://review.coreboot.org/1374
This patch fixes the breakage in static.c and also backs out a small
portion of Kyosti's patch (because it's not needed anymore)
Change-Id: I6fdea9cbb8c6041663bd36f68f1cae4b435c1f9b
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/1421
Tested-by: build bot (Jenkins)
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
KDE editors love to create tons of backup files every time a file is
edited. This makes it very hard to get useful information from git status
and very easy to commit the wrong stuff. Add those to the gitignore list.
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Change-Id: I2fbb907f21d85d6994caa8bbe32c4e9814b5f4b4
Reviewed-on: http://review.coreboot.org/1423
Reviewed-by: Peter Stuge <peter@stuge.se>
Tested-by: build bot (Jenkins)
These existed to provide a hook to add reserved memory regions
in the coreboot memory table. Reserved memory are now
added as resources.
Change-Id: I9f83df33845cfa6973b018a51cf9444dbf0f8667
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1414
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Peter Stuge <peter@stuge.se>
Use of lb_add_memory_region() is reduntant with the MMCONF
resource being set as reserved.
Change-Id: I747ea34823692b6966b2e50d22aea1fb89c73c25
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1394
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The function is a noop for all but amd/serengeti_cheetah.
Change-Id: I09e2e710aa964c2f31e35fcea4f14856cc1e1dca
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1184
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
The names were set at various times during development, but
the way the code works, you might end up with the wrong name
being displayed in the logs. Instead of doing magic, just
display both names for each component
Change-Id: I1f8ce44d156442f5f7d717e1a2b47ed1218d4527
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/1413
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Move beep commands to board-specific area as they need to be different for
different codecs.
Change-Id: I2a1ac938c49827cc816a95df10793a7e234942bf
Signed-off-by: Dylan Reid <dgreid@chromium.org>
Reviewed-on: http://review.coreboot.org/1410
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Use IORESOURCE_RESERVE to exclude the region from system RAM table.
Change-Id: I61b51022165e1304a41554f67af75b3089d892af
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1393
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
Apply the change
http://review.coreboot.org/1390
to all the AMD southbridge.
Change-Id: I8e94014f8883a0408b68355d9aa33aea4373881f
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: zbao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/1406
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Move the POST display to take place just before jumping
the payload, a bit later than before.
Change-Id: Ie1d1ff24dc6c1640e25681be7dc5740943c7f112
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1396
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Some mainboards have really nice capabilities for posting, beyond
simple POST cards. Further, some can not use a POST card. This
change defines a weak symbol (mainboard_post) that can be overridden
by a real mainboard_post function.
If, for example, you'd like to do something fancy before the payload starts,
you can add this to mainboard.c:
void mainboard_post(u8 value)
{
switch(value){
case POST_TIME_TO_PARTY: some_fancy_lights();
break;
}
}
Maybe the post function should be an entry in the device. We're beginning to over-use
weak symbols.
BUG=None
TEST=Build and boot a google chromebook. Observe that it still works. Use it to drive
some pretty lights.
Change-Id: I3512d2ec34a66c747287191851c3f68b6a7cc1b2
Signed-off-by: Ronald G. Minnich <rminnich@gmail.com>
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-on: http://review.coreboot.org/1397
Tested-by: build bot (Jenkins)
In accordance to PCH EDS 14.1.35.1
Change-Id: I2e6cec6d4f49f404e33a171a8fbd6e4880327896
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/1411
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
We thought about two ways to do this change. The way we decided to try
was to
1. drop all ops from devices in romstage
2. constify all devices in romstage (make them read-only) so we can
compile static.c into romstage
3. the device tree "devices" can be used to read configuration from
the device tree (and nothing else, really)
4. the device tree devices are accessed through struct device * in
romstage only. device_t stays the typedef to int in romstage
5. Use the same static.c file in ramstage and romstage
We declare structs as follows:
ROMSTAGE_CONST struct bus dev_root_links[];
ROMSTAGE_CONST is const in romstage and empty in ramstage; This
forces all of the device tree into the text area.
So a struct looks like this:
static ROMSTAGE_CONST struct device _dev21 = {
#ifndef __PRE_RAM__
.ops = 0,
#endif
.bus = &_dev7_links[0],
.path = {.type=DEVICE_PATH_PCI,{.pci={ .devfn = PCI_DEVFN(0x1c,3)}}},
.enabled = 0,
.on_mainboard = 1,
.subsystem_vendor = 0x1ae0,
.subsystem_device = 0xc000,
.link_list = NULL,
.sibling = &_dev22,
#ifndef __PRE_RAM__
.chip_ops = &southbridge_intel_bd82x6x_ops,
#endif
.chip_info = &southbridge_intel_bd82x6x_info_10,
.next=&_dev22
};
Change-Id: I722454d8d3c40baf7df989f5a6891f6ba7db5727
Signed-off-by: Ronald G. Minnich <rminnich@chromium.org>
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Reviewed-on: http://review.coreboot.org/1398
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Apply the change
http://review.coreboot.org/1265
to all the AMD northbridge.
Change-Id: Idf3994c1e9ec76cd19db9f740d825cf24059884f
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: zbao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/1404
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Apply the change
http://review.coreboot.org/1264
to all the AMD northbridge.
Change-Id: Ied74d6f579d2c0350288e2619d7810f8d44fa574
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: zbao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/1403
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
The board had HAVE_MAINBOARD_RESOURCES=0 so this was never
called. Drop unnecessary includes too.
Change-Id: Ia7bddf29a16966c052b5cabbb47029299e6dbd12
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1392
Tested-by: build bot (Jenkins)
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
Detection for a hyper-threading CPU was not compatible with multicore
CPUs. When using CPUID eax==4, also need to set ecx=0.
CAR init tested on real hardware with hyper-threading model_f25 and
under qemu 0.15.1 with multicore CPU.
Change-Id: I28ac8790f94652e4ba8ff88fe7812c812f967608
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1172
Tested-by: build bot (Jenkins)
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
Commit 188e3c2ff0 dropped mainboard
out of the static device tree. This left dev_root->chip_ops unset,
and mainboard_ops.enable_dev() was no longer called.
Change-Id: I6d447c8049a66041b8bb36ec9aac3e7e0d20a99b
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1374
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
If the CMOS is cleared or someone writes some random date/time
on purpose, the CMOS date register has a invalid date. This will
hurts some OS, like Windows 7, which hangs at MS logo forever.
When we detect that, we need to write a reasonable date in CMOS.
Alexandru Gagniuc:
Hmm, it would be interesting to use the date the coreboot image
was built and set that as the default date. At least until time
travel is invented.
Change-Id: Ic1c7a2d60e711265686441c77bdf7891a7efb42e
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: zbao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/1389
Tested-by: build bot (Jenkins)
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
The field device in PCI_ADDRESS only takes 5 bits. So if the device number is
more than 32, it will truncated to 5 bits. Before this patch, other pci devices
will be incorrectly probed as processor node.
Change-Id: I64dcd4f4fda7b7080a9905dce580feb829584b94
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: zbao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/1264
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
Parmer and thather hang at windows 7 booting process. Setting the
valid date in CMOS can fix that.
Change-Id: I5e427cfb42430ebebdb4c1e48bd25860c0fec45f
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: zbao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/1390
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Thatcher features: Family 15 trinity FP2. Hudson.
close to Parmer.
This board and parmer both need to revert the change
http://review.coreboot.org/#/c/1359/, and add thatcher's own
chip.h,otherwise the mainboard_enable can not be called.
Change-Id: I54e1cfca845fbcea1d3aad5eff08d760d0d215c9
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: zbao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/1382
Tested-by: build bot (Jenkins)
Reviewed-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Before (data flag ignored -> broken):
66 DATA:
e944f1 JMP 1ff6
After (fixed):
66 DATA:
e944f1ffff JMP 00001ff8
This subtle difference in the length of decoded instruction meant
that the VBE call jumped to the routine setting AX=0x14F (VBE Failed)
instead of the routine that set AX=0x4F (VBE success).
The ability to run the same code in vm86 significantly aided the
debugging of this issue. Those X.org developers who would like to drop
vm86 better take special care towards _all_ vesa bugs, as those will
expose further issues.
Imported from:
http://cgit.freedesktop.org/xorg/xserver/commit/hw/xfree86/x86emu?id=cc2c73ddcb4370a7c3ad439cda4da825156c26c9
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Change-Id: Id08ead9b17468cf19ede45508e5dcc50e45b5acf
Signed-off-by: Luc Verhaegen <libv@skynet.be>
Tested-by: Luc Verhaegen <libv@skynet.be>
Reviewed-by: Adam Jackson <ajax@redhat.com>
Signed-off-by: Keith Packard <keithp@keithp.com>
Reviewed-on: http://review.coreboot.org/1365
Tested-by: build bot (Jenkins)
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
cc2c73ddcb4370a7c3ad439cda4da825156c26c9's three-cent titanium tax
doesn't go too far enough. Fix the rest of the call and jmp
instructions to handle the data prefix correctly.
Reference: Intel 64 and IA-32 Architectures Software Developer's Manual
Volume 2A: Instruction Set Reference, A-M
http://www.intel.com/Assets/PDF/manual/253666.pdf
Imported from:
http://cgit.freedesktop.org/xorg/xserver/commit/hw/xfree86/x86emu?id=bb18f277156c08be028a6e12d8987fb1593e9168
Signed-off-by: Stefan Reinauer <reinauer@google.com>
Change-Id: I83e6245d9748ee86722cfb7d8ac65258c35c013c
Reviewed-by: Julien Cristau <jcristau@debian.org>
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-on: http://review.coreboot.org/1366
Tested-by: build bot (Jenkins)
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
Reviewed-by: Patrick Georgi <patrick@georgi-clan.de>
This patch validates the previous "drop uma_memory_base" patches;
there are no more references to uma_memory_base when GFXUMA is not
selected.
Change-Id: I735b5e765b0c5cb4af1b4a7470cfe1af2bda7d38
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1385
Tested-by: build bot (Jenkins)
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
UMA region can be determined at any time after the amount
of RAM is known and before the uma_resource() call.
Change-Id: I2a0bf2d3cad55ee70e889c88846f962b7faa0c7e
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1379
Reviewed-by: Zheng Bao <zheng.bao@amd.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
Without GFXUMA, variables were not referenced anywhere.
Fail builds on Family10 if GFXUMA is selected, because the northbridge
code does not set UMA base or size.
Change-Id: I15b91cf6241e9a890398eed03824b753828a0a51
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1247
Reviewed-by: Zheng Bao <zheng.bao@amd.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
The code in rs690 or rs780 is always used with K8 or AMDFAM10
northbridge. Without GFXUMA, both of these set the same static value
indirectly using the variable uma_memory_base.
Make the register setting with immediate value, to remove the obscure
use of variable uma_memory_base.
Change-Id: I5354684457a76e73013b4e34a4538a6d122eee8d
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1246
Reviewed-by: Zheng Bao <zheng.bao@amd.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
I dont known if missed something, but why an extra 0x100 was added to limit?
My board would get the wrong memory table entry 7f000000-7fffffff as RAM, which
is higher than TOM.
coreboot memory table:
0. 0000000000000000-0000000000000fff: CONFIGURATION TABLES
1. 0000000000001000-000000000009ffff: RAM
2. 00000000000c0000-000000005e13efff: RAM
3. 000000005e13f000-000000005effffff: CONFIGURATION TABLES
4. 000000005f000000-000000007effffff: RESERVED
5. 000000007f000000-000000007fffffff: RAM
6. 00000000a0000000-00000000afffffff: RESERVED
Ronald G. Minnich:
I think someone who wrote the code was trying to round up the
next 0x100 boundary and did it incorrectly.
Here is code that would do it correctly:
limitk = ((resource_t)((d.mask + 0x00000ff) & 0x1fffff00)) << 9 ;
Zheng:
Plus 0xFF is correct, but the d.mask take bit 0 as enable it.
This bit should be clear when we try to calculate the limitk.
Change-Id: I3848ed5f23001e5bd61a19833650fe13df26eef3
Signed-off-by: Zheng Bao <zheng.bao@amd.com>
Signed-off-by: zbao <fishbaozi@gmail.com>
Reviewed-on: http://review.coreboot.org/1265
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Use of uma_resource() in northbridge code created a memory
resource marked as reserved. Such resources are removed
from system memory in write_coreboot_table().
Change-Id: I14bfd560140d8d30ec156562f23072bfae747bde
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1238
Tested-by: build bot (Jenkins)
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
With SandyBridge northbridge code, uma_memory_size was reset to
zero before variable MTRRs were set. This means MTRR setup routine
did not previously create a un-cacheable hole for uma.
Keep the behaviour that way, mmio_resource() has a prerequisuite that
the new region does not overlap with any cacheable ram_resource().
The result is not optimal setup in the number of used MTRRs, but
continue with this approach until MTRR algorithm is improved.
Change-Id: I63c8df19ad6b6350d46a3eca3055abf684b8b114
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1373
Tested-by: build bot (Jenkins)
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
Reserved memory resources will get removed from memory table at
the end of write_coreboot_table(),
Change-Id: I02711b4be4f25054bd3361295d8d4dc996b2eb3e
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1372
Tested-by: build bot (Jenkins)
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
This reverts commit 042c1461fb.
It turned out that sending IPIs via broadcast doesn't work on
Sandybridge. We tried to come up with a solution, but didn't
found any so far. So revert the code for now until we have
a working solution.
Change-Id: I7dd1cba5a4c1e4b0af366b20e8263b1f6f4b9714
Signed-off-by: Sven Schnelle <svens@stackframe.org>
Reviewed-on: http://review.coreboot.org/1381
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
This reverts commit 78efc4c36c.
The broadcast patch was reverted, so this commit should also
be reverted. The reason for reverting the broadcast patch:
It turned out that sending IPIs via broadcast doesn't work on
Sandybridge. We tried to come up with a solution, but didn't
found any so far. So revert the code for now until we have
a working solution.
Change-Id: I05c27dec55fa681f455215be56dcbc5f22808193
Signed-off-by: Sven Schnelle <svens@stackframe.org>
Reviewed-on: http://review.coreboot.org/1380
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Some usb debug devices don't respond fast enough. The linux kernel
(which uses almost the same usbdebug code) added a bit more
retry code, so let's copy that. Even if it might look stupid,
i pass the DBG_LOOPS argument through all functions to keep
the code at least a bit in sync with the linux kernel code.
Change-Id: I7c4b63b8bf1d2270fd6b8c8aa835e2cb324820bd
Signed-off-by: Sven Schnelle <svens@stackframe.org>
Reviewed-on: http://review.coreboot.org/1375
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
Compilation fails with set_debug_port undeclared in ramstage and
smm code. Fix that by adding usb_debug.c to the appropriate stages.
Change-Id: I2a037d3c5fab76ae6ea65c3a7f4d4e7561bb6d34
Signed-off-by: Sven Schnelle <svens@stackframe.org>
Reviewed-on: http://review.coreboot.org/1376
Tested-by: build bot (Jenkins)
Reviewed-by: Stefan Reinauer <stefan.reinauer@coreboot.org>
MRC messes with USB devices, so we have to reinitialize
USB debug after MRC has finished.
Change-Id: I45c0a687cebd69d0a31235bb870f8c455f42d4f2
Signed-off-by: Sven Schnelle <svens@stackframe.org>
Reviewed-on: http://review.coreboot.org/1377
Reviewed-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Tested-by: build bot (Jenkins)
Reviewed-by: Ronald G. Minnich <rminnich@gmail.com>
Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
Commit 2d42b34003 changed the
variable MTRR setup and removed compensation of uma_memory_size in
the cacheable memory resources.
Since the cacheable region size was no longer divisible by a large
power of 2, like 256 MB, this caused excessive use of MTRRs.
As first symptoms, slow boot with grub and poor user response.
As a solution, register the actual top of low ram with ram_resource(),
and do not subtract the UMA/TSEG regions from it.
TSEG may require further work as the original did not appear exactly
right to begin with. To have UMA as un-cacheable, use uma_resource().
Change-Id: I4ca99b5c2ca4e474296590b3d0c6ef5d09550d80
Signed-off-by: Kyösti Mälkki <kyosti.malkki@gmail.com>
Reviewed-on: http://review.coreboot.org/1239
Reviewed-by: Anton Kochkov <anton.kochkov@gmail.com>
Tested-by: build bot (Jenkins)
Stefan Reinauer