Skip to content

GitLab

Switch branch/tag
  1. 30 Nov, 2017 1 commit
    • David Cunado's avatar
      Do not enable SVE on pre-v8.2 platforms · 3872fc2d
      David Cunado authored 
      

Pre-v8.2 platforms such as the Juno platform does not have
the Scalable Vector Extensions implemented and so the build
option ENABLE_SVE is set to zero.

This has a minor performance improvement with no functional
impact.

Change-Id: Ib072735db7a0247406f8b60e325b7e28b1e04ad1
Signed-off-by: default avatarDavid Cunado <david.cunado@arm.com>
      3872fc2d
  3. 06 Nov, 2017 1 commit
  5. 01 Nov, 2017 1 commit
  7. 29 Aug, 2017 7 commits
  9. 15 Aug, 2017 1 commit
    • Julius Werner's avatar
      Add new alignment parameter to func assembler macro · 64726e6d
      Julius Werner authored 
      

Assembler programmers are used to being able to define functions with a
specific aligment with a pattern like this:

    .align X
  myfunction:

However, this pattern is subtly broken when instead of a direct label
like 'myfunction:', you use the 'func myfunction' macro that's standard
in Trusted Firmware. Since the func macro declares a new section for the
function, the .align directive written above it actually applies to the
*previous* section in the assembly file, and the function it was
supposed to apply to is linked with default alignment.

An extreme case can be seen in Rockchip's plat_helpers.S which contains
this code:

  [...]
  endfunc plat_crash_console_putc

  .align 16
  func platform_cpu_warmboot
  [...]

This assembles into the following plat_helpers.o:

  Sections:
  Idx Name                             Size  [...]  Algn
   9 .text.plat_crash_console_putc 00010000  [...]  2**16
  10 .text.platform_cpu_warmboot   00000080  [...]  2**3

As can be seen, the *previous* function actually got the alignment
constraint, and it is also 64KB big even though it contains only two
instructions, because the .align directive at the end of its section
forces the assembler to insert a giant sled of NOPs. The function we
actually wanted to align has the default constraint. This code only
works at all because the linker just happens to put the two functions
right behind each other when linking the final image, and since the end
of plat_crash_console_putc is aligned the start of platform_cpu_warmboot
will also be. But it still wastes almost 64KB of image space
unnecessarily, and it will break under certain circumstances (e.g. if
the plat_crash_console_putc function becomes unused and its section gets
garbage-collected out).

There's no real way to fix this with the existing func macro. Code like

 func myfunc
 .align X

happens to do the right thing, but is still not really correct code
(because the function label is inserted before the .align directive, so
the assembler is technically allowed to insert padding at the beginning
of the function which would then get executed as instructions if the
function was called). Therefore, this patch adds a new parameter with a
default value to the func macro that allows overriding its alignment.

Also fix up all existing instances of this dangerous antipattern.

Change-Id: I5696a07e2fde896f21e0e83644c95b7b6ac79a10
Signed-off-by: default avatarJulius Werner <jwerner@chromium.org>
      64726e6d
  11. 14 Jul, 2017 1 commit
  13. 05 Jul, 2017 1 commit
  15. 30 Jun, 2017 1 commit
    • Caesar Wang's avatar
      rockchip/rk3399: fixes the typo and the WARNINGS during suspend/resume · c3710ee7
      Caesar Wang authored 
      

This patch fixes the two things as follows:

1) rk3399_flash_l2_b" seems to be a typo. That's "flush", not "flash".

2) fixes the warnings log.
We always hit the warnings thing during the suspend, as below log:
..
[   51.022334] CPU5: shutdown
[   51.025069] psci: CPU5 killed.
INFO:    sdram_params->ddr_freq = 928000000
WARNING: rk3399_flash_l2_b:reg 28830380,wait

When the L2 completes the clean and invalidate sequence, it asserts the
L2FLUSHDONE signal. The SoC can now deassert L2FLUSHREQ signal and then
the L2 deasserts L2FLUSHDONE.

Then, a loop without a delay isn't really great to measure time. We should
probably add a udelay(10) or so in there and then maybe replace the WARN()
after the loop. In the actual tests, the L2 cache will take ~4ms by
default for big cluster.

In the real world that give 10ms for the enough margin, like the
ddr/cpu/cci frequency and other factors that will affect it.

Change-Id: I55788c897be232bf72e8c7b0e10cf9b06f7aa50d
Signed-off-by: default avatarCaesar Wang <wxt@rock-chips.com>
      c3710ee7
  17. 28 Jun, 2017 1 commit
    • Caesar Wang's avatar
      rockchip: enable A53's erratum 855873 for rk3399 · dea1e8ee
      Caesar Wang authored 
      

For rk3399, the L2ACTLR[14] is 0 by default, as ACE CCI-500 doesn't
support WriteEvict. and you will hit the condition L2ACTLR[3] with 0,
as the Evict transactions should propagate to CCI-500 since it has
snoop filters.

Maybe this erratum applies to all Cortex-A53 cores so far, especially
if RK3399's A53 is a r0p4. we should enable it to avoid data corruption,

Change-Id: Ib86933f1fc84f8919c8e43dac41af60fd0c3ce2f
Signed-off-by: default avatarCaesar Wang <wxt@rock-chips.com>
      dea1e8ee
  19. 26 Jun, 2017 1 commit
  21. 14 Jun, 2017 1 commit
  23. 08 Jun, 2017 10 commits
  25. 15 May, 2017 1 commit
  27. 10 May, 2017 3 commits
  29. 03 May, 2017 1 commit
  31. 25 Apr, 2017 1 commit
    • tony.xie's avatar
      rockchip: rk3328: support rk3328 · 0d5ec955
      tony.xie authored 
      
rk3328 is a Quad-core soc and Cortex-a53 inside!
This patch supports the following functions:
1、power up/off cpus
2、suspend/resume cpus
3、suspend/resume system
4、reset system
5、power off system

Change-Id: I60687058d13912c6929293b06fed9c6bc72bdc84
Signed-off-by: default avatartony.xie <tony.xie@rock-chips.com>
      0d5ec955
  33. 07 Apr, 2017 1 commit
  35. 04 Apr, 2017 1 commit
  37. 20 Mar, 2017 1 commit
  39. 03 Mar, 2017 1 commit
  41. 01 Mar, 2017 1 commit
  43. 24 Feb, 2017 2 commits