Simple question about RTL

Tue Sep 6 03:48:53 UTC 2016

Hi Pavel,
it works on my Zynq platform. Thankyou. I will go through the patches you posted.

Best wishes,
Tim Tian
xuelin.tian at qkmtech.com

From: Pavel Pisa
Date: 2016-09-06 08:52
To: Chris Johns
CC: xuelin.tian at qkmtech.com; users at rtems.org
Subject: Re: Simple question about RTL
Hello Chris,

On Tuesday 06 of September 2016 01:52:05 Chris Johns wrote:
> On 05/09/2016 23:52, Pavel Pisa wrote:
> > I can prepare 4.11 branch on my GitHub account and if it is confirmed,
> > that it works, then discussion about inclusion to yet unreleased 4.11
> > should start. Other option is to release 4.11 and correct these for
> > 4.11.1.
>
> Please post the patches to devel for review. If they are fixing issues
> in libdl I am ok with pushing to 4.11 before 4.11.0 is released.

I have spent huge amount today on shuffling patches but I have
not 4.11 booting on RPi2 yet. Problem is that code before branching
cannot run with actual firmware and U-boot versions. Hypervisor
boot is required.

Fixing RTL on Zynq requires to at least these patches which I have
listed in previous e-mails. Problem is, that there are dependencies
on my previous generic cache management patches. The case of Zynq
and Altera is not so bad, cache manager has included specific support
for them. But there is significant problem that almost all other
ARM BSPs (including Raspberry Pi) with cache have not selected
any cache operations in the past due to the condition in generic
cache support. To do that functional requires backport another
patches.

So I have ended with something like 25 patches

  https://github.com/ppisa/rtems/tree/4.11-arm-update

and debugging still why RPi2 does not boot.

I consider patches marked by "Z" as critical for RTL on Zynq.
Patches marked by "O" as optional/may be required by dependencies
or Zynq boot. The rest is mostly RPi stuff and fix for rest of ARMs

  ae58402 bsps/arm: CP15 support for flush prefetch buffer and table base control.
  bbfee81 arm/raspberrypi: ensure that RTEMS application image can be started by U-boot.
  211ff5f arm/score and shared: define ARM hypervisor mode and alternate vector table base access.
Z 1d640b1 score/arm: Ensure that copile time alignment is 64 bytes for Cortex-A multilib.
Z 52d170a rtems+bsps/cache: Define cache manager operations for code synchronization and maximal alignment.
Z ada6141 bsp/arm: Report correct maximal cache line length for ARM Cortex-A + L2C-310.
  24b045a bsps/arm: Change code to explicit selection of cache implementation for ARM BSPs.
  303a314 bsps/arm: basic on core cache support changed to use l1 functions.
Z 07ed33e libdl/rtl-obj.c: synchronize cache after code relocation.
  12b2428 bsps/arm: do not disable MMU during translation table management operations.
  65739ba arm/bsps: CP15 and basic cache support entire cache clean for more architecture variants now.
  1b7c3cc arm/raspberrypi: use cache manager operations to flush/invalidate all cache levels.
O c9fb885 arm/xilinx_zynq: ensure that cache is cleaned and MMU disabled when initialization starts.
  0504357 bsps/arm: use defines for cache type register format field.
O 531ac9b bsps/arm: remove lock in arm_cp15_set_translation_table_entries().
  6519972 bsps/arm: reorganize CP15 code to allow clean and invalidate ARMv7 cache by level.
  8904d09 bsps/arm: Fix basic cache support for SMP
  4192010 libbsp/arm: Add the TTB table to the default MMU set up as read/write.
  22f5808 libbsp/arm: Fix ARM BSPs missing the bsp_translation_table_end symbol.
  7d143ff bsps/arm: Support recent bootloaders starting kernel in HYP mode
  e94cc10 arm/raspberrypi: Enable HYP to SVC switch for this BSP.
  fe85ebd arm/raspberrypi: reorder and update MMU config table to nor force RW section later to RO.
  08374dc arm/raspberrypi: ensure that correct RPI_PERIPHERAL_BASE is provided by raspberrypi.h
  cc6e7ae arm/raspberrypi: remove duplicate setup of IRQ handler in the main ARM exception table.
  4f362eb arm/raspberrypi: change interrupt dispatch and enable to generic vector id based approach.

I am not sure if that growing has chance to be acceptable for 4.11.
I have not even tested Zynq build. I do not have board there at home.
Incremental testing on Zynq has advantage that it runs and change breaking
code is immediately seen.

All patches are backports from mainline, some RPi ones have removed touching
these files and subsystems which are not present on 4.11. So generally,
they should be OK. But ....

Best wishes,

                Pavel

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/users/attachments/20160906/22f95223/attachment.html>