[Bug 1001796] New: AT91/ARM9 architecure extension

[Bug 1001796] New: AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

            Bug ID: 1001796
           Summary: AT91/ARM9 architecure extension
           Product: eCos
           Version: CVS
            Target: All
  Architecture/Host Other
                OS:
            Status: UNCONFIRMED
          Severity: enhancement
          Priority: low
         Component: HAL
          Assignee: unassigned@bugs.ecos.sourceware.org
          Reporter: bernd.edlinger@hotmail.de
                CC: ecos-bugs@ecos.sourceware.org

This is a spin-off of the enhancement bug# 1001649.

It contains all the extensions to the AT91 architecture that are necessary
to integrate the new ARM9 processor family members of the growing Atmel AT91
processor family.

One main objective of this patch is to ensure the existing ports do
not break because of these extensions.

This is done by changing as little as possible to the existing headers,
like var_io.h, while opening a way, how to integrate new processor without
touching these central header files all the time.

There are however a few bug-fixes included that may be of interest to other
ports as well:

- THUMB Mode was broken (did not compile)
  It does compile now, and is tested on AT91SAM9G45 platform.

- Fixed FIQ Enable Mode: FIQ CPSR Flag was task specific.

- Fixed Spurious Interrupts Crash.

- Fixed ARM_ICE_THREAD_SUPPORT. Must keep that at 0ffset 0x50.

- AT91SAM7s: Diag Port=UART2 did not compile.

- Fixed recv from debug port with serial line speed > 38400 baud.

- Enhanced accuracy in hal_delay_us() with PIT timer.

- Fixed overflow in hal_delay_us() with TC timer.

- Fixed crash in hal_if.c when no Debug port is configured.
  (by Number of communication channels on board = 0, Debug serial port = -1)

- Fixed compile error in arm_stub.c in some configurations.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug 1001796] AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

--- Comment #1 from Bernd Edlinger <bernd.edlinger@hotmail.de> ---
Created attachment 2125
  --> http://bugs.ecos.sourceware.org/attachment.cgi?id=2125&action=edit
enhancements to the AT91/ARM9 hal

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug 1001796] AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

Bernd Edlinger <bernd.edlinger@hotmail.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|low                         |normal

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug 1001796] AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

Bernd Edlinger <bernd.edlinger@hotmail.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Architecture/Host|Other                       |ARM
                 OS|                            |

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug 1001796] AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

--- Comment #2 from Bernd Edlinger <bernd.edlinger@hotmail.de> ---
Created attachment 2126
  --> http://bugs.ecos.sourceware.org/attachment.cgi?id=2126&action=edit
fix for kernel/current/tests/kcache2.c

This fixes some collateral damage in kcache2.c:

Some very small targets failed to compile this test sample.
Although the targets do not have any cache, kcache2.c may overflow
the .bss segment with a 64K test variable.

Reason: in preparation to enabling the DATA Cache, the macro
HAL_DCACHE_LINE_SIZE is now always defined to a power of 2.

This is suitable to place cache aligned variables with
__attribute__((aligned(HAL_DCACHE_LINE_SIZE)))

If the hardware does not support DCACHE however, HAL_DCACHE_LINE_SIZE is 4,
thus the variable is simply word-aligned.

Unfortunately the kcache2.c has a conditional code like
#ifdef HAL_DCACHE_LINE_SIZE
but it should better use #ifdef HAL_DCACHE_SIZE for this purpose.

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug 1001796] AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

Ilija Kocho <ilijak@siva.com.mk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Blocks|                            |1001649

-- 
You are receiving this mail because:
You are on the CC list for the bug.

[Bug 1001796] AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

Ilija Kocho <ilijak@siva.com.mk> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Blocks|                            |1001649

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug 1001796] AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

--- Comment #2 from Bernd Edlinger <bernd.edlinger@hotmail.de> ---
Created attachment 2126
  --> http://bugs.ecos.sourceware.org/attachment.cgi?id=2126&action=edit
fix for kernel/current/tests/kcache2.c

This fixes some collateral damage in kcache2.c:

Some very small targets failed to compile this test sample.
Although the targets do not have any cache, kcache2.c may overflow
the .bss segment with a 64K test variable.

Reason: in preparation to enabling the DATA Cache, the macro
HAL_DCACHE_LINE_SIZE is now always defined to a power of 2.

This is suitable to place cache aligned variables with
__attribute__((aligned(HAL_DCACHE_LINE_SIZE)))

If the hardware does not support DCACHE however, HAL_DCACHE_LINE_SIZE is 4,
thus the variable is simply word-aligned.

Unfortunately the kcache2.c has a conditional code like
#ifdef HAL_DCACHE_LINE_SIZE
but it should better use #ifdef HAL_DCACHE_SIZE for this purpose.

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug 1001796] AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

Bernd Edlinger <bernd.edlinger@hotmail.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|low                         |normal

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug 1001796] AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

Bernd Edlinger <bernd.edlinger@hotmail.de> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Architecture/Host|Other                       |ARM
                 OS|                            |

-- 
You are receiving this mail because:
You are the assignee for the bug.

[Bug 1001796] AT91/ARM9 architecure extension

[bugzilla-daemon]

Please do not reply to this email, use the link below.

http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001796

--- Comment #1 from Bernd Edlinger <bernd.edlinger@hotmail.de> ---
Created attachment 2125
  --> http://bugs.ecos.sourceware.org/attachment.cgi?id=2125&action=edit
enhancements to the AT91/ARM9 hal

-- 
You are receiving this mail because:
You are the assignee for the bug.