[Bug 1001558] New: GDB stub fails to send some registers.

[Bug 1001558] New: GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

           Summary: GDB stub fails to send some registers.
           Product: eCos
           Version: CVS
          Platform: All
        OS/Version: Other
            Status: NEW
          Severity: enhancement
          Priority: low
         Component: GDB stubs
        AssignedTo: unassigned@bugs.ecos.sourceware.org
        ReportedBy: ilijak@siva.com.mk
                CC: jifl@ecoscentric.com, ecos-bugs@ecos.sourceware.org,
                    john@dallaway.org.uk
             Class: Advice Request


GDB stub sometimes fails to send all registers. I encountered this problem in a
course of providing support for Cortex-M4F VFP registers.

The cause is function  in hal/common/current/src/thread-packets.c that sends
NUMREGS words that may be not correct if registers are not all same size and/or
some registers have been skipped.

Ilija

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

--- Comment #1 from Ilija Kocho <ilijak@siva.com.mk> 2012-04-11 21:46:05 BST ---
(In reply to comment #0)
> GDB stub sometimes fails to send all registers. I encountered this problem in a
> course of providing support for Cortex-M4F VFP registers.
> 
> The cause is function  in hal/common/current/src/thread-packets.c that sends
> NUMREGS words that may be not correct if registers are not all same size and/or
> some registers have been skipped.
> 
> Ilija

Here is more detailed description:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001524#c14

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

--- Comment #2 from Ilija Kocho <ilijak@siva.com.mk> 2012-04-11 21:51:52 BST ---
Created an attachment (id=1693)
 --> (http://bugs.ecos.sourceware.org/attachment.cgi?id=1693)
thread-packets.c patch

This patch seems like solution but has been tested only with Cortex-M4F target
(VFP4-SP-D16). For the time being i mark it as "works for me".

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

--- Comment #3 from Jonathan Larmour <jifl@ecoscentric.com> 2012-04-12 17:15:29 BST ---
HAL_STUB_REGISTERS_SIZE is not guaranteed to be defined. See hal_stub.c for
example, which sets it to NUMREGS by default if undefined.

That bit (with its accompanying comment) should be moved to hal_stub.h,
somewhere after the inclusion of the platform stub header.

Apart from that, I think the change is ok.

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

Jonathan Larmour <jifl@ecoscentric.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #1693|                            |assignment+, review-
               Flag|                            |

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #1693|0                           |1
        is obsolete|                            |

--- Comment #4 from Ilija Kocho <ilijak@siva.com.mk> 2012-04-25 19:52:40 BST ---
Created an attachment (id=1717)
 --> (http://bugs.ecos.sourceware.org/attachment.cgi?id=1717)
hal_stub_120425

(In reply to comment #3)
> HAL_STUB_REGISTERS_SIZE is not guaranteed to be defined. See hal_stub.c for
> example, which sets it to NUMREGS by default if undefined.
> 
> That bit (with its accompanying comment) should be moved to hal_stub.h,
> somewhere after the inclusion of the platform stub header.
> 
> Apart from that, I think the change is ok.

Here it is. By the way, one warning "set but unused variable" gets fixed.

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

--- Comment #5 from Jonathan Larmour <jifl@ecoscentric.com> 2012-04-25 20:16:03 BST ---
(In reply to comment #4)
> Here it is. By the way, one warning "set but unused variable" gets fixed.

Thanks.

A really trivial thing, but you took out the separator lines in hal_stub.c too:

-//-----------------------------------------------------------------------------
-// Extra eCos data.
-

Also there was a spelling mistake in the original comment: suffucient  should
be sufficient. With those, please commit. No need to post an updated patch
here.

Jifl

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

--- Comment #6 from Ilija Kocho <ilijak@siva.com.mk> 2012-04-25 20:57:49 BST ---
(In reply to comment #5)
> (In reply to comment #4)
> > Here it is. By the way, one warning "set but unused variable" gets fixed.
> 
> Thanks.
> 
> A really trivial thing, but you took out the separator lines in hal_stub.c too:
> 
> -//-----------------------------------------------------------------------------
> -// Extra eCos data.
> -

Oops.
Thanks.

> 
> Also there was a spelling mistake in the original comment: suffucient  should
> be sufficient. With those, please commit. No need to post an updated patch
> here.
> 

Fixed both mistakes.

Then checked in.

Ilija

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

--- Comment #6 from Ilija Kocho <ilijak@siva.com.mk> 2012-04-25 20:57:49 BST ---
(In reply to comment #5)
> (In reply to comment #4)
> > Here it is. By the way, one warning "set but unused variable" gets fixed.
> 
> Thanks.
> 
> A really trivial thing, but you took out the separator lines in hal_stub.c too:
> 
> -//-----------------------------------------------------------------------------
> -// Extra eCos data.
> -

Oops.
Thanks.

> 
> Also there was a spelling mistake in the original comment: suffucient  should
> be sufficient. With those, please commit. No need to post an updated patch
> here.
> 

Fixed both mistakes.

Then checked in.

Ilija

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

--- Comment #5 from Jonathan Larmour <jifl@ecoscentric.com> 2012-04-25 20:16:03 BST ---
(In reply to comment #4)
> Here it is. By the way, one warning "set but unused variable" gets fixed.

Thanks.

A really trivial thing, but you took out the separator lines in hal_stub.c too:

-//-----------------------------------------------------------------------------
-// Extra eCos data.
-

Also there was a spelling mistake in the original comment: suffucient  should
be sufficient. With those, please commit. No need to post an updated patch
here.

Jifl

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #1693|0                           |1
        is obsolete|                            |

--- Comment #4 from Ilija Kocho <ilijak@siva.com.mk> 2012-04-25 19:52:40 BST ---
Created an attachment (id=1717)
 --> (http://bugs.ecos.sourceware.org/attachment.cgi?id=1717)
hal_stub_120425

(In reply to comment #3)
> HAL_STUB_REGISTERS_SIZE is not guaranteed to be defined. See hal_stub.c for
> example, which sets it to NUMREGS by default if undefined.
> 
> That bit (with its accompanying comment) should be moved to hal_stub.h,
> somewhere after the inclusion of the platform stub header.
> 
> Apart from that, I think the change is ok.

Here it is. By the way, one warning "set but unused variable" gets fixed.

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

Jonathan Larmour <jifl@ecoscentric.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #1693|                            |assignment+, review-
               Flag|                            |

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

--- Comment #3 from Jonathan Larmour <jifl@ecoscentric.com> 2012-04-12 17:15:29 BST ---
HAL_STUB_REGISTERS_SIZE is not guaranteed to be defined. See hal_stub.c for
example, which sets it to NUMREGS by default if undefined.

That bit (with its accompanying comment) should be moved to hal_stub.h,
somewhere after the inclusion of the platform stub header.

Apart from that, I think the change is ok.

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

--- Comment #2 from Ilija Kocho <ilijak@siva.com.mk> 2012-04-11 21:51:52 BST ---
Created an attachment (id=1693)
 --> (http://bugs.ecos.sourceware.org/attachment.cgi?id=1693)
thread-packets.c patch

This patch seems like solution but has been tested only with Cortex-M4F target
(VFP4-SP-D16). For the time being i mark it as "works for me".

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.

[Bug 1001558] GDB stub fails to send some registers.

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001558

--- Comment #1 from Ilija Kocho <ilijak@siva.com.mk> 2012-04-11 21:46:05 BST ---
(In reply to comment #0)
> GDB stub sometimes fails to send all registers. I encountered this problem in a
> course of providing support for Cortex-M4F VFP registers.
> 
> The cause is function  in hal/common/current/src/thread-packets.c that sends
> NUMREGS words that may be not correct if registers are not all same size and/or
> some registers have been skipped.
> 
> Ilija

Here is more detailed description:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001524#c14

-- 
Configure bugmail: http://bugs.ecos.sourceware.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.