[Bug 1001269] New: Via Rhine Ethernet driver is reading the MAC incorrectly

[Bug 1001269] New: Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

           Summary: Via Rhine Ethernet driver is reading the MAC
                    incorrectly
           Product: eCos
           Version: CVS
          Platform: pc (i386 PC target)
        OS/Version: IA32
            Status: UNCONFIRMED
          Severity: major
          Priority: low
         Component: Ethernet
        AssignedTo: unassigned@bugs.ecos.sourceware.org
        ReportedBy: stephen@centtech.com
                CC: ecos-bugs@ecos.sourceware.org
             Class: Advice Request


Created an attachment (id=1292)
 --> (http://bugs.ecos.sourceware.org/attachment.cgi?id=1292)
if_rhine.c patch

The current Via Rhine Ethernet driver was designed to access the PCI card by
way of I/O reads and writes.  Early in the code the driver acquires the I/O
base and stores in the cpd->base structure.  Then, the driver uses
HAL_PCI_IO_READs and HAL_PCI_IO_WRITEs macros to access the card.

However, for some unknown reason the driver reloads the MAC address using the
cpd->base variable as a memory address.  This causes the driver to load random
data for the MAC.  Later, REDBOOT fails because it passes a bad MAC to the DHCP
server.

The fix was to change the code to access the card via macros.

Old Code:

        // Use the address from the serial EEPROM
        p = cpd->base + RHINE_PAR0;
        for (i = 0; i < 6; i++)
            cpd->esa[i] = *p++;

New Code:

        // Use the address from the serial EEPROM
        for (i = 0; i < 6; i++) {
      HAL_PCI_IO_READ_UINT8(cpd->base + RHINE_PAR0 + i, cpd->esa[i]);
    }

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

Sergei Gavrikov <sergei.gavrikov@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #1292|0                           |1
        is obsolete|                            |

--- Comment #1 from Sergei Gavrikov <sergei.gavrikov@gmail.com> 2011-08-28 10:45:26 BST ---
Created an attachment (id=1356)
 --> (http://bugs.ecos.sourceware.org/attachment.cgi?id=1356)
via_rhine: Fixed ESA management and trivial warning fix.

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #2 from Sergei Gavrikov <sergei.gavrikov@gmail.com> 2011-08-28 11:00:19 BST ---
Hi Stephen,

Thank you for your bug report and workaround.

One thing. I have not seen a reason to suppress a debug message in
find_rhine_match_func(), so, I have rejected that tweak from your patch
(also we use 4 spaces indents).

When I tested build of the driver I found 1 annoying compiler warning
from via_rhine.h and I've decided to put the fix in cummulative patch.

If would be great if you could test the patch on real hardware. If you
could do it, let me know, please (I mean an attachment 1356).

Sergei

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #3 from stephen@centtech.com 2011-09-06 20:24:21 BST ---
Hi Sergei,


(In reply to comment #2)
> Hi Stephen,
> 
> Thank you for your bug report and workaround.
> 
> One thing. I have not seen a reason to suppress a debug message in
> find_rhine_match_func(), so, I have rejected that tweak from your patch
> (also we use 4 spaces indents).
> 
> When I tested build of the driver I found 1 annoying compiler warning
> from via_rhine.h and I've decided to put the fix in cummulative patch.
> 
> If would be great if you could test the patch on real hardware. If you
> could do it, let me know, please (I mean an attachment 1356 [details]).
> 
> Sergei

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #4 from stephen@centtech.com 2011-09-06 20:27:56 BST ---
Hi Sergei,

Yes, I will test the patch out.  I have the hardware in my office.
I will email you when the testing is done.

Thanks,

Stephen


(In reply to comment #3)
> Hi Sergei,
> 
> 
> (In reply to comment #2)
> > Hi Stephen,
> > 
> > Thank you for your bug report and workaround.
> > 
> > One thing. I have not seen a reason to suppress a debug message in
> > find_rhine_match_func(), so, I have rejected that tweak from your patch
> > (also we use 4 spaces indents).
> > 
> > When I tested build of the driver I found 1 annoying compiler warning
> > from via_rhine.h and I've decided to put the fix in cummulative patch.
> > 
> > If would be great if you could test the patch on real hardware. If you
> > could do it, let me know, please (I mean an attachment 1356 [details] [details]).
> > 
> > Sergei

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #5 from Sergei Gavrikov <sergei.gavrikov@gmail.com> 2011-10-02 15:51:59 BST ---
(In reply to comment #4)

> Yes, I will test the patch out.  I have the hardware in my office.
> I will email you when the testing is done.

Stephen, could you test the patch? We appreciate any feedback.

Sergei

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #6 from stephen@centtech.com 2011-10-03 15:36:00 BST ---
(In reply to comment #5)
> (In reply to comment #4)
> 
> > Yes, I will test the patch out.  I have the hardware in my office.
> > I will email you when the testing is done.
> 
> Stephen, could you test the patch? We appreciate any feedback.
> 
> Sergei

Sorry.  Yes.  I will test the patch.  

Thanks,

Stephen

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #8 from Sergei Gavrikov <sergei.gavrikov@gmail.com> 2012-01-17 18:48:55 GMT ---
(In reply to comment #7)
> Hello Sergei.  I am sorry for the delay.  My boss had me working on
> other stuff until recently.  I applied your patch to the latest ecos
> build tree. I have verified that your patch [attachment 1356] fixes
> the bug in the VIA Rhine driver.  
> 
> I built the VIA Rhine driver for the pc redboot package.  This is a
> new package that I will be uploading to ECOS very shortly.  I was
> successful in loading and running hello.c via GDB network debugging
> using a D-Link DFE-530TX Rev-A3-1 PCI card.

> Please commit your changes to the ECOS tree at your earliest
> convenience.
> 
> Thanks,
> 
> Stephen

Hello Stephen,

Thank you for testing and your contribution to eCos. Checked-in.

Sergei

-- 
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 1001269] New: Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

           Summary: Via Rhine Ethernet driver is reading the MAC
                    incorrectly
           Product: eCos
           Version: CVS
          Platform: pc (i386 PC target)
        OS/Version: IA32
            Status: UNCONFIRMED
          Severity: major
          Priority: low
         Component: Ethernet
        AssignedTo: unassigned@bugs.ecos.sourceware.org
        ReportedBy: stephen@centtech.com
                CC: ecos-bugs@ecos.sourceware.org
             Class: Advice Request


Created an attachment (id=1292)
 --> (http://bugs.ecos.sourceware.org/attachment.cgi?id=1292)
if_rhine.c patch

The current Via Rhine Ethernet driver was designed to access the PCI card by
way of I/O reads and writes.  Early in the code the driver acquires the I/O
base and stores in the cpd->base structure.  Then, the driver uses
HAL_PCI_IO_READs and HAL_PCI_IO_WRITEs macros to access the card.

However, for some unknown reason the driver reloads the MAC address using the
cpd->base variable as a memory address.  This causes the driver to load random
data for the MAC.  Later, REDBOOT fails because it passes a bad MAC to the DHCP
server.

The fix was to change the code to access the card via macros.

Old Code:

        // Use the address from the serial EEPROM
        p = cpd->base + RHINE_PAR0;
        for (i = 0; i < 6; i++)
            cpd->esa[i] = *p++;

New Code:

        // Use the address from the serial EEPROM
        for (i = 0; i < 6; i++) {
      HAL_PCI_IO_READ_UINT8(cpd->base + RHINE_PAR0 + i, cpd->esa[i]);
    }

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

Sergei Gavrikov <sergei.gavrikov@gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Attachment #1292|0                           |1
        is obsolete|                            |

--- Comment #1 from Sergei Gavrikov <sergei.gavrikov@gmail.com> 2011-08-28 10:45:26 BST ---
Created an attachment (id=1356)
 --> (http://bugs.ecos.sourceware.org/attachment.cgi?id=1356)
via_rhine: Fixed ESA management and trivial warning fix.

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #2 from Sergei Gavrikov <sergei.gavrikov@gmail.com> 2011-08-28 11:00:19 BST ---
Hi Stephen,

Thank you for your bug report and workaround.

One thing. I have not seen a reason to suppress a debug message in
find_rhine_match_func(), so, I have rejected that tweak from your patch
(also we use 4 spaces indents).

When I tested build of the driver I found 1 annoying compiler warning
from via_rhine.h and I've decided to put the fix in cummulative patch.

If would be great if you could test the patch on real hardware. If you
could do it, let me know, please (I mean an attachment 1356).

Sergei

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #3 from stephen@centtech.com 2011-09-06 20:24:21 BST ---
Hi Sergei,


(In reply to comment #2)
> Hi Stephen,
> 
> Thank you for your bug report and workaround.
> 
> One thing. I have not seen a reason to suppress a debug message in
> find_rhine_match_func(), so, I have rejected that tweak from your patch
> (also we use 4 spaces indents).
> 
> When I tested build of the driver I found 1 annoying compiler warning
> from via_rhine.h and I've decided to put the fix in cummulative patch.
> 
> If would be great if you could test the patch on real hardware. If you
> could do it, let me know, please (I mean an attachment 1356 [details]).
> 
> Sergei

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #4 from stephen@centtech.com 2011-09-06 20:27:56 BST ---
Hi Sergei,

Yes, I will test the patch out.  I have the hardware in my office.
I will email you when the testing is done.

Thanks,

Stephen


(In reply to comment #3)
> Hi Sergei,
> 
> 
> (In reply to comment #2)
> > Hi Stephen,
> > 
> > Thank you for your bug report and workaround.
> > 
> > One thing. I have not seen a reason to suppress a debug message in
> > find_rhine_match_func(), so, I have rejected that tweak from your patch
> > (also we use 4 spaces indents).
> > 
> > When I tested build of the driver I found 1 annoying compiler warning
> > from via_rhine.h and I've decided to put the fix in cummulative patch.
> > 
> > If would be great if you could test the patch on real hardware. If you
> > could do it, let me know, please (I mean an attachment 1356 [details] [details]).
> > 
> > Sergei

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #5 from Sergei Gavrikov <sergei.gavrikov@gmail.com> 2011-10-02 15:51:59 BST ---
(In reply to comment #4)

> Yes, I will test the patch out.  I have the hardware in my office.
> I will email you when the testing is done.

Stephen, could you test the patch? We appreciate any feedback.

Sergei

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #6 from stephen@centtech.com 2011-10-03 15:36:00 BST ---
(In reply to comment #5)
> (In reply to comment #4)
> 
> > Yes, I will test the patch out.  I have the hardware in my office.
> > I will email you when the testing is done.
> 
> Stephen, could you test the patch? We appreciate any feedback.
> 
> Sergei

Sorry.  Yes.  I will test the patch.  

Thanks,

Stephen

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

stephen@centtech.com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|                            |CURRENTRELEASE

--- Comment #7 from stephen@centtech.com 2012-01-17 17:49:11 GMT ---
(In reply to comment #5)
> (In reply to comment #4)
> 
> > Yes, I will test the patch out.  I have the hardware in my office.
> > I will email you when the testing is done.
> 
> Stephen, could you test the patch? We appreciate any feedback.
> 
> Sergei

Hello Sergei.  I am sorry for the delay.  My boss had me working on 
other stuff until recently.  I applied your patch to the latest ecos 
build tree. I have verified that your patch [attachment 1356]
fixes the bug in the VIA Rhine driver.  

I built the VIA Rhine driver for the pc redboot package.  This
is a new package that I will be uploading to ECOS very shortly.  
I was successful in loading and running hello.c via GDB network debugging
using a D-Link DFE-530TX Rev-A3-1 PCI card.

Please commit your changes to the ECOS tree at your earliest convenience.

Thanks,

Stephen

-- 
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 1001269] Via Rhine Ethernet driver is reading the MAC incorrectly

[bugzilla-daemon]

Please do not reply to this email. Use the web interface provided at:
http://bugs.ecos.sourceware.org/show_bug.cgi?id=1001269

--- Comment #8 from Sergei Gavrikov <sergei.gavrikov@gmail.com> 2012-01-17 18:48:55 GMT ---
(In reply to comment #7)
> Hello Sergei.  I am sorry for the delay.  My boss had me working on
> other stuff until recently.  I applied your patch to the latest ecos
> build tree. I have verified that your patch [attachment 1356] fixes
> the bug in the VIA Rhine driver.  
> 
> I built the VIA Rhine driver for the pc redboot package.  This is a
> new package that I will be uploading to ECOS very shortly.  I was
> successful in loading and running hello.c via GDB network debugging
> using a D-Link DFE-530TX Rev-A3-1 PCI card.

> Please commit your changes to the ECOS tree at your earliest
> convenience.
> 
> Thanks,
> 
> Stephen

Hello Stephen,

Thank you for testing and your contribution to eCos. Checked-in.

Sergei

-- 
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.