From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Ed Soloko" Subject: Intel 5400 EDAC Driver Issues Date: Thu, 3 Dec 2009 09:55:47 -0800 (PST) Message-ID: <013901ca7442$e96b0910$bc411b30$@com> Mime-Version: 1.0 Content-Type: multipart/mixed; boundary="===============1659517563674875923==" Return-path: Content-Language: en-us List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: bluesmoke-devel-bounces@lists.sourceforge.net To: bluesmoke-devel@lists.sourceforge.net Cc: mchehab@redhat.com List-Id: edac.vger.kernel.org This is a multi-part message in MIME format. --===============1659517563674875923== Content-Type: multipart/alternative; boundary="----=_NextPart_000_013A_01CA73FF.DB47C910" Content-Language: en-us This is a multi-part message in MIME format. ------=_NextPart_000_013A_01CA73FF.DB47C910 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Hi, I am working on the Intel 5400 EDAC driver as our product needs to monitor memory ecc errors and such. I have found a number of discrepancies between the code in [i5400_edac.c] and the 5400 MCH datasheet. I am wondering if anyone else has noticed them and more importantly has changes for them. Otherwise, I will embark upon the changes. The last commit to this code seems to be Roberson's of 2009-10-29. I noticed the following 1. The datasheet reference is for the 5000 MCH (minor, I know). 2. REDMEMB is an offset into Device 21,22, Function 1 but the code uses it as an offset into Device 16, function 1 via pvt->branchmap_werrors. 3. RECMEMA and RECMEMB usage is being incorrectly used similar to REDMEMB. 4. The values of RECMEMA, RECMEMB, and RECFB_DIMMA are 0xf0, 0xf4, 0xf8, respectively. These are incorrect. They should be 0xe0, 0xe4, 0xe8, respectively. 5. The pci device referenced by pvt_branch_0 is the device for Device 21, Function 0. There needs to be a pci_device in struct i5400_pvt for Device 21, Function 1. 6. Same comment as item #5 but for branch_1, Device 22, Function 1. 7. The function i5400_get_error_info() needs to obtain the values at offsets NRECMEMA, NRECMEMB, and REDMEMB, from Device 21,22, Function 1 instead of Device 16, Function 1. This is what I have found so far. If anyone can confirm what I have found or have changes already please let me know. Also, has this driver been used in production by anyone? Thanks, Ed Soloko GridIron Systems, Inc. ------=_NextPart_000_013A_01CA73FF.DB47C910 Content-Type: text/html; charset="us-ascii" Content-Transfer-Encoding: quoted-printable

Hi,

 

I am working on the Intel 5400 EDAC driver as our = product needs to monitor memory ecc errors and such. I have found a number of discrepancies between the code in [i5400_edac.c] and the 5400 MCH = datasheet. I am wondering if anyone else has noticed them and more importantly has = changes for them. Otherwise, I will embark upon the changes. The last commit to = this code seems to be Roberson’s of 2009-10-29.

 

I noticed the following

 

1.       The datasheet reference is for the 5000 MCH = (minor, I know).

2.       REDMEMB is an offset into Device 21,22, Function = 1 but the code uses it as an offset into Device 16, function 1 via pvt->branchmap_werrors.

3.       RECMEMA and RECMEMB usage is being incorrectly = used similar to REDMEMB.

4.       The values of RECMEMA, RECMEMB, and RECFB_DIMMA = are 0xf0, 0xf4, 0xf8, respectively. These are incorrect. They should be = 0xe0, 0xe4, 0xe8, respectively.

5.       The pci device referenced by pvt_branch_0 is the device  for Device 21, Function 0. There needs to be a pci_device = in struct i5400_pvt for Device 21, Function 1.

6.       Same comment as item #5 but for branch_1, Device = 22, Function 1.

7.       The function i5400_get_error_info() needs to = obtain the values at offsets NRECMEMA, NRECMEMB, and REDMEMB, from Device 21,22, = Function 1 instead of Device 16, Function 1.

 

This is what I have found so far.  If anyone = can  confirm what I have found or have changes already please let me = know.

 

Also, has this driver been used in production by = anyone?

 

Thanks,

 

Ed Soloko

GridIron Systems, Inc.

 

------=_NextPart_000_013A_01CA73FF.DB47C910-- --===============1659517563674875923== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline ------------------------------------------------------------------------------ Join us December 9, 2009 for the Red Hat Virtual Experience, a free event focused on virtualization and cloud computing. Attend in-depth sessions from your desk. Your couch. Anywhere. http://p.sf.net/sfu/redhat-sfdev2dev --===============1659517563674875923== Content-Type: text/plain; charset="us-ascii" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit Content-Disposition: inline _______________________________________________ bluesmoke-devel mailing list bluesmoke-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bluesmoke-devel --===============1659517563674875923==--