Talk:Silicon Image

Fair use rationale for Image:Silicon image logo.jpg
Image:Silicon image logo.jpg is being used on this article. I notice the image page specifies that the image is being used under fair use but there is no explanation or rationale as to why its use in this Wikipedia article constitutes fair use. In addition to the boilerplate fair use template, you must also write out on the image description page a specific explanation or rationale for why using this image in each article is consistent with fair use.

Please go to the image description page and edit it to include a fair use rationale. Using one of the templates at Fair use rationale guideline is an easy way to insure that your image is in compliance with Wikipedia policy, but remember that you must complete the template. Do not simply insert a blank template on an image page.

If there is other fair use media, consider checking that you have specified the fair use rationale on the other images used on this page. Note that any fair use images uploaded after 4 May, 2006, and lacking such an explanation will be deleted one week after they have been uploaded, as described on criteria for speedy deletion. If you have any questions please ask them at the Media copyright questions page. Thank you.

BetacommandBot 21:49, 25 October 2007 (UTC)

SiI3112
The disparaging comments about the 3112 are not supported by the citation given. The citation is to a email archive of a user group. Several posts indicated that they had trouble, but: Signal crostalk problems or power and ground plane issues in a motherboard design can cause errors even when the chips are not malfunctioning. In addition, many problems in systems have been traced to poor SATA cables.
 * The problems occurred with certain combinations of mother boards, disk drives, controller (3112), and software
 * There was no evidence presented to determine which component or components caused the problem
 * Several of the postings indicated that the problems they had were due to obsolete BIOS software or were the fault of the motherboard
 * Most importantly for this article, there was no evidence given that indicated that the 3112 itself was at fault

Since the claim in the main article rests, at best, on opinion, I believe that it should be removed.

UsersManual (talk) 08:57, 1 March 2008 (UTC)

"revision 1.50 date: 2003/12/08 09:22:20; author: sos; state: Exp; lines: +3 -1 More errata fixing for the SiI3112A disaster chip: Serialize access to the SATA channels, the chip messes up if both channels are used at the same time. The SiI3112 hereby takes the price as the most crappy SATA chip in existance by a significant amount. My advise to our userbase is to avoid this chip like the plague..." http://unix.derkeiler.com/Mailing-Lists/FreeBSD/stable/2005-07/0630.html ,And yes this guy knows his stuff. He's been writing ata drivers for MANY years. I'm inclined to take this denial of the problem as a censor attack from SII corporate PR machine. Electron9 (talk) 02:22, 2 March 2008 (UTC)

Based on your new information, it sounds like the chip does make it difficult to write a reliable driver for it, and that the driver programmers have been frustrated with it. But, if the driver can be written so that the device works, even at a performance loss, it still could still be a reasonable choice for some users, given the low price of boards made with it. You don't expect a Corolla to perform like a Ferrari.

Based on the information in the above link and some of the ones I found, it does seem that the 3112 can be made to work reliably. For example, in from 2004, it states that the Silicon Image driver doesn't exhibit the problems other drivers have had. And this more recent one says that the problems go away if the workarounds are properly applied.

At this point, I do believe you have made a good case that the performance under Linux and FreeBSD will be hurt by the workarounds, and that out of date drivers can cause serious problems. Would you be willing to change the wording on the page to reflect what is supported by the citations? Something like: "Released the SiI3112 SATA controller which, due to required workarounds in the driver, can cause significantly diminished performance in Linux and FreeBSD systems. In addition, users are strongly cautioned to ensure they always have the latest software, or data corruption can result."

UsersManual (talk) 08:00, 3 March 2008 (UTC)

I think "Responsible for the release of the SII3112 SATA controller chip which, due to required workarounds in the driver, will cause significantly diminished performance in a system. In addition, users are strongly cautioned to ensure that their drivers include the workaround for this deficiency." is a better formulation. So that the update rat race isn't encouraged. The 'SII' identifier tend to be used on actual chips. I don't see any reason why MS operating systems can't manage without this workaround so it should be considered in general, ie not specific to BSD/Linux. Only discovered there due the transparency of the coding process. However the problem with the workaround approach is:
 * 1) The confidence in that the chip manufacturer knows what they are doing is gone. And thus there might be plenty of other potential GOT-YOU (gotchas) just waiting to scramble your files.
 * 2) Disrespecting their end users by selling bad chips. For which they should have known that it's broken by design. And supplied an ERRATA.

This predatory behaviour from the manufactor must be frowned upon. Or they will have not reason to not repeat it. Electron9 (talk) 23:12, 3 March 2008 (UTC)

I agree with your comment about upgrades. I think the way you put it is better. However, I think that the "will cause significant ..." is not supported by the citations. As I understand it, the slowdowns occur only with certain disk drive models. How about "will often" instead? On the other hand, I think that there should be a mention of data corruption if the workarounds are not supplied. It should be clear that the consequences can be dire.

So how about: "Silicon Image is responsible for the release of the SII3112 SATA controller chip which, due to required workarounds in the driver, will often cause significantly diminished performance in a system. In addition, users are strongly cautioned to ensure that their drivers include the workaround for this deficiency, or data corruption can result."

By the way, I was thinking that it would be a good idea to make a separate section for this. I'd suggest "Product Notes" or something similar. I think it would be more prominent and easier to find there, and less likely to be deleted by later edits by PR types. What do you think?

In any case, you are right that it is irresponsible of them not to provide an ERRATA, making it very difficult for Linux and FreeBSD developer to produce reliable drivers.

UsersManual (talk) —Preceding comment was added at 07:51, 4 March 2008 (UTC)

I'm not sure why the SiI3112 problems (If any) should even be mentioned on the Wikipedia article for the company. Sure, the 3112 might be one of their more common chipsets, but Microsoft's Wikipedia article doesn't contain tips about working around buggy features, or advice for interoperability with Linux. At best, I reckon you can list some of their headline products, and, if you can find citations, mention that there are issues with them on non-Windows environments. Otherwise, this really just seems like anti-Silicon Image PR. Zelphar (talk) 20:20, 4 March 2008 (UTC)


 * Wikipedia is about sharing knowledge. And if I'm a potential buyer of equipment and find out that the company is a bad apple. I can make a more informed decision. Therefore information like this should be present. Microsoft's Wikipedia is run by corporate PR machine with lot's of cash. It's against their corporate interests to help Linux in any way, so no mystery to why that kind of missing info. On the MS-Win platform the manufacturer (SII) writes the driver, and their employees are not likely to point out that the product is crap. Not really a valid object of comparision. Besides that, the information is strongly relevant for anyone wishing to make an informed buying decision. The reasons to withhold that information is quite interesting. Electron9 (talk) 08:12, 6 March 2008 (UTC)


 * No, Wikipedia is an encyclopedia. It is not a how-to guide, a buyers guide, or a technical support resource. It is also unethical for those directly involved with a certain topic (Ie. MS employees for the article on MS) to edit that topic. Assuming Microsoft employees edit the MS article, which I don't believe you can prove (You're welcome to try), just because they do does not automatically mean you should fill the MS article with as much negative slightly-related information as possible. Likewise, it does not provide validation for putting such information in other topics, and then claim that the article was all filtered PR from the company in question, and that it needs such information to give it balance, and that any attempt to remove such information is all a huge cover up and conspiracy, etc. etc. To reiterate, I don't think that the information you want to add follows the Wikipedia guidelines, such as this one Zelphar (talk) 09:54, 9 March 2008 (UTC)

As I understand it, the Wikipedia guidelines say that any information that an editor wishes to add to a page is OK as long as it is factual, verifiable (via citations), fairly worded (neutral), and on topic. I think the wording that Electron9 and I have worked out meet the test. So, unless there are any further objections, I plan to replace the disputed section with the change described above in the next few days.

UsersManual (talk) 09:01, 8 March 2008 (UTC)

I have just discovered a fairly significant piece of information. In a message on osdir.com, it states that the problems arise "... because of faulty SATA implementation of the affected seagate hard drives combined with standard-compliant but peculiar behavior of silicon image controllers", and goes on to describe the details of what goes wrong. So, it seems that the fundamental fault lies with the drives, which don't meet the SATA standard. The 3112 does meet the standard, but operates within the standard in a way that triggers the bug in the disks (unlike other controllers). I think that the wording should be changed to reflect this fact.

UsersManual (talk) 18:53, 8 March 2008 (UTC)

I'm still suspicious regarding that controller. And their policy to not mention ANYTHING about it is enough to point it out. Anyway, I found some more info: home-tj.org - Sil_m15w. "workaround for 3512/3114 R_ERR on DMA activate FIS workaround." - Seems these controllers also have their share of silicon bugs. Electron9 (talk) 01:09, 9 March 2008 (UTC)

I agree that it is worth mentioning it, because there doesn't seem to be anywhere else for people to find out about such a serious compatibility issue. So I propose creating the "Product Notes" section, and in it saying: "Silicon Image released the SII3112 SATA controller chip, which requires driver workarounds to correct a compatibility problem with certain disk drive models. When activated on the affected drives, the workarounds will often cause significantly diminished performance in a system. In addition, users are strongly cautioned to ensure that their drivers include the workarounds, or data corruption can result.", along with links to citations for: the problem description, drive list, the performance loss, and data corruption instances. Can we agree to that?

UsersManual (talk) 07:56, 9 March 2008 (UTC)

"Silicon Image has released the SII3112, SII3114 and SII3512 S-ATA controller chips, which may cause corrupt data unless a driver workaround to correct compatibility problems with certain disk drive models is included. When activated with the affected disc drives, these workarounds will often cause significantly diminished performance in a system."

With the possible addition of: "In addition, users are strongly cautioned to ensure that their drivers include these workarounds, or data corruption can result."

More references:

"Buyer beware", or "Product notification" (sounds like PR coverup) heading should be ok. Electron9 (talk) 14:52, 9 March 2008 (UTC)

OK. I'll call the section "Product alerts" if that's alright with you. It seems more attention grabbing than "notification".

UsersManual (talk) 04:50, 10 March 2008 (UTC)

Sorry, I missed Zelphar's new comment and updated the page. I reverted it to the earlier contents until we can reach consensus here.

UsersManual (talk) 06:34, 10 March 2008 (UTC)

How about this: "Silicon Image has released the SII3112, SII3114 and SII3512 S-ATA controller chips, which may cause corrupt data storage unless a driver workaround to correct compatibility problems with certain disk drive models is included. When activated with the affected disc drives, these workarounds will often cause significantly diminished performance in a system." Electron9 (talk) 13:35, 10 March 2008 (UTC)
 * Product alerts **

That is mostly OK. However, I'm not comfortable saying that the controllers "may cause corrupt data", since the fault seems to lie more with the drives (at least for the Seagate ones). I would also change it to make the case of workaround(s) match. I.e. "Silicon Image has released the SII3112, SII3114 and SII3512 S-ATA controller chips, which, in order to avoid corrupt data storage, require driver workarounds to correct compatibility problems with certain disk drive models. When activated with the affected disc drives, these workarounds will often cause significantly diminished performance in a system." I would be willing to remove the dispute tag with that.

By the way, I found some direct links to the relevant information that you might want to include. They are:
 * For the performance loss:
 * For a description of the workarounds:
 * For the affected drives:

UsersManual (talk) 17:24, 11 March 2008 (UTC)

I've changed the description of the errata to be more specific and removed some links that were out of date or linked to information that was speculative or uninformed. —Preceding unsigned comment added by 70.76.47.20 (talk) 23:24, 31 December 2008 (UTC)


 * I find that the previous article used peacock words and had other puffiness issues and have done my best to correct them. I also felt that the product alerts section used mostly forum posts and other sources that don't qualify as a reliable source. From my perspective the content may be useful in a general sense, but not appropriate for Wikipedia, which is not a product review site. I've left a note on UsersManual's Talk page inviting him to discuss or potentially bring to the Reliable Sources Noticeboard for discussion with other experienced editors. User:King4057

Compromise
FYI, the source code bug was worked out here with Electron and I think we reached a reasonable compromise based on the current text. User:King4057 21:28, 26 May 2012 (UTC)