Intel 386 programmer's manual errata

[Pietro]

Hello to everyone. I briefly introduce myself, I'm Pietro, I'm from Italy and... I'm completely new to os development but willing to learn. Here is my first question.

Intel's 386 programmer reference manual has a lot of errors inside. Where do I find an errata-corrige for it? If none is available I'm interested in making one, so please send me your favorite bugs (schizophonic at tiscali dot it) and I will collect and publish them.

Thank you

Pietro

[Tim]

Intel's 386 programmer reference manual has a lot of errors inside.

Which errors? I've never seen any.

[Pietro]

Sorry, I've been imprecise. It is the *Italian translation* that has "lots of" errors, but now I also have downloaded from gaztek.sourceforge.net an English version (I had some hard time trying to find it at Intel's developer's page, but without success). I found in it just a couple of mistakes, i.e. F13 as opcode of STI in chapter 17 and 10 as type of executable segment descriptor in figure 6-1, but nothing more. Fool me. Anyways, does someone knows about other errors?

[Pietro]

Got another. At figure 5-6 (selector format) the three LS bits are said to be reserved for RPL when in facts the RPL bits are two. You can see the error as in the figure the index field starts at bit 4, when it really starts at bit 3 so that multiplied by 8 becomes the descriptor offset in the table.

[Tim]

Is that just in the Italian translation? (I don't have an English copy handy to check.)

Does that figure mention the TI bit at all, if RPL is erroneously described as taking 3 bits?

[Ready4Dis]

Ok, I brought up a dead topic, but I found this in google, so wanted to add:

Intel 386 manual, english version, I just found errata while programming my 386 emulator. If you look under the IN opcode, it says that
0xE5 (16-bit and 32-bit) are byte inputs, while it should say word and dword. That's what I get for using the official reference .

[Nessphoro]

Wow come on man, 2002.

[Ready4Dis]

Hey, it might be from 2002, but it's still valid for OS dev, and it's still in their official 386 reference (I downloaded directly from intel's site).

[Muneer]

In Page 91, the address translation overview diagram is wrong. The "Paging Enabled" pointer was wrongly placed.
And I cant remember but I think I did find an error in a table. I think it may be in page 111 "Useful Combination of E,G, And B bits",but I am not sure.