Hi list, We use pf+ALTQ for trafic shaping on some routers. We are switching to new servers: Dell PowerEdge R620 with 2 8-cores Intel. (E5-2650L), 8GB RAM and Intel I 350T4 > (quad port) using igb driver.
Click to expand.I'm having a lot of trouble parsing what you're trying to say, but I think what you're getting at is actually the 'subsystem vendor ID', not the 'device ID'. On any of these cards (legit or not) I would expect the vendor ID to be intel's, and the device ID to be an intel part (AM4). On a legit device using an AM4 chip on an OEM board the subsystem vendor ID should belong to the OEM, and the subsystem ID will be some random number that nobody except the OEM cares about. On a counterfeit the subsystem vendor ID & subsystem ID will be copied from an actual intel part. 'PCI ID' commonly refers to the combination of the device ID and the vendor ID.
The vendor ID and subsystem vendor ID are allocated by PCI-SIG and unique to each vendor; the device ID and subsystem ID are managed by the vendors, and are not unique across vendors. I'm having a lot of trouble parsing what you're trying to say, but I think what you're getting at is actually the 'subsystem vendor ID', not the 'device ID'. On any of these cards (legit or not) I would expect the vendor ID to be intel's, and the device ID to be an intel part (AM4). On a legit device using an AM4 chip on an OEM board the subsystem vendor ID should belong to the OEM, and the subsystem ID will be some random number that nobody except the OEM cares about. On a counterfeit the subsystem vendor ID & subsystem ID will be copied from an actual intel part. 'PCI ID' commonly refers to the combination of the device ID and the vendor ID. The vendor ID and subsystem vendor ID are allocated by PCI-SIG and unique to each vendor; the device ID and subsystem ID are managed by the vendors, and are not unique across vendors.
Click to expand.thanks for correction, Hard to focus while typing on my small screen moto x. The format Vendor ID: subsystemID deviceID, device ID can be existed or not. If you look on subsystem ID for I350, there are more than one belong to Intel. Faker copied Intel the same (subsystem) ID regardless of i340 or I350 since (on linux) the hardware uses igb module. On my understanding: Intel has release other subsystemID in the past due on hardware fix or make cheaper by removing/replace logics.
SubsystemID is needed (on linux) to make decision on the code to handle the hardware. Worst case scenario is compatibality and performance issue. Other story in USB ID: this has been an issue with fake USB-serial dongle. The fake dongle works on regular usage, but fail on specific commands. Most on ebay USB-Serial dongle are fakes. Click to expand.The Winyao cards are not copies of the Intel cards - take a look at -they use their own card designs, and the layouts are totally different. I don't know where Winyao source the Intel chips to drive their cards, but I don't class them the same as the 'counterfeit' cards.
![For For](/uploads/1/2/5/6/125650591/663448169.jpg)
Especially because Winyao is actually their real retail brand! DealExtreme sell both the 2 and 4 port versions, with free worldwide shipping - and if anyone is looking for a cheap not-made-by-Intel card. I've got a couple, and they work with the standard Intel ProSet driver packages.
The Winyao cards are not copies of the Intel cards - take a look at -they use their own card designs, and the layouts are totally different. I don't know where Winyao source the Intel chips to drive their cards, but I don't class them the same as the 'counterfeit' cards. Especially because Winyao is actually their real retail brand! DealExtreme sell both the 2 and 4 port versions, with free worldwide shipping - and if anyone is looking for a cheap not-made-by-Intel card. I've got a couple, and they work with the standard Intel ProSet driver packages.
I for one don't see much a problem with Winyao NICs, while it is not super clear due to their naming conventions bearing some resemblance to Intels NICs, they are in fact different. Mind you, I don't consider them of equal quality as cards from the larger brands (Intel, Supermicro, HP, IBM, etc.), but the fact that you can distinguish them, allows you to make an educated guess as to whether or not you think they are worth buying. I consider them to be of same quality as Realtek based NICs, but with a better chip. And for some application this may well be acceptable. The cards that I would avoid are the ones that are actually fake/counterfeit, ie. The ones that are sold as real i350-t4 etc., because for obvious reasons you have no idea what you are actually getting. On another note, it might be interesting to compare some of the various 'chinese' i350 quad port NICs, some like in post #0 are just clones of the i350-T4 but others are different, most use the MR-MACOM chips instead of Delta, but there are also these.
Click to expand.I am not sold on your method of identifying 'fake' cards, 8086:1521 appears to be very common for identifying i350 based cards. I have two i350 based NICs a Fujitsu quad port and a Winyao dual port (for PCIe X1 slot), both of which identify themselves as 8086:1521. The Fujitsu also mentions another value around Subsystem with 'lspci -v' '11d1'. If I am reading the output from pciconf (for pfsense) the Winyao uses an additional identifier of 1d1a. Of course even my Fujitsu card could be counterfeit, but I seriously doubt that is the case. If you think about it, it makes sense that the device first identifies itself as having an Intel device, then that it is uses an i350 chip and finally who actually made the card. All of these cards uses the same driver, and I see no reason why Intel would want to maintain a list of cards using this chip in their driver., also indicates that the Supermicro AOC-SGP-I2, also uses 8086:1521.
I for one don't see much a problem with Winyao NICs, while it is not super clear due to their naming conventions bearing some resemblance to Intels NICs, they are in fact different. Mind you, I don't consider them of equal quality as cards from the larger brands (Intel, Supermicro, HP, IBM, etc.), but the fact that you can distinguish them, allows you to make an educated guess as to whether or not you think they are worth buying. I consider them to be of same quality as Realtek based NICs, but with a better chip. And for some application this may well be acceptable.
The cards that I would avoid are the ones that are actually fake/counterfeit, ie. The ones that are sold as real i350-t4 etc., because for obvious reasons you have no idea what you are actually getting. On another note, it might be interesting to compare some of the various 'chinese' i350 quad port NICs, some like in post #0 are just clones of the i350-T4 but others are different, most use the MR-MACOM chips instead of Delta, but there are also these: I am not sold on your method of identifying 'fake' cards, 8086:1521 appears to be very common for identifying i350 based cards. I have two i350 based NICs a Fujitsu quad port and a Winyao dual port (for PCIe X1 slot), both of which identify themselves as 8086:1521.
The Fujitsu also mentions another value around Subsystem with 'lspci -v' '11d1'. If I am reading the output from pciconf (for pfsense) the Winyao uses an additional identifier of 1d1a. Of course even my Fujitsu card could be counterfeit, but I seriously doubt that is the case. If you think about it, it makes sense that the device first identifies itself as having an Intel device, then that it is uses an i350 chip and finally who actually made the card.
All of these cards uses the same driver, and I see no reason why Intel would want to maintain a list of cards using this chip in their driver., also indicates tthe Supermicro AOC-SGP-I2, also uses 8086:1521. Click to expand.Those are real Intel oem Those faker are using low quality components.
If you make your own product. You have to register your own pci I'd. Since tour products are not oem Looks further on device Id. I know Dell and hphas unique number. Please read my previous posting.
Please let me lend my flashligh, would not you. Hehehhe:. 11d1 - C600/X79 seriechipset PCI Express Root Port 1: your card is sitting on pci express root 1.
1d1a -C600/X79 series chipset PCI Express Root Port 6: your card is sitting on pci express root 6 8086:1521 has additional device ID. On Intel product mostly have none ( this faker using intel PCI ID) some examples: 8086:1521:1028 0602, 1028 - Dell product 8086:1521:103c 2003, 103c 0 HP product 8086:1521:17aa 4005, 17AA - Lenov Product 8086:1521:8086 0002 - Intel product ( faker uses this pci id to piggy back intel driver update) some OEM product (I know IBM) using this 8086:1521:8086 XXXX, I believe they pay licenses directly to Intel instead of registering unique sub-vendor ID. On WinYao, I already posted, check active componets, ethernet transformer ballon on WinYao is not good as the real Intel OEM, you can find the spec in pdf. At the end all goes to you!, also indicates tthe Supermicro AOC-SGP-I2, also uses 8086:1521. 8086:1521 is intel OEM product for i350.
This is needed for os to load appropirate kernel module/drive. EX: on linux will load igb kernel module. Click to expand.My WinYao cards look like the attached - their SUBSYS of 00001D1A doesn't appear in your list above.
I don't know if it's unique to WinYao - anyone else have a match with a none-WinYao card? There's some bad information a couple of pages back in this thread about the VendorID and the DeviceID - you should expect them to match the Intel IDs on every I350 card - genuine or fake. These two values are what your operating system uses to pick the correct driver - if a card is using an Intel I350 chip, then it can't change the VendorID or DeviceID - otherwise the drivers won't work! If you want a complete list of 'supported by Intel' IDs, then you can grab the Windows drivers from Intel (I'm using the Windows 10 drivers at ) and check out the driver.inf files. My I350 cards are using the e1r65x64.sys driver, and it's corresponding e1r65x64.inf file supports the following hardware. Note the very first line is a 'catch all' of 8086 and 1521, so the driver is going to work no matter what SUBSYS lives on your card.
My WinYao cards look like the attached - their SUBSYS of 00001D1A doesn't appear in your list above. I don't know if it's unique to WinYao - anyone else have a match with a none-WinYao card? There's some bad information a couple of pages back in this thread about the VendorID and the DeviceID - you should expect them to match the Intel IDs on every I350 card - genuine or fake. These two values are what your operating system uses to pick the correct driver - if a card is using an Intel I350 chip, then it can't change the VendorID or DeviceID - otherwise the drivers won't work!
If you want a complete list of 'supported by Intel' IDs, then you can grab the Windows drivers from Intel (I'm using the Windows 10 drivers at ) and check out the driver.inf files. My I350 cards are using the e1r65x64.sys driver, and it's corresponding e1r65x64.inf file supports the following hardware. Note the very first line is a 'catch all' of 8086 and 1521, so the driver is going to work no matter what SUBSYS lives on your card. Click to expand.pci id list can be query on the net, go to linux pci id. You can see all vendor ID: product ID: device ID based on. Device ID (DID) - 16-bit value located at offset 2. PCI device - An electrical component that conforms to the PCI Local Bus Specification, Revision 2.2.
Revision ID (REV) - 8-bit value located at offset 8. Subsystem ID (SID) - 16-bit value located at offset 2E. The SID is the ID assigned by the manufacturer to uniquely identify the device. Subsystem Vendor ID (SVID) - 16-bit value located at offset 2C. Note: The SVID is the ID assigned to the manufacturer by the PCI Special Interest Group (PCI SIG). Vendor ID (VID) - 16-bit value located at offset 0. The following is an example of a typical PnP ID section of an INF file, with VID/DID and VID/DID/SID/SVIDentries: you can not find sub vendor ID (subsystem ID in Windows): 00001D1A information.
Most driver when failed on lookup in driver installation, will land to VID/DID (microsoft way) Intel can lockdown, but they do not do. Please look on the link on below note. On linux, the kernel only care vendor and product ID to load correct kernel module I2XX/I3XX will be loaded igb kernel module Iwould assume the sameway on other UN.X variances. Note: I think, you should read that is for USB, chinese maker clone well know USB ID, and the real chipmaker put a locked down. Well at the end. The lockdown was removed due on complains.
Click to expand.haha, I bought a bunch of fake USB to serial, was scary hahah all is possible, when someone know on how to flash the hardware, many ghost factories in china and working officially on the real factory that produce intel products. One beauty working real and ghost area. Click to expand.Agreed - the Windows 10 drivers above support VEN8086&DEV1521, VEN8086&DEV1522, VEN8086&DEV1523 and VEN8086&DEV1524. The 1521 is 'Intel(R) I350 Gigabit Network Connection'. The 1522 are the Fibre cards - they're listed as 'Intel(R) I350 Gigabit Fiber Network Connection' and 'Intel(R) Ethernet Server Adapter I350-F1/I350-F2/I350-F4'.
The 1523 is listed as 'Intel(R) I350 Gigabit Backplane Connection'. And the 1524 is just 'Intel(R) I350 Gigabit Connection'.
A fake card can potentially choose any of those four DeviceIDs and still work fine with the Intel drivers. Agreed - the Windows 10 drivers above support VEN8086&DEV1521, VEN8086&DEV1522, VEN8086&DEV1523 and VEN8086&DEV1524. The 1521 is 'Intel(R) I350 Gigabit Network Connection'. The 1522 are the Fibre cards - they're listed as 'Intel(R) I350 Gigabit Fiber Network Connection' and 'Intel(R) Ethernet Server Adapter I350-F1/I350-F2/I350-F4'.
The 1523 is listed as 'Intel(R) I350 Gigabit Backplane Connection'. And the 1524 is just 'Intel(R) I350 Gigabit Connection'. A fake card can potentially choose any of those four DeviceIDs and still work fine with the Intel drivers.