Does it look like this picture ? Not all IDE drives like to be master-slaved to each other, what make/model HD's are you trying to use ?
I suggest one HD on primary, other on secondary and slave the cd to one of them....
Ohh and if that is like your card it is a Winbond ref design and the jumpers are silkscreened....

Are you sure your BIOS supports secondary IDE? Many VLB era ones don't. One way to know is if it refers to the disks as C: and D: rather than primary/secondary master/slave, it only supports primary, and the secondary interface would need to be supported via a DOS driver.

oh and by the way: pre-El Torito BIOSes won't detect IDE CDROM; the ATAPI spec is specifically designed to hide from pre-ATAPI IDE device probing

Yes! It is the same card, but there are no tables on the silkscreen, so thank you very much! Also the bios really only tries to auto-detect C and D drive, the CD-ROM option is not there, but it works on the secondary channel fine with a DOS driver. But still I would like to use two HDDs in the machine and putting them on the primary channel as master and slave doesn't seem to work. One of the HDDs is some HP 1GB drive, I have WFW 3.11 there with networking up and running and the other one is an old CONNER CP-2064 64 MB laptop hdd that I'am trying to back up trough this machine.

You likely need xtide xub to enable this.

An NE2000 compatible nic with a 32k bootrom socket is your ally to make this happen.

You'll likely get overall better lba and multiword io support that way anyhow.

I have this card, do you think I can solder a DIL 28 socket to those holes and use the XTIDE ROM? There is some SRAM chip on the same traces, is it possible that it's used instead of the ROM chip?

That is a 3c509b/c

Aka, 'etherlink iii'

https://en.m.wikipedia.org/wiki/3Com_3c509

And, yes, it is known to work.
XT-IDE BIOS with 3com 3c509B

You will need an (E)EPROM programmer.

Yes, but there are many versions of the pcb, this one is 3C509B-TPC and weird thing is that it has no socket for the ROM chip, the holes are filled with solder and there is a CY6264-70SNC SRAM chip next to it that shares the traces wit the ROM chip position, can it be that they used this chip instead of the EPROM and it has to be desoldered before using the custom ROM?

OK, anwser found on the wiki page, thanks! It should work...