Fix DDR5 SPD misidentification as DDR4

[chenx-dust]

DDR5 hubs expose their device revision at offset 0x02, where DDR4 EEPROMs store the memory type. A revision such as 0x10 can therefore be mistaken for LPDDR4 by the DDR4-first detection path.

Extract the existing SPD5118 identity check into a shared read-only helper and use it to exclude DDR5 hubs from DDR4 detection. Verify the hub identity before DDR5 page selection to avoid treating a non-DDR5 EEPROM byte as MR11. Keep the existing detection order, retry logic, and failure logging.

Assisted by GPT 6 Astra

[Blacktempel]

Thank you for your PR.
Due to many upcoming changes all non-critical PRs for this repository will currently be put on hold.
Your change will be reviewed and considered for a patch version, when possible.

[chenx-dust]

Thank you for your PR. Due to many upcoming changes all non-critical PRs for this repository will currently be put on hold. Your change will be reviewed and considered for a patch version, when possible.

I understand. However, the bug fixed by this PR is a critical issue that incorrectly identifies DDR5 memory as DDR4, which has affected the proper operation of downstream applications, e.g. librehardwaremonitor. I hope this issue can be resolved as soon as possible. Thank you! :)