Remove Hygon value from usbdevs.
- Queries
- All Stories
- Search
- Advanced Search
- Transactions
- Transaction Logs
Advanced Search
Feb 10 2020
Show more context.
Sort Hygon constants in order.
Show more context.
Show more context.
Show more context.
Show more context.
Show more context.
Feb 9 2020
Show more context.
In D23560#517197, @cem wrote:Generally please upload phabricator diffs with more context. The arcanist tool can help do that easily. Or if you prefer the manual process, please generate patches with -U99999.
Remove the redundant include.
In D23559#517201, @cem wrote:My questions on the initial review were never really answered...
The next generation Hygon core will evolve based on current architecture; there is no needed to be entirely different with AMD.
Keep the Hygon constants in upper case.
Create a new function vmm_is_svm(), and roll the AMD and Hygon check into that.
In D23564#517113, @hselasky wrote:Do you have a reference for the new PCI ID values?
Yes, please reference to:
http://pci-ids.ucw.cz/read/PC/1d94
Sort the new entry in alpha order.
Feb 8 2020
In D23556#517250, @emaste wrote:please upload with context and presumably the new entry should be sorted in alpha oder
In D23556#517181, @mav wrote:Hi. Does the Hygon really have only one AHCI ID where AMD had 5?
In D23553#517246, @emaste wrote:I agree with @grehan, vmm_is_svm is preferable rather than changing all cases to vmm_is_amd || vmm_is_hygon
Feb 7 2020
Jan 21 2020
In D23163#510442, @kib wrote:Do you want me to commit this now ?
Jan 20 2020
Or Hygon with the old condition in db_show_sysregs.
Jan 17 2020
Jan 15 2020
Jan 14 2020
In D23163#507793, @cem wrote:Is this chip something we will be able to obtain outside of China? Will hardware documentation be published in English, other than that provided by AMD for its not-quite-the-same-hardware? What plans does Hygon have going forward? Will they continue to just license future AMD designs or will there be any differentiation?
It seems that Hygon is just a licensed AMD Zen core with zero differentiation. Why not just set cpu_vendor_id to CPU_VENDOR_AMD if Hygon cpuid string is detected? That way we can avoid rototilling almost all of this code by treating hygon / family 0x18 as a weird AMD (which it is).
My concern is that this adds maintenance burden to the AMD MD codebase for an obscure chip no one actually has, uses, or has documentation for.
In D23163#507790, @imp wrote:src/stand looks good for sure. I eyeballed the rest and didn't see problems