User Details
- User Since
- May 16 2014, 7:29 PM (265 w, 2 d)
Fri, Jun 14
Fix _ctx_start relocation for abort() call.
Wed, Jun 12
Add @dim, for the clang part. The clang part will be upstreamed once this is all done and stable.
Address feedback, fix a crash of syncicache() too big for the PLT.
Tue, Jun 11
Mon, Jun 10
Sat, Jun 8
Fri, Jun 7
Good catch. We should probably consider adding !hw_direct_map support sometime in the future, too, for use in Qemu/KVM.
Thanks for the explanation. It looks fine either way. I just recommend making the change for consistency, but it's not required if it's not needed.
Thu, Jun 6
Overall looks good, just one question.
Sat, Jun 1
Thu, May 30
This change will probably also be needed for DPCPU. Can that be tested as well?
Use the right VPN mask.
Wed, May 29
Mon, May 27
Looks good. A trivial fix.
Sat, May 25
Fri, May 24
Thu, May 23
Wed, May 22
Looks good now!
Tue, May 21
Mon, May 20
@bdrewery knows the Makefiles much better than probably anybody else on the project.
Looks fine to me now. Good work!
This looks fine to me. Hoping we can obliterate GNU ld before too long, even on 32-bit.
Sun, May 19
Third time's a charm.
Makes sense. Should've thought of that when I made this change.
Address feedback from markj.
May 18 2019
May 17 2019
May 16 2019
May 13 2019
Will take this to completion.
May 11 2019
May 10 2019
May 8 2019
May 6 2019
May 5 2019
May 4 2019
May 3 2019
See above comment. It shouldn't be gated on LP64
May 2 2019
The 64-bit Book-E pmap overhead is around 32kB (12 bits worth of PP2D_NENTRIES, so 4096 8-byte entries), and 4kB on 32-bit. This is not small by any stretch, and I do want to rework it, and plan to over the next couple months. For now I can #ifdef within the union (I do want to keep it a union, so that we can eventually have a GENERIC that boots on both AIM and Book-E).
May 1 2019
From my count it saves 8 bytes, not 64. I don't think it's a significant issue, given the size of the structure already, but it's not a bad change, and does allow for better packing.
Apr 30 2019
Apr 29 2019
We might want to do this with other devices as well.
Apr 27 2019
Apr 26 2019
That idea sounds fine to me. After the MFC, just remove the whole aliasing thing entirely. I have a feeling the number of PowerNV users (the only ones who would see msdosfs /boot) running 12 will be very small compared to those running head.
As @rgrimes mentioned on the email thread, can the symlink just summarily be jettisoned, since it was a compat shim within 12? Or am I mistaken, and it's a compat shim in 13?
Apr 24 2019
I need to think about this more. According to the PowerISA reference, 'cmpb' is 'new' with PowerISA 2.05, which means this will break the PowerPC 970 and POWER5 and earlier. I already broke it last night with the commit of the strcmp optimization, so don't want to break it further now, until we better reason this.