In D51175#1168258, @rmacklem wrote:

Put .Dv before UF_HIDDEN and UF_SYSTEM as suggested
by ziaee@ and replaced va_flags with flags as suggested
by kib@. (I did not put .Ar in front of flags, since it is not
an argument to pathconf. Should I?)

Loading module which is not configured against the same kernel config as current kernel is dangerous. It does not guarantee that the KBI of module and kernel match, and then all bets are off.

The var should be added to libsys/pathconf.2

In D51060#1167398, @dougm wrote:

In D51060#1167241, @kib wrote:

The name pt_root and the term root is confusing IMO. At least it is worth adding a comment explaining the pt_root is some node, not necessary root.

I don't mind changing the name from 'root' to ... what? 'start'? 'cursor'? Any suggestions are welcome.

Do you want me to rebase the changes?

Did you considered writing a coccinelle script to do the transformation? We even have the place for it, tools/coccinelle.

Exclude errc() from extended error printing.

Add vexterr() helper.

Rework retry.
Relock sx between retries.
Check for pending signals.

Update comments.

I believe I already agreed with note 1.

Mention the owned reference in the comment.
Limit the number of restarts for sysctl handler.

Add KVA map.
Use loops instead of memcpy/memset.

Instead of exposing INP_FREED, add in_pcbrele_rlock().

The name pt_root and the term root is confusing IMO. At least it is worth adding a comment explaining the pt_root is some node, not necessary root.

I suggest to split the diff into two.
One for doing something with the signal list, if you decide to move it (or rather merge with the list in signal(3)).
And second for the .Tn removal etc (BTW why?). There is also .St macro, which might be appropriate in some places.

Beg me pardon please for the naive question.
What this revision basically does is changing the semantic of the pt_root. Instead being a real root of the trie, it becomes the starting point for search. The argument for not needed the real root is that we can always climb up to it from any node, following pn_parent. Am I right?

Why the man page added to libsys? It must go into libc/gen.

Use vm_offset_t for kva_layout members

Rebase.

Correct diff, with all 's.' eliminated.

Track D51127

Drop internal union member name.
Handle MD_PRELOAD in mddestroy().

I proposed the patch to support unmapped preloaded images in D51128, untested.
This patch would require a trivial adaptation, in particular, remove the pmap_map() line, and pass paddr as the second arg to md_preloaded():

md_preloaded(0, paddr, len, scratch, false);

Fix assert.

Properly handle freeing of the allocated pt page.

I only read the sys/kern bits.

In D37885#1166278, @des wrote:

Well then what was the point of adding vn_getsize() if it wasn't eventually going to be backed by a VOP?

It would be useful to provide a reasoning why the setting is safe.

pmap_demote_PDEP(): Add WITNESS_WARN, contain operations with m into global if.

Or, do you want me to integrate this into the series of patches for D50970?

Pass vm_page instead of &vm_page as pt to pmap_demote_pdpe().

Update KASSERT message.
Remove redundant MPASS().

This effectively allows to boot on >4Tb machines. I tried it on qemu with 8TB configured, the machine boot, but eventually qemu dies on my 64GB w/s due to OOM (really out of swap space).
The sysctl vm.vmtotal looks funny:

System wide totals computed every five seconds: (values in kilobytes)
===============================================
Processes:		(RUNQ: 1 Disk Wait: 0 Page Wait: 0 Sleep: 1)
Virtual Memory:		(Total: 4295023404K Active: 4295023404K)
Real Memory:		(Total: 6192K Active: 6192K)
Shared Virtual Memory:	(Total: 2704K Active: 2704K)
Shared Real Memory:	(Total: 2224K Active: 2224K)
Free Memory:	8402748544K

and top memory usage line cannot display the free memory size.

In D37885#1165729, @des wrote:

Hi @mjg & @kib, if the plan was to eventually add a VOP_GETSIZE(), I would suggest getting it done in the next few weeks, before we branch stable/15. Do you have something planned already or would you like me to cook up a patch?

In D50920#1165865, @jrtc27 wrote:

In D50920#1165848, @kib wrote:

Consider increasing RTLD_STATIC_TLS_EXTRA, we now use 16 bytes there unconditionally.

This is done before allocate_initial_tls so before we fix the size of the static TLS block, like any initially loaded object. RTLD_STATIC_TLS_EXTRA remains entirely free for dlopened objects.