♟ rmacklem

Fix build issue introduced by r361699.

Looks fine to me, too.
I assume this fixes your crashes, good work.

Oops, I didn't notice the "cd" is needed for each install line.

Update the Makefile to copy rpcsec_tls.h to /usr/include/rpc.

Update rpcsec_tls.h to the one just added to head/current.

Add the .h file that describes the operations for the rpctls_syscall.

Oops two, missed syscall.mk as well.

Oops, missed syscall.h and sysproto.h for r361602.

Add an entry to Symbol.map for the rpctls_syscall added by r361599.

Update the files created from the new syscalls.master from r361599.

Add a syscall for the nfs-over-tls daemons to use.

Modified compat/freebsd32/syscalls.master entry
as suggested by brooks@.
It worked and did not create a proto entry.

Fix sosend() for the case where mbufs are passed in while doing ktls.

To enable building a kernel without having the rest of /usr/src around,

Sorry, I know nothing about this stuff, so I can't review it.
(This code has been written by others, possibly long ago.)

Fix a few problems caused during the code merge of the up-to-date head kernel.

Fix one case of an ext_pgs mbuf change that I missed.

Merge in an up-to-date kernel from head and make the changes

Oops, forgot to "make sysent" before testing the last one.
Since the entry on compat/freebsd32/syscalls.master is NOSTD,
I needed to prepend "freebsd_" to the prototype name to
avoid a conflict.

Removed the blank line from syscalls.master and
added the definition to compat/freebsd32/syscalls.master,
as suggested by brooks@.

Fix the daemons so that they use the preferred calls for openssl3

Modify the client side non-application data record upcall so that it is

Fix the daemons so that they do not terminate upon SIGPIPE.

Add a function nfsm_set() to initialize "struct nfsrv_descript" for building

Assorted error handling fixups plus a fix for client side handling of

Separate errors doing the rpctlscd/rpctlssd upcalls from RPC errors.

Make nfsm_set() a void function, since the return value is no longer used.

Add support for doing upcalls to the rpctlscd daemon to the krpc client

Add fdrop()s to the rpctls syscall for the "fp" reference.

Add code that does upcalls to the daemons when an non-application data record

Get rid of the changes done to the krpc and xdr for dissecting ext_pgs mbufs.

Get rid of two fields that were only used for dissecting ext_pgs mbufs.

Sure. Actually I thought of mentioning this when you got
rid of the #ifdef APPLE junk, but figured you'd be tired of
large but trivial edits.

Revert one more change that was done to handle dissection of ext_pgs mbufs.

Get rid of some more of the dissect code for ext_pgs mbufs.

Comments inline. Mostly questions.

Simplify nfsm_set() so that it does not do the dissect case.

Revert some of the code for dissecting the ext_pgs mbufs, since it is

After you have updated your patch, I'll look again.
Thanks for doing this, rick

Update the nfs-over-tls-setup.txt document slightly.

Modify ktls_decrypt() slightly, so that a small iov is allocated on

Add some very basic handling of TLS_GET_RECORD control mbufs.

Add some very basic handling of TLS_GET_RECORD control mbufs for the

Make both rpctlssd and rpctlscd load the ktls_ocf module, as required.

Update krpc.h as needed for the new clnt_vc_soupcall() just committed.

Rewrite clnt_vc_soupcall() so that it does soreceive() for an

Yes, they are useless.
(As an historic note, long ago there was a Mac OS/X port, where the code
was built as a module and these #ifdef's were used for that.)

Update nfs-over-tls-setup.txt.

Get rid of MYTEST and modify GENERIC, so that builds are easier for testers.

Do not do SSL_shutdown() for now, since it can cause crashes.

Delete unused function newnfs_trimleading.

I just made another inline comment. It looks like SOCKBUF_LOCK(sb) gets called
twice in the if (error) { } block?

I haven't looked at the patch in detail (and I'm far from an expert in this
area). However, the overall approach sounds fine to me and I have only
found the one little problem (leak of malloc'd "iov" in ktls_decrypt())
during testing sofar. (See inline comment above.)

Revert r360514, to avoid unnecessary churn of the sources.

Put a copy of nfs-over-tls-setup.txt here.

Fix a leak in ktls_decrypt().

Replace openssl3.patch with a slightly cleaner one.

Add a little function to acquire the ktls sysctls.

Fix handling of ktls not enabled in the daemons.

Add the patch file for /usr/ktls/include/openssl.

Get rid of mountd subdirectory that I somehow created.

Clear out all the cruft I created under usr.sbin, so that only the

Disable the code that copied received mbufs to ext_pgs mbufs.

Add the kernel config I use, which includes "options KERN_TLS", plus

Add jhb@'s ktls rx patch to the kernel.

Fix the daemons so they actually work with jhb@'s patched openssl3.

Modify nfsm_copyfrommbuf() so that it calls nfsm_copfrommbuf_extpgs() for

Factor some code out of nfsm_dissct() into separate functions.

Merge in an up to date kernel from head.

Merge in the recent change to replace the Mac OS/X uio_XX macros from head.

Get rid of uio_XXX macros used for the Mac OS/X port.

Merge in an up to date kernel from head.

Fix sosend_generic() so that it can handle a list of ext_pgs mbufs.

Fix the krpc.ko build.

Remove Mac OS/X macros that did nothing for FreeBSD.

Make the code work with post-r359919 ext_pgs mbufs.

Fix the updated kernel so that it will build post-r359919.

Merge in an up to date kernel from head.

Make the NFSv4.n client's recovery from NFSERR_BADSESSION RFC5661 conformant.

Change the type of "len" to avoid warnings.

Replace all occurrences of mbuf_t with "struct mbuf *".

Replace all instances of the typedef mbuf_t with "struct mbuf *".

Add a sanity check for nes_numsecflavor to the NFS server.

This looks fine to me.

rmacklem (Rick Macklem)User

Projects

User Details

Recent ActivityView All

Yesterday

Mon, Jun 1

Sun, May 31

Fri, May 29

Thu, May 28

Wed, May 27

Mon, May 25

Sun, May 24

Sat, May 23

Thu, May 21

Wed, May 20

Tue, May 19

Mon, May 18

Sun, May 17

Sat, May 16

Thu, May 14

Wed, May 13

Tue, May 12

Mon, May 11

Sun, May 10

Sat, May 9

Fri, May 8

Wed, May 6

Tue, May 5

Mon, May 4

May 3 2020

May 1 2020

Apr 30 2020

Apr 28 2020

Apr 27 2020

Apr 25 2020

Apr 24 2020

Apr 23 2020

Apr 22 2020

Apr 21 2020

Apr 20 2020

Apr 18 2020

Apr 17 2020

rmacklem (Rick Macklem)
User

Recent Activity
View All