]> git.baikalelectronics.ru Git - kernel.git/commit
SUNRPC: Revert c650c6eb6f6f ("SUNRPC: Remove xdr_buf_trim()")
authorChuck Lever <chuck.lever@oracle.com>
Wed, 15 Apr 2020 21:36:22 +0000 (17:36 -0400)
committerChuck Lever <chuck.lever@oracle.com>
Mon, 27 Apr 2020 14:58:30 +0000 (10:58 -0400)
commitcf0cc5b509edb7ac9331f4faf06191f18c046bd6
tree5d970e9546b444a329fc41b466ceb0d31ebf4240
parent1dfc2243a44a91f2af3857cbb7f9942b44f4139b
SUNRPC: Revert c650c6eb6f6f ("SUNRPC: Remove xdr_buf_trim()")

I've noticed that when krb5i or krb5p security is in use,
retransmitted requests are missing the server's duplicate reply
cache. The computed checksum on the retransmitted request does not
match the cached checksum, resulting in the server performing the
retransmitted request again instead of returning the cached reply.

The assumptions made when removing xdr_buf_trim() were not correct.
In the send paths, the upper layer has already set the segment
lengths correctly, and shorting the buffer's content is simply a
matter of reducing buf->len.

xdr_buf_trim() is the right answer in the receive/unwrap path on
both the client and the server. The buffer segment lengths have to
be shortened one-by-one.

On the server side in particular, head.iov_len needs to be updated
correctly to enable nfsd_cache_csum() to work correctly. The simple
buf->len computation doesn't do that, and that results in
checksumming stale data in the buffer.

The problem isn't noticed until there's significant instability of
the RPC transport. At that point, the reliability of retransmit
detection on the server becomes crucial.

Fixes: c650c6eb6f6f ("SUNRPC: Remove xdr_buf_trim()")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
include/linux/sunrpc/xdr.h
net/sunrpc/auth_gss/gss_krb5_wrap.c
net/sunrpc/auth_gss/svcauth_gss.c
net/sunrpc/xdr.c