Hi Linus,
Could you apply this, please?
Thanks,
David
---
In the same spirit as commit ca57f02295f1 ("afs: Fix fileserver
probe RTT handling"), don't rule out using a vlserver just because
there haven't been enough packets yet to calculate a real rtt.
Always set the server's probe rtt from the estimate provided by
rxrpc_kernel_get_srtt, which is capped at 1 second.
This could lead to EDESTADDRREQ errors when accessing a cell for
the first time, even though the vl servers are known and have
responded to a probe.
Fixes: 1d4adfaf6574 ("rxrpc: Make rxrpc_kernel_get_srtt() indicate validity")
Signed-off-by: Marc Dionne <[email protected]>
Signed-off-by: David Howells <[email protected]>
cc: [email protected]
Link: http://lists.infradead.org/pipermail/linux-afs/2023-June/006746.html
---
fs/afs/vl_probe.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/afs/vl_probe.c b/fs/afs/vl_probe.c
index d1c7068b4346..58452b86e672 100644
--- a/fs/afs/vl_probe.c
+++ b/fs/afs/vl_probe.c
@@ -115,8 +115,8 @@ void afs_vlserver_probe_result(struct afs_call *call)
}
}
- if (rxrpc_kernel_get_srtt(call->net->socket, call->rxcall, &rtt_us) &&
- rtt_us < server->probe.rtt) {
+ rxrpc_kernel_get_srtt(call->net->socket, call->rxcall, &rtt_us);
+ if (rtt_us < server->probe.rtt) {
server->probe.rtt = rtt_us;
server->rtt = rtt_us;
alist->preferred = index;