[
https://issues.apache.org/jira/browse/TS-4309?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15307321#comment-15307321
]
Oknet Xu commented on TS-4309:
------------------------------
{{buf.writer()->write_avail();}} will extended the iobufferblock chains if no
space to save data
{{SSLReadBuffer}} is used to read and decrypt data, but only received a full
{{SSL Data Record}} and then decrypt it.
And a {{SSL Data Record}} could very large and trigger READ many times.
{{SSLReadBuffer}} return {{WANT READ}} before received a full {{SSL Data
Record}}.
{{SSLReadBuffer}} could return all data of {{SSL Data Record}} when decrypt it.
the rest data keeps in {{rBIO}} if {{SSLReadBuffer}} return a parts of data
depend on the size of read buffer.
Due to the data already decrypted and saved in {{rBIO}}, {{bytes_read}} could
bigger than {{toread}} if there has enough {{write_avail}} space to save more
data in {{rBIO}}.
{code}
diff --git a/iocore/net/SSLNetVConnection.cc b/iocore/net/SSLNetVConnection.cc
index c6e0e5a..dbe082b 100644
--- a/iocore/net/SSLNetVConnection.cc
+++ b/iocore/net/SSLNetVConnection.cc
@@ -201,16 +201,12 @@ ssl_read_from_net(SSLNetVConnection *sslvc, EThread
*lthread, int64_t &ret)
bool trace = sslvc->getSSLTrace();
Debug("ssl", "trace=%s", trace ? "TRUE" : "FALSE");
- bytes_read = 0;
- while (sslErr == SSL_ERROR_NONE) {
- int64_t block_write_avail = buf.writer()->block_write_avail();
+ int64_t toread = buf.writer()->write_avail();
+
+ while (sslErr == SSL_ERROR_NONE && bytes_read < toread) {
+ block_write_avail = buf.writer()->block_write_avail();
if (block_write_avail <= 0) {
- buf.writer()->add_block();
- block_write_avail = buf.writer()->block_write_avail();
- if (block_write_avail <= 0) {
- Warning("Cannot add new block");
- break;
- }
+ break;
}
Debug("ssl", "[SSL_NetVConnection::ssl_read_from_net] b->write_avail()=%"
PRId64, block_write_avail);
{code}
> Reduced SSL upload/download speed after event loop change (TS-4260)
> -------------------------------------------------------------------
>
> Key: TS-4309
> URL: https://issues.apache.org/jira/browse/TS-4309
> Project: Traffic Server
> Issue Type: Bug
> Components: Core
> Reporter: Susan Hinrichs
> Assignee: Susan Hinrichs
> Fix For: 7.0.0
>
>
> With the Eventloop changes of TS-4260 there are fewer spurious event loop
> kicks. This reveals stalls during upload and download on a lightly loaded
> system. Tidying up the net read/write loops and reactivating a never
> triggered signal fixed the problem.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
