Review Request for JMC-6621 - Handle event size 0
Erik Gahlin
erik.gahlin at oracle.com
Thu Oct 31 14:44:22 UTC 2019
Thanks for the info. I will look into it.
Erik
> On 31 Oct 2019, at 15:40, Henrik Dafgård <hdafgard at gmail.com> wrote:
>
> Hi Erik,
>
> Using "jfr print" on a recording with a zero event size works as expected, but running "jfr summary" on it doesn't produce a result. It seems to also be stuck in an infinite loop.
>
>
> Cheers,
> Henrik Dafgård
>
>
> On Thu, 31 Oct 2019 at 15:29, Erik Gahlin <erik.gahlin at oracle.com <mailto:erik.gahlin at oracle.com>> wrote:
> Looks good.
>
> The JDK parser checks for a zero size, so the JMC parser should do the same,
>
> Erik
>
> > On 31 Oct 2019, at 15:05, Henrik Dafgård <hdafgard at gmail.com <mailto:hdafgard at gmail.com>> wrote:
> >
> > Hi all,
> >
> > Should JMC attempt to parse a recording with an event size of 0 (this
> > should not be possible) it will currently enter an infinite loop. The patch
> > outlined below simply adds a check for size 0 and then throws a
> > CouldNotLoadRecordingException if that is the case.
> >
> > diff -r b153474a15a6
> > core/org.openjdk.jmc.flightrecorder/src/main/java/org/openjdk/jmc/flightrecorder/internal/parser/v1/ChunkLoaderV1.java
> > ---
> > a/core/org.openjdk.jmc.flightrecorder/src/main/java/org/openjdk/jmc/flightrecorder/internal/parser/v1/ChunkLoaderV1.java
> > Fri Oct 25 15:48:20 2019 +0200
> > +++
> > b/core/org.openjdk.jmc.flightrecorder/src/main/java/org/openjdk/jmc/flightrecorder/internal/parser/v1/ChunkLoaderV1.java
> > Thu Oct 31 15:00:27 2019 +0100
> > @@ -84,6 +84,9 @@
> > input.seek(index);
> > int size = input.readInt();
> > long type = input.readLong();
> > + if (size == 0) {
> > + throw new CouldNotLoadRecordingException("Found event with invalid size
> > (0)");
> > + }
> > if (type != CONSTANT_POOL_EVENT_TYPE && type !=
> > ChunkMetadata.METADATA_EVENT_TYPE) {
> > manager.readEvent(type, input);
> > }
> >
> >
> > Cheers,
> > Henrik Dafgård
>
More information about the jmc-dev
mailing list