From: Het Gala <het.gala@nutanix.com>
To: "Daniel P. Berrangé" <berrange@redhat.com>
Cc: qemu-devel@nongnu.org, prerna.saxena@nutanix.com,
quintela@redhat.com, dgilbert@redhat.com, pbonzini@redhat.com,
armbru@redhat.com, eblake@redhat.com, manish.mishra@nutanix.com,
aravind.retnakaran@nutanix.com
Subject: Re: [PATCH v4 4/8] migration: converts rdma backend to accept MigrateAddress struct
Date: Mon, 15 May 2023 20:08:57 +0530 [thread overview]
Message-ID: <ae609793-7d9b-a305-d0bf-e8e9eb59d66c@nutanix.com> (raw)
In-Reply-To: <ZGIIVc83VbEMgUhB@redhat.com>
On 15/05/23 3:54 pm, Daniel P. Berrangé wrote:
> On Fri, May 12, 2023 at 02:32:36PM +0000, Het Gala wrote:
>> RDMA based transport backend for 'migrate'/'migrate-incoming' QAPIs
>> accept new wire protocol of MigrateAddress struct.
>>
>> It is achived by parsing 'uri' string and storing migration parameters
>> required for RDMA connection into well defined InetSocketAddress struct.
>>
>> Suggested-by: Aravind Retnakaran <aravind.retnakaran@nutanix.com>
>> Signed-off-by: Het Gala <het.gala@nutanix.com>
>> ---
>> migration/migration.c | 8 ++++----
>> migration/rdma.c | 38 ++++++++++++++++----------------------
>> migration/rdma.h | 6 ++++--
>> 3 files changed, 24 insertions(+), 28 deletions(-)
>>
>> @@ -3360,10 +3346,12 @@ static int qemu_rdma_accept(RDMAContext *rdma)
>> .private_data_len = sizeof(cap),
>> };
>> RDMAContext *rdma_return_path = NULL;
>> + InetSocketAddress *isock = g_new0(InetSocketAddress, 1);
>> struct rdma_cm_event *cm_event;
>> struct ibv_context *verbs;
>> int ret = -EINVAL;
>> int idx;
>> + char arr[8];
>>
>> ret = rdma_get_cm_event(rdma->channel, &cm_event);
>> if (ret) {
>> @@ -3375,13 +3363,17 @@ static int qemu_rdma_accept(RDMAContext *rdma)
>> goto err_rdma_dest_wait;
>> }
>>
>> + isock->host = rdma->host;
>> + sprintf(arr,"%d", rdma->port);
>> + isock->port = arr;
> While Inet ports are 16-bit, and so 65535 fits in a char[8], nothing
> at the QAPI parser level is enforcing this.
>
> IOW, someone can pass QEMU a QAPI config with port = 235252353253253253232
> and casue this sprintf to smash the stack.
>
> Also this is assigning a stack variable to isock->port which
> expects a heap variable. qapi_free_InetSocketAddress() will
> call free(isock->port) which will again crash.
>
> Just do
>
> g_autoptr(InetSocketAddress) isock = g_new0(InetSocketAddress, 1);
>
> isock->port = g_strdup_printf("%d", rdma->port);
Thanks Daniel. Will change this in next version of patchset. Is a
protection for isock->host and isock->port needed here ?
>> +
>> /*
>> * initialize the RDMAContext for return path for postcopy after first
>> * connection request reached.
>> */
>> if ((migrate_postcopy() || migrate_return_path())
>> && !rdma->is_return_path) {
>> - rdma_return_path = qemu_rdma_data_init(rdma->host_port, NULL);
>> + rdma_return_path = qemu_rdma_data_init(isock, NULL);
>> if (rdma_return_path == NULL) {
>> rdma_ack_cm_event(cm_event);
>> goto err_rdma_dest_wait;
>> @@ -3506,6 +3498,8 @@ static int qemu_rdma_accept(RDMAContext *rdma)
>> err_rdma_dest_wait:
>> rdma->error_state = ret;
>> qemu_rdma_cleanup(rdma);
>> + qapi_free_InetSocketAddress(isock);
>> + g_free(arr);
> Free'ing a stack variable
Ack, will delete both statements from here.
>> g_free(rdma_return_path);
>> return ret;
>> }
>> @@ -4114,7 +4108,8 @@ static void rdma_accept_incoming_migration(void *opaque)
>> }
>> }
>>
>> -void rdma_start_incoming_migration(const char *host_port, Error **errp)
>> +void rdma_start_incoming_migration(InetSocketAddress *host_port,
>> + Error **errp)
>> {
>> int ret;
>> RDMAContext *rdma;
>> @@ -4160,13 +4155,12 @@ err:
>> error_propagate(errp, local_err);
>> if (rdma) {
>> g_free(rdma->host);
>> - g_free(rdma->host_port);
>> }
>> g_free(rdma);
>> }
>>
>> void rdma_start_outgoing_migration(void *opaque,
>> - const char *host_port, Error **errp)
>> + InetSocketAddress *host_port, Error **errp)
>> {
>> MigrationState *s = opaque;
>> RDMAContext *rdma_return_path = NULL;
>> diff --git a/migration/rdma.h b/migration/rdma.h
>> index de2ba09dc5..ee89296555 100644
>> --- a/migration/rdma.h
>> +++ b/migration/rdma.h
>> @@ -14,12 +14,14 @@
>> *
>> */
>>
>> +#include "qemu/sockets.h"
>> +
>> #ifndef QEMU_MIGRATION_RDMA_H
>> #define QEMU_MIGRATION_RDMA_H
>>
>> -void rdma_start_outgoing_migration(void *opaque, const char *host_port,
>> +void rdma_start_outgoing_migration(void *opaque, InetSocketAddress *host_port,
>> Error **errp);
>>
>> -void rdma_start_incoming_migration(const char *host_port, Error **errp);
>> +void rdma_start_incoming_migration(InetSocketAddress *host_port, Error **errp);
>>
>> #endif
>> --
>> 2.22.3
>>
> With regards,
> Daniel
Regards,
Het Gala
next prev parent reply other threads:[~2023-05-15 14:46 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-05-12 14:32 [PATCH v4 0/8] migration: Modified 'migrate' and 'migrate-incoming' QAPI commands for migration Het Gala
2023-05-12 14:32 ` [PATCH v4 1/8] migration: introduced 'MigrateAddress' in QAPI for migration wire protocol Het Gala
2023-05-15 8:37 ` Juan Quintela
2023-05-15 10:06 ` Daniel P. Berrangé
2023-05-12 14:32 ` [PATCH v4 2/8] migration: Converts uri parameter into 'MigrateAddress' struct Het Gala
2023-05-15 8:43 ` Juan Quintela
2023-05-15 10:12 ` Daniel P. Berrangé
2023-05-15 11:45 ` Het Gala
2023-05-15 11:55 ` Juan Quintela
2023-05-15 12:17 ` Daniel P. Berrangé
2023-05-15 12:25 ` Juan Quintela
2023-05-12 14:32 ` [PATCH v4 3/8] migration: converts socket backend to accept MigrateAddress struct Het Gala
2023-05-15 8:55 ` Juan Quintela
2023-05-15 10:17 ` Daniel P. Berrangé
2023-05-15 14:22 ` Het Gala
2023-05-15 14:46 ` Juan Quintela
2023-05-15 15:16 ` Het Gala
2023-05-15 16:28 ` Het Gala
2023-05-15 16:42 ` Daniel P. Berrangé
2023-05-12 14:32 ` [PATCH v4 4/8] migration: converts rdma " Het Gala
2023-05-15 9:51 ` Juan Quintela
2023-05-15 10:24 ` Daniel P. Berrangé
2023-05-15 14:38 ` Het Gala [this message]
2023-05-15 14:58 ` Daniel P. Berrangé
2023-05-15 15:17 ` Het Gala
2023-05-12 14:32 ` [PATCH v4 5/8] migration: converts exec " Het Gala
2023-05-15 9:58 ` Juan Quintela
2023-05-15 10:29 ` Daniel P. Berrangé
2023-05-15 15:04 ` Het Gala
2023-05-12 14:32 ` [PATCH v4 6/8] migration: modified 'migrate' QAPI to accept 'channels' argument for migration Het Gala
2023-05-15 10:36 ` Daniel P. Berrangé
2023-05-16 5:48 ` Het Gala
2023-05-16 8:57 ` Daniel P. Berrangé
2023-05-16 10:14 ` Het Gala
2023-05-12 14:32 ` [PATCH v4 7/8] migration: modified 'migrate-incoming' " Het Gala
2023-05-15 10:01 ` Juan Quintela
2023-05-15 10:38 ` Daniel P. Berrangé
2023-05-12 14:32 ` [PATCH v4 8/8] migration: Introduced MigrateChannelList struct to migration code flow Het Gala
2023-05-15 10:04 ` Juan Quintela
2023-05-15 10:42 ` Daniel P. Berrangé
2023-05-16 17:18 ` Het Gala
2023-05-17 8:34 ` Juan Quintela
2023-05-16 10:32 ` [PATCH v4 0/8] migration: Modified 'migrate' and 'migrate-incoming' QAPI commands for migration Markus Armbruster
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=ae609793-7d9b-a305-d0bf-e8e9eb59d66c@nutanix.com \
--to=het.gala@nutanix.com \
--cc=aravind.retnakaran@nutanix.com \
--cc=armbru@redhat.com \
--cc=berrange@redhat.com \
--cc=dgilbert@redhat.com \
--cc=eblake@redhat.com \
--cc=manish.mishra@nutanix.com \
--cc=pbonzini@redhat.com \
--cc=prerna.saxena@nutanix.com \
--cc=qemu-devel@nongnu.org \
--cc=quintela@redhat.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).