migration/rdma: Enable use of g_autoptr with struct rdma_cm_event
cover migration/rdma: Enable use of g_autoptr with struct rdma_cm_event
Use g_autoptr to free rdma_cm_event to avoid missing to use the not so obvious rdma_ack_cm_event() instead:
https://lists.gnu.org/archive/html/qemu-devel/2021-06/msg00511.html
Li Zhijian (1):
migration/rdma: Fix cm event use after free
Philippe Mathieu-Daudé (1):
migration/rdma: Enable use of g_autoptr with struct rdma_cm_event
migration/rdma.c | 33 ++++++++++++++------------------- 1 file changed, 14 insertions(+), 19 deletions(-)
Signature · 1 line
2.26.3
PATCH 1/2 migration/rdma: Fix cm event use after free
From: Li Zhijian <[email protected]>
Message-Id: <[email protected]>
--- migration/rdma.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-)
migration/rdma.c +4 -2
@@ -1539,8 +1539,10 @@ static int qemu_rdma_wait_comp_channel(RDMAContext *rdma)if (pfds[1 ].revents) { ret =rdma_get_cm_event (rdma->channel, &cm_event); -if (!ret) { -rdma_ack_cm_event (cm_event); +if (ret) { +error_report ("failed to get cm event while wait " +"completion channel" ); +return -EPIPE ; } error_report("receive cm event while wait comp channel,"
Signature · 1 line
2.26.3
RFC PATCH 2/2 migration/rdma: Enable use of g_autoptr with struct rdma_cm_event
In reply to Philippe Mathieu-DaudéSince 00f2cfbbec6 ("glib: bump min required glib library version to 2.48") we can use g_auto/g_autoptr to have the compiler automatically free an allocated variable when it goes out of scope, removing this burden on the developers.
Per rdma_cm(7) and rdma_ack_cm_event(3) man pages:
"rdma_ack_cm_event() - Free a communication event.
All events which are allocated by rdma_get_cm_event() must be released, there should be a one-to-one correspondence between successful gets and acks. This call frees the event structure and any memory that it references."
Since the 'ack' description doesn't explicit the event is also released (free'd), it is safer to use the GLib g_autoptr feature. The G_DEFINE_AUTOPTR_CLEANUP_FUNC() macro expects a single word for the type name, so add a type definition to achieve this. Convert to use g_autoptr and remove the rdma_ack_cm_event() calls.
---
RFC: build-tested only
--- migration/rdma.c | 27 ++++++++++----------------- 1 file changed, 10 insertions(+), 17 deletions(-)
migration/rdma.c +10 -17
@@ -38,6 +38,9 @@#include "qom/object.h" #include <poll.h> +typedef struct rdma_cm_event rdma_cm_event ; +G_DEFINE_AUTOPTR_CLEANUP_FUNC (rdma_cm_event, rdma_ack_cm_event) +/* * Print and error on both the Monitor and the Log file. */
@@ -939,7 +942,7 @@ static int qemu_rdma_resolve_host(RDMAContext *rdma, Error **errp)int ret;struct rdma_addrinfo *res;char port_str[16 ]; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;char ip[40 ] ="unknown" ;struct rdma_addrinfo *e;
@@ -1007,11 +1010,11 @@ route:ERROR (errp,"result not equal to event_addr_resolved %s" ,rdma_event_str (cm_event->event));perror ("rdma_resolve_addr" ); -rdma_ack_cm_event (cm_event); ret = -EINVAL ; goto err_resolve_get_addr; }rdma_ack_cm_event (cm_event); + cm_event =NULL ;/* resolve route */ ret =rdma_resolve_route (rdma->cm_id,RDMA_RESOLVE_TIMEOUT_MS );
@@ -1028,11 +1031,9 @@ route:if (cm_event->event !=RDMA_CM_EVENT_ROUTE_RESOLVED ) {ERROR (errp,"result not equal to event_route_resolved: %s" ,rdma_event_str (cm_event->event)); -rdma_ack_cm_event (cm_event); ret = -EINVAL ; goto err_resolve_get_addr; } -rdma_ack_cm_event (cm_event); rdma->verbs = rdma->cm_id->verbs;qemu_rdma_dump_id ("source_resolve_host" , rdma->cm_id->verbs);qemu_rdma_dump_gid ("source_resolve_host" , rdma->cm_id);
@@ -1501,7 +1502,7 @@ static uint64_t qemu_rdma_poll(RDMAContext *rdma, uint64_t *wr_id_out,*/ static int qemu_rdma_wait_comp_channel (RDMAContext *rdma) { -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;int ret =-1 ;/*
@@ -2503,7 +2504,7 @@ static int qemu_rdma_connect(RDMAContext *rdma, Error **errp, bool return_path) .private_data = &cap, .private_data_len = sizeof(cap), }; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;int ret;/*
@@ -2544,7 +2545,6 @@ static int qemu_rdma_connect(RDMAContext *rdma, Error **errp, bool return_path)if (cm_event->event !=RDMA_CM_EVENT_ESTABLISHED ) {perror ("rdma_get_cm_event != EVENT_ESTABLISHED after rdma_connect" );ERROR (errp,"connecting to destination!" ); -rdma_ack_cm_event (cm_event); goto err_rdma_source_connect; } rdma->connected = true;
@@ -2564,8 +2564,6 @@ static int qemu_rdma_connect(RDMAContext *rdma, Error **errp, bool return_path)trace_qemu_rdma_connect_pin_all_outcome (rdma->pin_all); -rdma_ack_cm_event (cm_event); - rdma->control_ready_expected =1 ; rdma->nb_sent =0 ;return 0 ;
@@ -3279,7 +3277,7 @@ static void rdma_cm_poll_handler(void *opaque) {RDMAContext *rdma = opaque;int ret; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;MigrationIncomingState *mis =migration_incoming_get_current (); ret =rdma_get_cm_event (rdma->channel, &cm_event);
@@ -3287,7 +3285,6 @@ static void rdma_cm_poll_handler(void *opaque)error_report ("get_cm_event failed %d" , errno);return ; } -rdma_ack_cm_event (cm_event);if (cm_event->event ==RDMA_CM_EVENT_DISCONNECTED || cm_event->event ==RDMA_CM_EVENT_DEVICE_REMOVAL ) {
@@ -3317,7 +3314,7 @@ static int qemu_rdma_accept(RDMAContext *rdma) .private_data_len = sizeof(cap), };RDMAContext *rdma_return_path =NULL ; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;struct ibv_context *verbs;int ret = -EINVAL ;int idx;
@@ -3328,7 +3325,6 @@ static int qemu_rdma_accept(RDMAContext *rdma) }if (cm_event->event !=RDMA_CM_EVENT_CONNECT_REQUEST ) { -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; }
@@ -3339,7 +3335,6 @@ static int qemu_rdma_accept(RDMAContext *rdma)if (migrate_postcopy () && !rdma->is_return_path) { rdma_return_path =qemu_rdma_data_init (rdma->host_port,NULL );if (rdma_return_path ==NULL ) { -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; }
@@ -3353,7 +3348,6 @@ static int qemu_rdma_accept(RDMAContext *rdma)if (cap.version <1 || cap.version >RDMA_CONTROL_VERSION_CURRENT ) {error_report ("Unknown source RDMA version: %d, bailing..." , cap.version); -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; }
@@ -3374,6 +3368,7 @@ static int qemu_rdma_accept(RDMAContext *rdma) verbs = cm_event->id->verbs;rdma_ack_cm_event (cm_event); + cm_event =NULL ;trace_qemu_rdma_accept_pin_state (rdma->pin_all);
@@ -3441,11 +3436,9 @@ static int qemu_rdma_accept(RDMAContext *rdma)if (cm_event->event !=RDMA_CM_EVENT_ESTABLISHED ) {error_report ("rdma_accept not event established" ); -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; } -rdma_ack_cm_event (cm_event); rdma->connected = true; ret =qemu_rdma_post_recv_control (rdma,RDMA_WRID_READY );
Signature · 1 line
2.26.3
Quoted 3 lines · Philippe Mathieu-Daudé
Since
00f2cfbbec6("glib: bump min required glib library version to 2.48") we can use g_auto/g_autoptr to have the compiler automatically free an allocated variable when it goes out of scope,
Glad to know this feature.
However per its code, a 'ack' does much more than just free the memory. not sure g_autoptr have the ability to do the same.
2212 static void ucma_complete_event(struct cma_id_private *id_priv) 2213 {
2214 pthread_mutex_lock(&id_priv->mut);
2215 id_priv->events_completed++;
2216 pthread_cond_signal(&id_priv->cond);
2217 pthread_mutex_unlock(&id_priv->mut);
2218 }
2219
2220 static void ucma_complete_mc_event(struct cma_multicast *mc) 2221 {
2222 pthread_mutex_lock(&mc->id_priv->mut);
2223 mc->events_completed++;
2224 pthread_cond_signal(&mc->cond);
2225 mc->id_priv->events_completed++;
2226 pthread_cond_signal(&mc->id_priv->cond);
2227 pthread_mutex_unlock(&mc->id_priv->mut);
2228 }
2229
2230 int rdma_ack_cm_event(struct rdma_cm_event *event) 2231 {
2232 struct cma_event *evt;
2233
2234 if (!event)
2235 return ERR(EINVAL);
2236
2237 evt = container_of(event, struct cma_event, event); 2238
2239 if (evt->mc)
2240 ucma_complete_mc_event(evt->mc);
2241 else
2242 ucma_complete_event(evt->id_priv);
2243 free(evt);
2244 return 0;
2245 }
Thanks
Zhijian
Quoted 179 lines
removing this
burden on the developers.Per
rdma_cm(7)andrdma_ack_cm_event(3)man pages:"
rdma_ack_cm_event()- Free a communication event.All events which are allocated by
rdma_get_cm_event()must be released, there should be a one-to-one correspondence between successful gets and acks. This call frees the event structure and any memory that it references."Since the 'ack' description doesn't explicit the event is also released (free'd), it is safer to use the GLib
g_autoptrfeature. TheG_DEFINE_AUTOPTR_CLEANUP_FUNC()macro expects a single word for the type name, so add a type definition to achieve this. Convert to useg_autoptrand remove therdma_ack_cm_event()calls.Inspired-by Li Zhijian <[email protected]>Signed-off-by Philippe Mathieu-Daudé <[email protected]>---
RFC: build-tested only--- migration/rdma.c | 27 ++++++++++----------------- 1 file changed, 10 insertions(+), 17 deletions(-)migration/rdma.c +10
-17@@ -38,6 +38,9 @@#include "qom/object.h" #include <poll.h> +typedef struct rdma_cm_event rdma_cm_event ; +G_DEFINE_AUTOPTR_CLEANUP_FUNC (rdma_cm_event, rdma_ack_cm_event) +/* * Print and error on both the Monitor and the Log file. */ @@ -939,7 +942,7 @@ static int qemu_rdma_resolve_host(RDMAContext *rdma, Error **errp)int ret;struct rdma_addrinfo *res;char port_str[16 ]; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;char ip[40 ] ="unknown" ;struct rdma_addrinfo *e;@@ -1007,11 +1010,11 @@ route:ERROR (errp,"result not equal to event_addr_resolved %s" ,rdma_event_str (cm_event->event));perror ("rdma_resolve_addr" ); -rdma_ack_cm_event (cm_event); ret = -EINVAL ; goto err_resolve_get_addr; }rdma_ack_cm_event (cm_event); + cm_event =NULL ;/* resolve route */ ret =rdma_resolve_route (rdma->cm_id,RDMA_RESOLVE_TIMEOUT_MS );@@ -1028,11 +1031,9 @@ route:if (cm_event->event !=RDMA_CM_EVENT_ROUTE_RESOLVED ) {ERROR (errp,"result not equal to event_route_resolved: %s" ,rdma_event_str (cm_event->event)); -rdma_ack_cm_event (cm_event); ret = -EINVAL ; goto err_resolve_get_addr; } -rdma_ack_cm_event (cm_event); rdma->verbs = rdma->cm_id->verbs;qemu_rdma_dump_id ("source_resolve_host" , rdma->cm_id->verbs);qemu_rdma_dump_gid ("source_resolve_host" , rdma->cm_id);@@ -1501,7 +1502,7 @@ static uint64_t qemu_rdma_poll(RDMAContext *rdma, uint64_t *wr_id_out,*/ static int qemu_rdma_wait_comp_channel (RDMAContext *rdma) { -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;int ret =-1 ;/* @@ -2503,7 +2504,7 @@ static int qemu_rdma_connect(RDMAContext *rdma, Error **errp, bool return_path) .private_data = &cap, .private_data_len = sizeof(cap), }; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;int ret;/* @@ -2544,7 +2545,6 @@ static int qemu_rdma_connect(RDMAContext *rdma, Error **errp, bool return_path)if (cm_event->event !=RDMA_CM_EVENT_ESTABLISHED ) {perror ("rdma_get_cm_event != EVENT_ESTABLISHED after rdma_connect" );ERROR (errp,"connecting to destination!" ); -rdma_ack_cm_event (cm_event); goto err_rdma_source_connect; } rdma->connected = true;@@ -2564,8 +2564,6 @@ static int qemu_rdma_connect(RDMAContext *rdma, Error **errp, bool return_path)trace_qemu_rdma_connect_pin_all_outcome (rdma->pin_all); -rdma_ack_cm_event (cm_event); - rdma->control_ready_expected =1 ; rdma->nb_sent =0 ;return 0 ;@@ -3279,7 +3277,7 @@ static void rdma_cm_poll_handler(void *opaque) {RDMAContext *rdma = opaque;int ret; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;MigrationIncomingState *mis =migration_incoming_get_current (); ret =rdma_get_cm_event (rdma->channel, &cm_event);@@ -3287,7 +3285,6 @@ static void rdma_cm_poll_handler(void *opaque)error_report ("get_cm_event failed %d" , errno);return ; } -rdma_ack_cm_event (cm_event);if (cm_event->event ==RDMA_CM_EVENT_DISCONNECTED || cm_event->event ==RDMA_CM_EVENT_DEVICE_REMOVAL ) {@@ -3317,7 +3314,7 @@ static int qemu_rdma_accept(RDMAContext *rdma) .private_data_len = sizeof(cap), };RDMAContext *rdma_return_path =NULL ; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;struct ibv_context *verbs;int ret = -EINVAL ;int idx;@@ -3328,7 +3325,6 @@ static int qemu_rdma_accept(RDMAContext *rdma) }if (cm_event->event !=RDMA_CM_EVENT_CONNECT_REQUEST ) { -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; }@@ -3339,7 +3335,6 @@ static int qemu_rdma_accept(RDMAContext *rdma)if (migrate_postcopy () && !rdma->is_return_path) { rdma_return_path =qemu_rdma_data_init (rdma->host_port,NULL );if (rdma_return_path ==NULL ) { -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; }@@ -3353,7 +3348,6 @@ static int qemu_rdma_accept(RDMAContext *rdma)if (cap.version <1 || cap.version >RDMA_CONTROL_VERSION_CURRENT ) {error_report ("Unknown source RDMA version: %d, bailing..." , cap.version); -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; }@@ -3374,6 +3368,7 @@ static int qemu_rdma_accept(RDMAContext *rdma) verbs = cm_event->id->verbs;rdma_ack_cm_event (cm_event); + cm_event =NULL ;trace_qemu_rdma_accept_pin_state (rdma->pin_all);@@ -3441,11 +3436,9 @@ static int qemu_rdma_accept(RDMAContext *rdma)if (cm_event->event !=RDMA_CM_EVENT_ESTABLISHED ) {error_report ("rdma_accept not event established" ); -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; } -rdma_ack_cm_event (cm_event); rdma->connected = true; ret =qemu_rdma_post_recv_control (rdma,RDMA_WRID_READY );
Quoted 10 lines · On 6/3/21 3:34 AM, [email protected] wrote:
Quoted 3 lines · Philippe Mathieu-Daudé
Since
00f2cfbbec6("glib: bump min required glib library version to 2.48") we can use g_auto/g_autoptr to have the compiler automatically free an allocated variable when it goes out of scope,Glad to know this feature.
However per its code, a 'ack' does much more than just free the memory. not sure
g_autoptrhave the ability to do the same.
Defines the appropriate cleanup function for a pointer type.
The function will not be called if the variable to be cleaned up contains NULL.
This will typically be the _free() or _unref() function for the given type.
This does not change the code to call free(ptr), but to call the registered cleanup function, which is rdma_ack_cm_event().
Quoted 96 lines
2212 static void
ucma_complete_event(struct cma_id_private *id_priv)2213 {
2214pthread_mutex_lock(&id_priv->mut);
2215 id_priv->events_completed++;
2216pthread_cond_signal(&id_priv->cond);
2217pthread_mutex_unlock(&id_priv->mut);
2218 }
2219
2220 static voiducma_complete_mc_event(struct cma_multicast *mc)2221 {
2222pthread_mutex_lock(&mc->id_priv->mut);
2223 mc->events_completed++;
2224pthread_cond_signal(&mc->cond);
2225 mc->id_priv->events_completed++;
2226pthread_cond_signal(&mc->id_priv->cond);
2227pthread_mutex_unlock(&mc->id_priv->mut);
2228 }
2229
2230 intrdma_ack_cm_event(struct rdma_cm_event *event)2231 {
2232 structcma_event*evt;
2233
2234 if (!event)
2235 return ERR(EINVAL);
2236
2237 evt =container_of(event, struct cma_event, event); 2238
2239 if (evt->mc)
2240ucma_complete_mc_event(evt->mc);
2241 else
2242ucma_complete_event(evt->id_priv);
2243 free(evt);
2244 return 0;
2245 }Thanks
ZhijianQuoted 57 lines
removing this
burden on the developers.Per
rdma_cm(7)andrdma_ack_cm_event(3)man pages:"
rdma_ack_cm_event()- Free a communication event.All events which are allocated by
rdma_get_cm_event()must be released, there should be a one-to-one correspondence between successful gets and acks. This call frees the event structure and any memory that it references."Since the 'ack' description doesn't explicit the event is also released (free'd), it is safer to use the GLib
g_autoptrfeature. TheG_DEFINE_AUTOPTR_CLEANUP_FUNC()macro expects a single word for the type name, so add a type definition to achieve this. Convert to useg_autoptrand remove therdma_ack_cm_event()calls.Inspired-by Li Zhijian <[email protected]>Signed-off-by Philippe Mathieu-Daudé <[email protected]>---
RFC: build-tested only--- migration/rdma.c | 27 ++++++++++----------------- 1 file changed, 10 insertions(+), 17 deletions(-)migration/rdma.c +4
-2@@ -38,6 +38,9 @@#include "qom/object.h" #include <poll.h> +typedef struct rdma_cm_event rdma_cm_event ; +G_DEFINE_AUTOPTR_CLEANUP_FUNC (rdma_cm_event, rdma_ack_cm_event) +/* * Print and error on both the Monitor and the Log file. */ @@ -939,7 +942,7 @@ static int qemu_rdma_resolve_host(RDMAContext *rdma, Error **errp)int ret;struct rdma_addrinfo *res;char port_str[16 ]; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;char ip[40 ] ="unknown" ;struct rdma_addrinfo *e;@@ -1007,11 +1010,11 @@ route:ERROR (errp,"result not equal to event_addr_resolved %s" ,rdma_event_str (cm_event->event));perror ("rdma_resolve_addr" ); -rdma_ack_cm_event (cm_event); ret = -EINVAL ; goto err_resolve_get_addr; }
[email protected]
Tested-by Li ZhijianQuoted 15 lines · Philippe Mathieu-Daudé
Quoted 9 lines · On 6/3/21 3:34 AM, [email protected] wrote:
Quoted 3 lines · Philippe Mathieu-Daudé
Since
00f2cfbbec6("glib: bump min required glib library version to 2.48") we can use g_auto/g_autoptr to have the compiler automatically free an allocated variable when it goes out of scope,Glad to know this feature.
However per its code, a 'ack' does much more than just free the memory. not sure
g_autoptrhave the ability to do the same.Defines the appropriate cleanup function for a pointer type.
Quoted 8 lines
The function will not be called if the variable to be cleaned up contains NULL.
This will typically be the
_free()or_unref()function for the given type.This does not change the code to call free(ptr), but to call the registered cleanup function, which is
rdma_ack_cm_event().
*
Thanks for your explanation.
*
Quoted 96 lines
Quoted 96 lines
2212 static void
ucma_complete_event(struct cma_id_private *id_priv)2213 {
2214pthread_mutex_lock(&id_priv->mut);
2215 id_priv->events_completed++;
2216pthread_cond_signal(&id_priv->cond);
2217pthread_mutex_unlock(&id_priv->mut);
2218 }
2219
2220 static voiducma_complete_mc_event(struct cma_multicast *mc)2221 {
2222pthread_mutex_lock(&mc->id_priv->mut);
2223 mc->events_completed++;
2224pthread_cond_signal(&mc->cond);
2225 mc->id_priv->events_completed++;
2226pthread_cond_signal(&mc->id_priv->cond);
2227pthread_mutex_unlock(&mc->id_priv->mut);
2228 }
2229
2230 intrdma_ack_cm_event(struct rdma_cm_event *event)2231 {
2232 structcma_event*evt;
2233
2234 if (!event)
2235 return ERR(EINVAL);
2236
2237 evt =container_of(event, struct cma_event, event); 2238
2239 if (evt->mc)
2240ucma_complete_mc_event(evt->mc);
2241 else
2242ucma_complete_event(evt->id_priv);
2243 free(evt);
2244 return 0;
2245 }Thanks
ZhijianQuoted 57 lines
removing this
burden on the developers.Per
rdma_cm(7)andrdma_ack_cm_event(3)man pages:"
rdma_ack_cm_event()- Free a communication event.All events which are allocated by
rdma_get_cm_event()must be released, there should be a one-to-one correspondence between successful gets and acks. This call frees the event structure and any memory that it references."Since the 'ack' description doesn't explicit the event is also released (free'd), it is safer to use the GLib
g_autoptrfeature. TheG_DEFINE_AUTOPTR_CLEANUP_FUNC()macro expects a single word for the type name, so add a type definition to achieve this. Convert to useg_autoptrand remove therdma_ack_cm_event()calls.Inspired-by Li Zhijian <[email protected]>Signed-off-by Philippe Mathieu-Daudé <[email protected]>---
RFC: build-tested only--- migration/rdma.c | 27 ++++++++++----------------- 1 file changed, 10 insertions(+), 17 deletions(-)migration/rdma.c +4
-2@@ -38,6 +38,9 @@#include "qom/object.h" #include <poll.h> +typedef struct rdma_cm_event rdma_cm_event ; +G_DEFINE_AUTOPTR_CLEANUP_FUNC (rdma_cm_event, rdma_ack_cm_event) +/* * Print and error on both the Monitor and the Log file. */ @@ -939,7 +942,7 @@ static int qemu_rdma_resolve_host(RDMAContext *rdma, Error **errp)int ret;struct rdma_addrinfo *res;char port_str[16 ]; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;char ip[40 ] ="unknown" ;struct rdma_addrinfo *e;@@ -1007,11 +1010,11 @@ route:ERROR (errp,"result not equal to event_addr_resolved %s" ,rdma_event_str (cm_event->event));perror ("rdma_resolve_addr" ); -rdma_ack_cm_event (cm_event); ret = -EINVAL ; goto err_resolve_get_addr; }
Quoted 14 lines · * Philippe Mathieu-Daudé ([email protected]) wrote:
Since
00f2cfbbec6("glib: bump min required glib library version to 2.48") we can use g_auto/g_autoptr to have the compiler automatically free an allocated variable when it goes out of scope, removing this burden on the developers.Per
rdma_cm(7)andrdma_ack_cm_event(3)man pages:"
rdma_ack_cm_event()- Free a communication event.All events which are allocated by
rdma_get_cm_event()must be released, there should be a one-to-one correspondence between successful gets and acks. This call frees the event structure and any memory that it references."
Quoted 8 lines
Since the 'ack' description doesn't explicit the event is also released (free'd), it is safer to use the GLib
g_autoptrfeature. TheG_DEFINE_AUTOPTR_CLEANUP_FUNC()macro expects a single word for the type name, so add a type definition to achieve this. Convert to useg_autoptrand remove therdma_ack_cm_event()calls.Inspired-by Li Zhijian <[email protected]>Signed-off-by Philippe Mathieu-Daudé <[email protected]>
It's unclear to me whether the changes in qemu_rdma_accept are legal rdma or not.
If we look at the err_rdma_dest_wait: path, it does a qemu_rdma_cleanup(rdma) before the exit; that does rdma_disconnect's and calls loads of other rdma/ibv cleanup calls to destroy state - is there any ordering requirement saying you're supposed to clean up your cm_event's before you nuke the rest of the channel? It feels like you probably should - but I have no idea if it's a requirement.
Dave
Quoted 162 lines
---
RFC: build-tested only--- migration/rdma.c | 27 ++++++++++----------------- 1 file changed, 10 insertions(+), 17 deletions(-)migration/rdma.c +10
-17@@ -38,6 +38,9 @@#include "qom/object.h" #include <poll.h> +typedef struct rdma_cm_event rdma_cm_event ; +G_DEFINE_AUTOPTR_CLEANUP_FUNC (rdma_cm_event, rdma_ack_cm_event) +/* * Print and error on both the Monitor and the Log file. */ @@ -939,7 +942,7 @@ static int qemu_rdma_resolve_host(RDMAContext *rdma, Error **errp)int ret;struct rdma_addrinfo *res;char port_str[16 ]; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;char ip[40 ] ="unknown" ;struct rdma_addrinfo *e;@@ -1007,11 +1010,11 @@ route:ERROR (errp,"result not equal to event_addr_resolved %s" ,rdma_event_str (cm_event->event));perror ("rdma_resolve_addr" ); -rdma_ack_cm_event (cm_event); ret = -EINVAL ; goto err_resolve_get_addr; }rdma_ack_cm_event (cm_event); + cm_event =NULL ;/* resolve route */ ret =rdma_resolve_route (rdma->cm_id,RDMA_RESOLVE_TIMEOUT_MS );@@ -1028,11 +1031,9 @@ route:if (cm_event->event !=RDMA_CM_EVENT_ROUTE_RESOLVED ) {ERROR (errp,"result not equal to event_route_resolved: %s" ,rdma_event_str (cm_event->event)); -rdma_ack_cm_event (cm_event); ret = -EINVAL ; goto err_resolve_get_addr; } -rdma_ack_cm_event (cm_event); rdma->verbs = rdma->cm_id->verbs;qemu_rdma_dump_id ("source_resolve_host" , rdma->cm_id->verbs);qemu_rdma_dump_gid ("source_resolve_host" , rdma->cm_id);@@ -1501,7 +1502,7 @@ static uint64_t qemu_rdma_poll(RDMAContext *rdma, uint64_t *wr_id_out,*/ static int qemu_rdma_wait_comp_channel (RDMAContext *rdma) { -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;int ret =-1 ;/* @@ -2503,7 +2504,7 @@ static int qemu_rdma_connect(RDMAContext *rdma, Error **errp, bool return_path) .private_data = &cap, .private_data_len = sizeof(cap), }; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;int ret;/* @@ -2544,7 +2545,6 @@ static int qemu_rdma_connect(RDMAContext *rdma, Error **errp, bool return_path)if (cm_event->event !=RDMA_CM_EVENT_ESTABLISHED ) {perror ("rdma_get_cm_event != EVENT_ESTABLISHED after rdma_connect" );ERROR (errp,"connecting to destination!" ); -rdma_ack_cm_event (cm_event); goto err_rdma_source_connect; } rdma->connected = true;@@ -2564,8 +2564,6 @@ static int qemu_rdma_connect(RDMAContext *rdma, Error **errp, bool return_path)trace_qemu_rdma_connect_pin_all_outcome (rdma->pin_all); -rdma_ack_cm_event (cm_event); - rdma->control_ready_expected =1 ; rdma->nb_sent =0 ;return 0 ;@@ -3279,7 +3277,7 @@ static void rdma_cm_poll_handler(void *opaque) {RDMAContext *rdma = opaque;int ret; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;MigrationIncomingState *mis =migration_incoming_get_current (); ret =rdma_get_cm_event (rdma->channel, &cm_event);@@ -3287,7 +3285,6 @@ static void rdma_cm_poll_handler(void *opaque)error_report ("get_cm_event failed %d" , errno);return ; } -rdma_ack_cm_event (cm_event);if (cm_event->event ==RDMA_CM_EVENT_DISCONNECTED || cm_event->event ==RDMA_CM_EVENT_DEVICE_REMOVAL ) {@@ -3317,7 +3314,7 @@ static int qemu_rdma_accept(RDMAContext *rdma) .private_data_len = sizeof(cap), };RDMAContext *rdma_return_path =NULL ; -struct rdma_cm_event *cm_event; +g_autoptr (rdma_cm_event) cm_event =NULL ;struct ibv_context *verbs;int ret = -EINVAL ;int idx;@@ -3328,7 +3325,6 @@ static int qemu_rdma_accept(RDMAContext *rdma) }if (cm_event->event !=RDMA_CM_EVENT_CONNECT_REQUEST ) { -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; }@@ -3339,7 +3335,6 @@ static int qemu_rdma_accept(RDMAContext *rdma)if (migrate_postcopy () && !rdma->is_return_path) { rdma_return_path =qemu_rdma_data_init (rdma->host_port,NULL );if (rdma_return_path ==NULL ) { -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; }@@ -3353,7 +3348,6 @@ static int qemu_rdma_accept(RDMAContext *rdma)if (cap.version <1 || cap.version >RDMA_CONTROL_VERSION_CURRENT ) {error_report ("Unknown source RDMA version: %d, bailing..." , cap.version); -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; }@@ -3374,6 +3368,7 @@ static int qemu_rdma_accept(RDMAContext *rdma) verbs = cm_event->id->verbs;rdma_ack_cm_event (cm_event); + cm_event =NULL ;trace_qemu_rdma_accept_pin_state (rdma->pin_all);@@ -3441,11 +3436,9 @@ static int qemu_rdma_accept(RDMAContext *rdma)if (cm_event->event !=RDMA_CM_EVENT_ESTABLISHED ) {error_report ("rdma_accept not event established" ); -rdma_ack_cm_event (cm_event); goto err_rdma_dest_wait; } -rdma_ack_cm_event (cm_event); rdma->connected = true; ret =qemu_rdma_post_recv_control (rdma,RDMA_WRID_READY );Signature · 1 line
2.26.3
Signature · 1 line
Dr. David Alan Gilbert / [email protected] / Manchester, UK