When ->poll() succeeds the AioHandler is placed on the ready list with
revents set to the magic value 0. This magic value causes
aio_dispatch_handler() to invoke ->poll_ready() instead of ->io_read()
for G_IO_IN or ->io_write() for G_IO_OUT.

This magic value 0 hack works for the IOThread where AioHandlers are
placed on ->ready_list and processed by aio_dispatch_ready_handlers().
It does not work for the main loop where all AioHandlers are processed
by aio_dispatch_handlers(), even those that are not ready and have a
revents value of 0.

As a result the main loop invokes ->poll_ready() on AioHandlers that are
not ready. These spurious ->poll_ready() calls waste CPU cycles and
could lead to crashes if the code assumes ->poll() must have succeeded
before ->poll_ready() is called (a reasonable asumption but I haven't
seen it in practice).

Stop using revents to track whether ->poll_ready() will be called on an
AioHandler. Introduce a separate AioHandler->poll_ready field instead.
This eliminates spurious ->poll_ready() calls in the main loop.

Fixes: 826cc32423db2a99d184dbf4f507c737d7e7a4ae ("aio-posix: split poll check from ready handler")
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
---
 util/aio-posix.h |  1 +
 util/aio-posix.c | 32 ++++++++++++++++++--------------
 2 files changed, 19 insertions(+), 14 deletions(-)

diff --git a/util/aio-posix.h b/util/aio-posix.h
index 7f2c37a684..80b927c7f4 100644
--- a/util/aio-posix.h
+++ b/util/aio-posix.h
@@ -37,6 +37,7 @@ struct AioHandler {
     unsigned flags; /* see fdmon-io_uring.c */
 #endif
     int64_t poll_idle_timeout; /* when to stop userspace polling */
+    bool poll_ready; /* has polling detected an event? */
     bool is_external;
 };
 
diff --git a/util/aio-posix.c b/util/aio-posix.c
index 7b9f629218..be0182a3c6 100644
--- a/util/aio-posix.c
+++ b/util/aio-posix.c
@@ -23,15 +23,6 @@
 #include "trace.h"
 #include "aio-posix.h"
 
-/*
- * G_IO_IN and G_IO_OUT are not appropriate revents values for polling, since
- * the handler may not need to access the file descriptor. For example, the
- * handler doesn't need to read from an EventNotifier if it polled a memory
- * location and a read syscall would be slow. Define our own unique revents
- * value to indicate that polling determined this AioHandler is ready.
- */
-#define REVENTS_POLL_READY 0
-
 /* Stop userspace polling on a handler if it isn't active for some time */
 #define POLL_IDLE_INTERVAL_NS (7 * NANOSECONDS_PER_SECOND)
 
@@ -49,6 +40,14 @@ void aio_add_ready_handler(AioHandlerList *ready_list,
     QLIST_INSERT_HEAD(ready_list, node, node_ready);
 }
 
+static void aio_add_poll_ready_handler(AioHandlerList *ready_list,
+                                       AioHandler *node)
+{
+    QLIST_SAFE_REMOVE(node, node_ready); /* remove from nested parent's list */
+    node->poll_ready = true;
+    QLIST_INSERT_HEAD(ready_list, node, node_ready);
+}
+
 static AioHandler *find_aio_handler(AioContext *ctx, int fd)
 {
     AioHandler *node;
@@ -76,6 +75,7 @@ static bool aio_remove_fd_handler(AioContext *ctx, AioHandler *node)
     }
 
     node->pfd.revents = 0;
+    node->poll_ready = false;
 
     /* If the fd monitor has already marked it deleted, leave it alone */
     if (QLIST_IS_INSERTED(node, node_deleted)) {
@@ -247,7 +247,7 @@ static bool poll_set_started(AioContext *ctx, AioHandlerList *ready_list,
 
         /* Poll one last time in case ->io_poll_end() raced with the event */
         if (!started && node->io_poll(node->opaque)) {
-            aio_add_ready_handler(ready_list, node, REVENTS_POLL_READY);
+            aio_add_poll_ready_handler(ready_list, node);
             progress = true;
         }
     }
@@ -282,6 +282,7 @@ bool aio_pending(AioContext *ctx)
     QLIST_FOREACH_RCU(node, &ctx->aio_handlers, node) {
         int revents;
 
+        /* TODO should this check poll ready? */
         revents = node->pfd.revents & node->pfd.events;
         if (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR) && node->io_read &&
             aio_node_check(ctx, node->is_external)) {
@@ -323,11 +324,15 @@ static void aio_free_deleted_handlers(AioContext *ctx)
 static bool aio_dispatch_handler(AioContext *ctx, AioHandler *node)
 {
     bool progress = false;
+    bool poll_ready;
     int revents;
 
     revents = node->pfd.revents & node->pfd.events;
     node->pfd.revents = 0;
 
+    poll_ready = node->poll_ready;
+    node->poll_ready = false;
+
     /*
      * Start polling AioHandlers when they become ready because activity is
      * likely to continue.  Note that starvation is theoretically possible when
@@ -344,7 +349,7 @@ static bool aio_dispatch_handler(AioContext *ctx, AioHandler *node)
         QLIST_INSERT_HEAD(&ctx->poll_aio_handlers, node, node_poll);
     }
     if (!QLIST_IS_INSERTED(node, node_deleted) &&
-        revents == 0 &&
+        poll_ready && revents == 0 &&
         aio_node_check(ctx, node->is_external) &&
         node->io_poll_ready) {
         node->io_poll_ready(node->opaque);
@@ -432,7 +437,7 @@ static bool run_poll_handlers_once(AioContext *ctx,
     QLIST_FOREACH_SAFE(node, &ctx->poll_aio_handlers, node_poll, tmp) {
         if (aio_node_check(ctx, node->is_external) &&
             node->io_poll(node->opaque)) {
-            aio_add_ready_handler(ready_list, node, REVENTS_POLL_READY);
+            aio_add_poll_ready_handler(ready_list, node);
 
             node->poll_idle_timeout = now + POLL_IDLE_INTERVAL_NS;
 
@@ -491,8 +496,7 @@ static bool remove_idle_poll_handlers(AioContext *ctx,
                  * this causes progress.
                  */
                 if (node->io_poll(node->opaque)) {
-                    aio_add_ready_handler(ready_list, node,
-                                          REVENTS_POLL_READY);
+                    aio_add_poll_ready_handler(ready_list, node);
                     progress = true;
                 }
             }
-- 
2.34.1

On Wed, Feb 23, 2022 at 03:57:03PM +0000, Stefan Hajnoczi wrote:
> When ->poll() succeeds the AioHandler is placed on the ready list with
> revents set to the magic value 0. This magic value causes
> aio_dispatch_handler() to invoke ->poll_ready() instead of ->io_read()
> for G_IO_IN or ->io_write() for G_IO_OUT.
> 
> This magic value 0 hack works for the IOThread where AioHandlers are
> placed on ->ready_list and processed by aio_dispatch_ready_handlers().
> It does not work for the main loop where all AioHandlers are processed
> by aio_dispatch_handlers(), even those that are not ready and have a
> revents value of 0.
> 
> As a result the main loop invokes ->poll_ready() on AioHandlers that are
> not ready. These spurious ->poll_ready() calls waste CPU cycles and
> could lead to crashes if the code assumes ->poll() must have succeeded
> before ->poll_ready() is called (a reasonable asumption but I haven't
> seen it in practice).
> 
> Stop using revents to track whether ->poll_ready() will be called on an
> AioHandler. Introduce a separate AioHandler->poll_ready field instead.
> This eliminates spurious ->poll_ready() calls in the main loop.
> 
> Fixes: 826cc32423db2a99d184dbf4f507c737d7e7a4ae ("aio-posix: split poll check from ready handler")
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
> ---
>  util/aio-posix.h |  1 +
>  util/aio-posix.c | 32 ++++++++++++++++++--------------
>  2 files changed, 19 insertions(+), 14 deletions(-)

Applied to my block tree:
https://gitlab.com/stefanha/qemu/commits/block

Stefan

在 2022/2/23 下午11:57, Stefan Hajnoczi 写道:
> When ->poll() succeeds the AioHandler is placed on the ready list with
> revents set to the magic value 0. This magic value causes
> aio_dispatch_handler() to invoke ->poll_ready() instead of ->io_read()
> for G_IO_IN or ->io_write() for G_IO_OUT.
>
> This magic value 0 hack works for the IOThread where AioHandlers are
> placed on ->ready_list and processed by aio_dispatch_ready_handlers().
> It does not work for the main loop where all AioHandlers are processed
> by aio_dispatch_handlers(), even those that are not ready and have a
> revents value of 0.
>
> As a result the main loop invokes ->poll_ready() on AioHandlers that are
> not ready. These spurious ->poll_ready() calls waste CPU cycles and
> could lead to crashes if the code assumes ->poll() must have succeeded
> before ->poll_ready() is called (a reasonable asumption but I haven't
> seen it in practice).
>
> Stop using revents to track whether ->poll_ready() will be called on an
> AioHandler. Introduce a separate AioHandler->poll_ready field instead.
> This eliminates spurious ->poll_ready() calls in the main loop.
>
> Fixes: 826cc32423db2a99d184dbf4f507c737d7e7a4ae ("aio-posix: split poll check from ready handler")
> Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>


Reported-by: Jason Wang <jasowang@redhat.com>

Tested-by: Jason Wang <jasowang@redhat.com>

Thanks


> ---
>   util/aio-posix.h |  1 +
>   util/aio-posix.c | 32 ++++++++++++++++++--------------
>   2 files changed, 19 insertions(+), 14 deletions(-)
>
> diff --git a/util/aio-posix.h b/util/aio-posix.h
> index 7f2c37a684..80b927c7f4 100644
> --- a/util/aio-posix.h
> +++ b/util/aio-posix.h
> @@ -37,6 +37,7 @@ struct AioHandler {
>       unsigned flags; /* see fdmon-io_uring.c */
>   #endif
>       int64_t poll_idle_timeout; /* when to stop userspace polling */
> +    bool poll_ready; /* has polling detected an event? */
>       bool is_external;
>   };
>   
> diff --git a/util/aio-posix.c b/util/aio-posix.c
> index 7b9f629218..be0182a3c6 100644
> --- a/util/aio-posix.c
> +++ b/util/aio-posix.c
> @@ -23,15 +23,6 @@
>   #include "trace.h"
>   #include "aio-posix.h"
>   
> -/*
> - * G_IO_IN and G_IO_OUT are not appropriate revents values for polling, since
> - * the handler may not need to access the file descriptor. For example, the
> - * handler doesn't need to read from an EventNotifier if it polled a memory
> - * location and a read syscall would be slow. Define our own unique revents
> - * value to indicate that polling determined this AioHandler is ready.
> - */
> -#define REVENTS_POLL_READY 0
> -
>   /* Stop userspace polling on a handler if it isn't active for some time */
>   #define POLL_IDLE_INTERVAL_NS (7 * NANOSECONDS_PER_SECOND)
>   
> @@ -49,6 +40,14 @@ void aio_add_ready_handler(AioHandlerList *ready_list,
>       QLIST_INSERT_HEAD(ready_list, node, node_ready);
>   }
>   
> +static void aio_add_poll_ready_handler(AioHandlerList *ready_list,
> +                                       AioHandler *node)
> +{
> +    QLIST_SAFE_REMOVE(node, node_ready); /* remove from nested parent's list */
> +    node->poll_ready = true;
> +    QLIST_INSERT_HEAD(ready_list, node, node_ready);
> +}
> +
>   static AioHandler *find_aio_handler(AioContext *ctx, int fd)
>   {
>       AioHandler *node;
> @@ -76,6 +75,7 @@ static bool aio_remove_fd_handler(AioContext *ctx, AioHandler *node)
>       }
>   
>       node->pfd.revents = 0;
> +    node->poll_ready = false;
>   
>       /* If the fd monitor has already marked it deleted, leave it alone */
>       if (QLIST_IS_INSERTED(node, node_deleted)) {
> @@ -247,7 +247,7 @@ static bool poll_set_started(AioContext *ctx, AioHandlerList *ready_list,
>   
>           /* Poll one last time in case ->io_poll_end() raced with the event */
>           if (!started && node->io_poll(node->opaque)) {
> -            aio_add_ready_handler(ready_list, node, REVENTS_POLL_READY);
> +            aio_add_poll_ready_handler(ready_list, node);
>               progress = true;
>           }
>       }
> @@ -282,6 +282,7 @@ bool aio_pending(AioContext *ctx)
>       QLIST_FOREACH_RCU(node, &ctx->aio_handlers, node) {
>           int revents;
>   
> +        /* TODO should this check poll ready? */
>           revents = node->pfd.revents & node->pfd.events;
>           if (revents & (G_IO_IN | G_IO_HUP | G_IO_ERR) && node->io_read &&
>               aio_node_check(ctx, node->is_external)) {
> @@ -323,11 +324,15 @@ static void aio_free_deleted_handlers(AioContext *ctx)
>   static bool aio_dispatch_handler(AioContext *ctx, AioHandler *node)
>   {
>       bool progress = false;
> +    bool poll_ready;
>       int revents;
>   
>       revents = node->pfd.revents & node->pfd.events;
>       node->pfd.revents = 0;
>   
> +    poll_ready = node->poll_ready;
> +    node->poll_ready = false;
> +
>       /*
>        * Start polling AioHandlers when they become ready because activity is
>        * likely to continue.  Note that starvation is theoretically possible when
> @@ -344,7 +349,7 @@ static bool aio_dispatch_handler(AioContext *ctx, AioHandler *node)
>           QLIST_INSERT_HEAD(&ctx->poll_aio_handlers, node, node_poll);
>       }
>       if (!QLIST_IS_INSERTED(node, node_deleted) &&
> -        revents == 0 &&
> +        poll_ready && revents == 0 &&
>           aio_node_check(ctx, node->is_external) &&
>           node->io_poll_ready) {
>           node->io_poll_ready(node->opaque);
> @@ -432,7 +437,7 @@ static bool run_poll_handlers_once(AioContext *ctx,
>       QLIST_FOREACH_SAFE(node, &ctx->poll_aio_handlers, node_poll, tmp) {
>           if (aio_node_check(ctx, node->is_external) &&
>               node->io_poll(node->opaque)) {
> -            aio_add_ready_handler(ready_list, node, REVENTS_POLL_READY);
> +            aio_add_poll_ready_handler(ready_list, node);
>   
>               node->poll_idle_timeout = now + POLL_IDLE_INTERVAL_NS;
>   
> @@ -491,8 +496,7 @@ static bool remove_idle_poll_handlers(AioContext *ctx,
>                    * this causes progress.
>                    */
>                   if (node->io_poll(node->opaque)) {
> -                    aio_add_ready_handler(ready_list, node,
> -                                          REVENTS_POLL_READY);
> +                    aio_add_poll_ready_handler(ready_list, node);
>                       progress = true;
>                   }
>               }