Skip to contentSkip to conversation

main-loop: Disable GLOBAL_STATE_CODE() assertions

PATCH In review Hanna Reitz 4 participants 5 messages Mar 29, 2022

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 main-loop: Disable GLOBAL_STATE_CODE() assertions

These assertions are very useful for developers to find bugs, and so they have indeed pointed us towards bugs already. For users, it is not so useful to find these bugs. We should probably not enable them in releases until we are sufficiently certain that they will not fire during normal operation, unless something is going seriously wrong.

For example, we have received a bug report that you cannot add an NBD server on a BDS in an I/O thread with -incoming defer. I am sure this is a real bug that needs investigation, but we do not really have that time right now, so close to release, and so I would rather disable the assertions to get time to investigate such reports.

(I am just putting the link as "buglink" below, not "closes", because disabling the assertion will not fix the likely underlying bug.)

---
 include/qemu/main-loop.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
include/qemu/main-loop.h +2 -1
index 7a4d6a0920..3bf8aeb3cd 100644
@@ -272,7 +272,8 @@ bool qemu_in_main_thread(void);
 /* Mark and check that the function is part of the global state API. */
 #define GLOBAL_STATE_CODE()                                         \
     do {                                                            \
-        assert(qemu_in_main_thread());                              \
+        /* FIXME: Re-enable after 7.0 release */                    \
+        /* assert(qemu_in_main_thread()); */                        \
     } while (0)
 
 /* Mark and check that the function is part of the I/O API. */
Signature · 1 line
2.35.1
Reviewed-by Tested-by on v1 1/1
#1 lore raw
Quoted 12 lines · Hanna Reitz

These assertions are very useful for developers to find bugs, and so they have indeed pointed us towards bugs already. For users, it is not so useful to find these bugs. We should probably not enable them in releases until we are sufficiently certain that they will not fire during normal operation, unless something is going seriously wrong.

For example, we have received a bug report that you cannot add an NBD server on a BDS in an I/O thread with -incoming defer. I am sure this is a real bug that needs investigation, but we do not really have that time right now, so close to release, and so I would rather disable the assertions to get time to investigate such reports.

Quoted 4 lines

(I am just putting the link as "buglink" below, not "closes", because disabling the assertion will not fix the likely underlying bug.)

Also helps:
https://gitlab.com/qemu-project/qemu/-/issues/926

Quoted 10 lines
---
  include/qemu/main-loop.h | 3 ++-
  1 file changed, 2 insertions(+), 1 deletion(-)
include/qemu/main-loop.h +0 -0
index 7a4d6a0920..3bf8aeb3cd 100644

include/qemu/main-loop.h bool qemu_in_main_thread(void); · line 279

Quoted 9 lines
include/qemu/main-loop.h +2 -1
@@ -272,7 +272,8 @@ bool qemu_in_main_thread(void);
  /* Mark and check that the function is part of the global state API. */
  #define GLOBAL_STATE_CODE()                                         \
      do {                                                            \
-        assert(qemu_in_main_thread());                              \
+        /* FIXME: Re-enable after 7.0 release */                    \
+        /* assert(qemu_in_main_thread()); */                        \
      } while (0)
 
  /* Mark and check that the function is part of the I/O API. */

Do you want me to cancel/repost my PR without this patch?

https://lore.kernel.org/qemu-devel/[email protected]/

Otherwise:

#2 lore raw
Quoted 42 lines · Philippe Mathieu-Daudé
Quoted 12 lines · Hanna Reitz

These assertions are very useful for developers to find bugs, and so they have indeed pointed us towards bugs already.  For users, it is not so useful to find these bugs.  We should probably not enable them in releases until we are sufficiently certain that they will not fire during normal operation, unless something is going seriously wrong.

For example, we have received a bug report that you cannot add an NBD server on a BDS in an I/O thread with -incoming defer.  I am sure this is a real bug that needs investigation, but we do not really have that time right now, so close to release, and so I would rather disable the assertions to get time to investigate such reports.

Quoted 4 lines

(I am just putting the link as "buglink" below, not "closes", because disabling the assertion will not fix the likely underlying bug.)

Also helps:
https://gitlab.com/qemu-project/qemu/-/issues/926

Quoted 10 lines

---
  include/qemu/main-loop.h | 3 ++-

  1 file changed, 2 insertions(+), 1 deletion(-)
include/qemu/main-loop.h +0 -0
index 7a4d6a0920..3bf8aeb3cd 100644

include/qemu/main-loop.h bool qemu_in_main_thread(void); · line 278

Quoted 8 lines
include/qemu/main-loop.h +2 -1
@@ -272,7 +272,8 @@ bool qemu_in_main_thread(void);
   /* Mark and check that the function is part of the global state API. */
   #define GLOBAL_STATE_CODE()                                         \
       do {                                                            \
- assert(qemu_in_main_thread());                              \
+        /* FIXME: Re-enable after 7.0 release */                    \
+        /* assert(qemu_in_main_thread()); */                        \
       } while (0)
     /* Mark and check that the function is part of the I/O API. */

Do you want me to cancel/repost my PR without this patch?

Quoted 1 line

https://lore.kernel.org/qemu-devel/[email protected]/

I think we should let Peter take your PR first, as long as the discussion on this is still out.  I’d like to give it a couple more hours, and as far as I understand, we definitely want yours.

(Taking mine will then cause me rebase conflicts and make it look weird, but that’s not too bad for something that’s just a temporary band-aid anyway.)

Hanna

Reviewed-by on v1 1/1
#3 lore raw
In reply to Hanna Reitz

Am 29/03/2022 um 11:35 schrieb Hanna Reitz:

Quoted 27 lines

These assertions are very useful for developers to find bugs, and so they have indeed pointed us towards bugs already. For users, it is not so useful to find these bugs. We should probably not enable them in releases until we are sufficiently certain that they will not fire during normal operation, unless something is going seriously wrong.

For example, we have received a bug report that you cannot add an NBD server on a BDS in an I/O thread with -incoming defer. I am sure this is a real bug that needs investigation, but we do not really have that time right now, so close to release, and so I would rather disable the assertions to get time to investigate such reports.

(I am just putting the link as "buglink" below, not "closes", because disabling the assertion will not fix the likely underlying bug.)

---
 include/qemu/main-loop.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
include/qemu/main-loop.h +0 -0
index 7a4d6a0920..3bf8aeb3cd 100644

include/qemu/main-loop.h bool qemu_in_main_thread(void); · line 279

Quoted 9 lines
include/qemu/main-loop.h +2 -1
@@ -272,7 +272,8 @@ bool qemu_in_main_thread(void);
 /* Mark and check that the function is part of the global state API. */
 #define GLOBAL_STATE_CODE()                                         \
     do {                                                            \
-        assert(qemu_in_main_thread());                              \
+        /* FIXME: Re-enable after 7.0 release */                    \
+        /* assert(qemu_in_main_thread()); */                        \
     } while (0)
 
 /* Mark and check that the function is part of the I/O API. */
Reviewed-by on v1 1/1
#4 lore raw
In reply to Hanna Reitz
Quoted 12 lines · Hanna Reitz

These assertions are very useful for developers to find bugs, and so they have indeed pointed us towards bugs already. For users, it is not so useful to find these bugs. We should probably not enable them in releases until we are sufficiently certain that they will not fire during normal operation, unless something is going seriously wrong.

For example, we have received a bug report that you cannot add an NBD server on a BDS in an I/O thread with -incoming defer. I am sure this is a real bug that needs investigation, but we do not really have that time right now, so close to release, and so I would rather disable the assertions to get time to investigate such reports.

Quoted 8 lines

(I am just putting the link as "buglink" below, not "closes", because disabling the assertion will not fix the likely underlying bug.)

---
 include/qemu/main-loop.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Sounds reasonable to me.