summaryrefslogtreecommitdiffstats
path: root/results/classifier/111/review/1869426
diff options
context:
space:
mode:
Diffstat (limited to 'results/classifier/111/review/1869426')
-rw-r--r--results/classifier/111/review/1869426113
1 files changed, 113 insertions, 0 deletions
diff --git a/results/classifier/111/review/1869426 b/results/classifier/111/review/1869426
new file mode 100644
index 00000000..c74c2f46
--- /dev/null
+++ b/results/classifier/111/review/1869426
@@ -0,0 +1,113 @@
+device: 0.114
+PID: 0.093
+other: 0.092
+permissions: 0.086
+semantic: 0.083
+files: 0.066
+boot: 0.063
+vnc: 0.061
+debug: 0.060
+performance: 0.060
+graphic: 0.058
+socket: 0.057
+network: 0.056
+KVM: 0.050
+debug: 0.655
+files: 0.089
+device: 0.042
+PID: 0.039
+other: 0.036
+KVM: 0.025
+semantic: 0.020
+performance: 0.017
+graphic: 0.016
+vnc: 0.014
+boot: 0.013
+socket: 0.013
+network: 0.012
+permissions: 0.011
+
+5.0rc0->4.2 serial migraiton
+
+Migrating from 5.0rc0->4.2 with pc-q35-4.2 we get an error:
+
+Unknown savevm section or instance 'serial' 1
+
+dumping the migration streams it looks like 5.0 is duplicating the serial migration data:
+
+ "serial (26)": {
+ "divider": "0x000c",
+ "rbr": "0x00",
+ "ier": "0x00",
+ "iir": "0x01",
+ "lcr": "0x00",
+ "mcr": "0x00",
+ "lsr": "0x60",
+ "msr": "0xb0",
+ "scr": "0x00",
+ "fcr_vmstate": "0x00"
+ },
+ "serial (27)": {
+ "state": {
+ "divider": "0x000c",
+ "rbr": "0x00",
+ "ier": "0x00",
+ "iir": "0x01",
+ "lcr": "0x00",
+ "mcr": "0x00",
+ "lsr": "0x60",
+ "msr": "0xb0",
+ "scr": "0x00",
+ "fcr_vmstate": "0x00"
+ }
+ },
+
+git bisect says:
+
+c9808d602813bce4fada7bf9ecc463aa779b73f7 is the first bad commit
+commit c9808d602813bce4fada7bf9ecc463aa779b73f7
+Author: Marc-André Lureau <email address hidden>
+Date: Tue Oct 22 01:02:50 2019 +0200
+
+ serial: realize the serial device
+
+ Instead of calling serial_realize_core(), use the QDev realize
+ callback.
+
+ Signed-off-by: Marc-André Lureau <email address hidden>
+ Reviewed-by: Philippe Mathieu-Daudé <email address hidden>
+
+
+Marc-Andre: I think you're ending up with two top level objects with vmsd's
+
+I thought backward migration wasn't supported.
+
+Isn't it this commit?
+
+commit 4cc017e505df7e5344e4dfe7fc17711117c5f11f
+Author: Marc-André Lureau <email address hidden>
+Date: Tue Oct 22 00:32:41 2019 +0200
+
+ serial: register vmsd with DeviceClass
+
+ Migration from old to new code works, however the other way fails for
+ devices that use serial_init/serial_mm_init with "base", used as
+ instance_id previously.
+
+ (with qdev_set_legacy_instance_id, the alias_id is only used in
+ savevm.c:find_se(), and thus can only be used to match against
+ "legacy" instance id values. On new code, instance_id is generated
+ incrementally from 0 with calculate_new_instance_id(), based on
+ "qdev-path/vmsd-name")
+
+ Signed-off-by: Marc-André Lureau <email address hidden>
+ Reviewed-by: xiaoqiang zhao <zxq_yx_007@163.com>
+
+
+Fix posted:
+https://lists.gnu.org/archive/html/qemu-devel/2020-03/msg08803.html
+
+Fixed here:
+https://git.qemu.org/?p=qemu.git;a=commitdiff;h=f602d047ac21
+
+