diff options
Diffstat (limited to '')
| -rw-r--r-- | results/classifier/accel-gemma3:12b/kvm/51 | 2 | ||||
| -rw-r--r-- | results/classifier/accel-gemma3:12b/kvm/511 | 2 | ||||
| -rw-r--r-- | results/classifier/accel-gemma3:12b/kvm/512 | 2 | ||||
| -rw-r--r-- | results/classifier/accel-gemma3:12b/kvm/513 | 2 | ||||
| -rw-r--r-- | results/classifier/accel-gemma3:12b/kvm/516 | 43 | ||||
| -rw-r--r-- | results/classifier/accel-gemma3:12b/kvm/517 | 2 | ||||
| -rw-r--r-- | results/classifier/accel-gemma3:12b/kvm/519 | 34 |
7 files changed, 87 insertions, 0 deletions
diff --git a/results/classifier/accel-gemma3:12b/kvm/51 b/results/classifier/accel-gemma3:12b/kvm/51 new file mode 100644 index 000000000..4e8d87b4b --- /dev/null +++ b/results/classifier/accel-gemma3:12b/kvm/51 @@ -0,0 +1,2 @@ + +Linux kernel oops on Malta board while accessing pflash diff --git a/results/classifier/accel-gemma3:12b/kvm/511 b/results/classifier/accel-gemma3:12b/kvm/511 new file mode 100644 index 000000000..28a2b44c3 --- /dev/null +++ b/results/classifier/accel-gemma3:12b/kvm/511 @@ -0,0 +1,2 @@ + +usbredirparser: bulk transfer length exceeds limits (can't use any USB storage) diff --git a/results/classifier/accel-gemma3:12b/kvm/512 b/results/classifier/accel-gemma3:12b/kvm/512 new file mode 100644 index 000000000..2597364ff --- /dev/null +++ b/results/classifier/accel-gemma3:12b/kvm/512 @@ -0,0 +1,2 @@ + +6.1.0-rc1 New regression (not in 6.1.0-rc0): Freezes using UEFI firmware without acceleration diff --git a/results/classifier/accel-gemma3:12b/kvm/513 b/results/classifier/accel-gemma3:12b/kvm/513 new file mode 100644 index 000000000..6ced9a17a --- /dev/null +++ b/results/classifier/accel-gemma3:12b/kvm/513 @@ -0,0 +1,2 @@ + +Qemu/WHPX fails on applying UEFI firmware with -pflash diff --git a/results/classifier/accel-gemma3:12b/kvm/516 b/results/classifier/accel-gemma3:12b/kvm/516 new file mode 100644 index 000000000..4c707b5af --- /dev/null +++ b/results/classifier/accel-gemma3:12b/kvm/516 @@ -0,0 +1,43 @@ + +Configure option `--enable-plugins` makes modules in shared library not loadable on macOS +Description of problem: +The title mentions `--enable-plugins` option, however as it's enabled by default, not providing `--disable-plugins` would also cause this to happen. + +If TCG plugin support is enabled, symbols in `qemu-system-*` binaries will be missing, and module libraries would fail to load as they expect those symbols to exist in the main binary. + +Configure options used: `STRIP="strip -x" ./configure --enable-user --enable-tools --enable-parallels --enable-libxml2 --enable-spice --enable-hvf --enable-cocoa --enable-guest-agent --enable-curses --enable-plugins --enable-modules --objcc=gcc --enable-libusb --enable-usb-redir` + +After inspecting the compiler command line, I've found the linker option `-Wl,-exported_symbols_list,qemu-plugins-ld64.symbols` is causing this to happen: only symbols listed in `qemu-plugins-ld64.symbols` would be kept in `qemu-system-*` binaries and all other symbols will be hidden. + +Note that this is not caused by stripping (although I had to use custom strip command line on macOS to successfully compile qemu); the option `-exported_symbols_list` works by only exposing the provided symbols and treating all other symbols as `visibility=hidden`. + +Replacing `--enable-plugins` to `--disable-plugins` in the above configure command line would "fix" it, although it means TCG plugins will not be supported. +Steps to reproduce: +1. Build QEMU on macOS with plugin support enabled +2. Try to use modules in shared library like qxl +Additional information: +Some examples: + +``` +$ qemu-system-x86_64 -device qxl +Failed to open module: dlopen(/usr/local/bin/../lib/qemu/ui-spice-core.dylib, 10): Symbol not found: __TRACE_QEMU_SPICE_ADD_MEMSLOT_DSTATE + Referenced from: /usr/local/bin/../lib/qemu/ui-spice-core.dylib + Expected in: flat namespace + in /usr/local/bin/../lib/qemu/ui-spice-core.dylib +Failed to open module: dlopen(/usr/local/bin/../lib/qemu/hw-display-qxl.dylib, 2): Symbol not found: __TRACE_QXL_CLIENT_MONITORS_CONFIG_CAPPED_DSTATE + Referenced from: /usr/local/bin/../lib/qemu/hw-display-qxl.dylib + Expected in: flat namespace + in /usr/local/bin/../lib/qemu/hw-display-qxl.dylib +qemu-system-x86_64: -device qxl: 'qxl' is not a valid device model name +``` + +``` +$ qemu-system-x86_64 -spice port=5901 +Failed to open module: dlopen(/usr/local/bin/../lib/qemu/ui-spice-core.dylib, 10): Symbol not found: __TRACE_QEMU_SPICE_ADD_MEMSLOT_DSTATE + Referenced from: /usr/local/bin/../lib/qemu/ui-spice-core.dylib + Expected in: flat namespace + in /usr/local/bin/../lib/qemu/ui-spice-core.dylib +qemu-system-x86_64: -spice port=5901: spice support is disabled +``` + +After disabling plugin support I could run virtual machines locally through libvirt with full spice and qxl video support. diff --git a/results/classifier/accel-gemma3:12b/kvm/517 b/results/classifier/accel-gemma3:12b/kvm/517 new file mode 100644 index 000000000..6343d76ae --- /dev/null +++ b/results/classifier/accel-gemma3:12b/kvm/517 @@ -0,0 +1,2 @@ + +Abort in vmxnet3_setup_tx_offloads diff --git a/results/classifier/accel-gemma3:12b/kvm/519 b/results/classifier/accel-gemma3:12b/kvm/519 new file mode 100644 index 000000000..a36febedd --- /dev/null +++ b/results/classifier/accel-gemma3:12b/kvm/519 @@ -0,0 +1,34 @@ + +xive: An extra '0x' prefix when printing hex values in the trace. +Description of problem: +The trace functions corresponding to the functions below print certain +parameters with a double "0x" prefix, i.e., with "0x0x". + + +- xive_source_esb_read +- xive_source_esb_write +- xive_tctx_tm_write +- xive_tctx_tm_read +Steps to reproduce: +1. Execute the command line on a terminal +2. Watch the terminal for the output. +Additional information: +Sample output: + + xive_end_source_read END 0x0/0xf @0x0x1e0800 + xive_source_esb_read @0x0x210c00 IRQ 0x10 val=0x0x1 + xive_tctx_tm_read @0x0x10038 sz=1 val=0x0 + xive_tctx_tm_write @0x0x10038 sz=1 val=0x80 + xive_tctx_tm_read @0x0x10038 sz=1 val=0x80 + xive_end_source_read END 0x0/0xf @0x0x1e0800 + xive_source_esb_read @0x0x210c00 IRQ 0x10 val=0x0x1 + xive_tctx_tm_read @0x0x10038 sz=1 val=0x0 + xive_tctx_tm_write @0x0x10038 sz=1 val=0x80 + xive_tctx_tm_read @0x0x10038 sz=1 val=0x80 + +The source lines at fault: + + xive_source_esb_read(uint64_t addr, uint32_t srcno, uint64_t value) "@0x0x%"PRIx64" IRQ 0x%x val=0x0x%"PRIx64 + xive_source_esb_write(uint64_t addr, uint32_t srcno, uint64_t value) "@0x0x%"PRIx64" IRQ 0x%x val=0x0x%"PRIx64 + xive_tctx_tm_write(uint64_t offset, unsigned int size, uint64_t value) "@0x0x%"PRIx64" sz=%d val=0x%" PRIx64 + xive_tctx_tm_read(uint64_t offset, unsigned int size, uint64_t value) "@0x0x%"PRIx64" sz=%d val=0x%" PRIx64 |