| Message ID | 20221228130439.80341-1-philmd@linaro.org |
|---|---|
| State | New |
| Headers | show |
| Series | memory: Do not print MR priority in flatview HMP output | expand |
Is there any chance between 7.1 and 7.2 ‘something’ happened to make it so that Qemu ‘cares more’ about e.g. when memory regions are added/removed?
I seem to get an abort because a memory region has not been completely setup in 7.2 (while it is being flattened actually) - In 7.1 that never seemed to happen….?
(Putting it all async work seems ‘excessive’ but maybe it’s needed now?)
Cheers
Mark
On 28/12/2022, 14:04, "Philippe Mathieu-Daudé" <philmd@linaro.org> wrote:
WARNING: This email originated from outside of Qualcomm. Please be wary of any links or attachments, and do not enable macros.
When requesting the flatview output with 'info mtree -f',
the MemoryRegion priority is irrelevant and noise. Remove it.
Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org>
---
softmmu/memory.c | 6 ++----
1 file changed, 2 insertions(+), 4 deletions(-)
diff --git a/softmmu/memory.c b/softmmu/memory.c
index e05332d07f..89713dd5ce 100644
--- a/softmmu/memory.c
+++ b/softmmu/memory.c
@@ -3330,22 +3330,20 @@ static void mtree_print_flatview(gpointer key, gpointer value,
mr = range->mr;
if (range->offset_in_region) {
qemu_printf(MTREE_INDENT TARGET_FMT_plx "-" TARGET_FMT_plx
- " (prio %d, %s%s): %s @" TARGET_FMT_plx,
+ " (%s%s): %s @" TARGET_FMT_plx,
int128_get64(range->addr.start),
int128_get64(range->addr.start)
+ MR_SIZE(range->addr.size),
- mr->priority,
range->nonvolatile ? "nv-" : "",
range->readonly ? "rom" : memory_region_type(mr),
memory_region_name(mr),
range->offset_in_region);
} else {
qemu_printf(MTREE_INDENT TARGET_FMT_plx "-" TARGET_FMT_plx
- " (prio %d, %s%s): %s",
+ " (%s%s): %s",
int128_get64(range->addr.start),
int128_get64(range->addr.start)
+ MR_SIZE(range->addr.size),
- mr->priority,
range->nonvolatile ? "nv-" : "",
range->readonly ? "rom" : memory_region_type(mr),
memory_region_name(mr));
--
2.38.1
On 28/12/22 17:26, Mark Burton wrote: > Is there any chance between 7.1 and 7.2 ‘something’ happened to make it > so that Qemu ‘cares more’ about e.g. when memory regions are added/removed? > > I seem to get an abort because a memory region has not been completely > setup in 7.2 (while it is being flattened actually) - In 7.1 that never > seemed to happen….? I couldn't find any relevant changes in softmmu/. Maybe TCG related? $ git diff --stat v7.1.0..v7.2.0 -- softmmu/ softmmu/cpus.c | 18 +++++++++++++----- softmmu/device_tree.c | 58 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ softmmu/dirtylimit.c | 4 ++-- softmmu/main.c | 10 +++++----- softmmu/memory.c | 72 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ softmmu/physmem.c | 36 ++++++++++++++++++------------------ softmmu/qemu-seccomp.c | 13 +++++++++++++ softmmu/qtest.c | 12 +++++++++--- softmmu/runstate.c | 11 ++++++++--- softmmu/vl.c | 85 +++++++++++++++++++++++++++++++++++++++++++++++++++---------------------------------- 10 files changed, 249 insertions(+), 70 deletions(-) $ git diff --stat v7.1.0..v7.2.0 -- accel/tcg/ accel/tcg/cpu-exec-common.c | 2 +- accel/tcg/cpu-exec.c | 229 +++++++++++++++------------ accel/tcg/cputlb.c | 350 +++++++++++++++++++++--------------------- accel/tcg/internal.h | 101 ++++++++++++ accel/tcg/meson.build | 1 + accel/tcg/plugin-gen.c | 22 +-- accel/tcg/tb-hash.h | 1 + accel/tcg/tb-jmp-cache.h | 65 ++++++++ accel/tcg/tb-maint.c | 704 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ accel/tcg/tcg-accel-ops-mttcg.c | 5 +- accel/tcg/tcg-accel-ops-rr.c | 9 +- accel/tcg/tcg-accel-ops.c | 98 ++++++++++++ accel/tcg/tcg-all.c | 17 +++ accel/tcg/translate-all.c | 1249 +++++++++++++++++++--------------------------------------------------------------------------------------------------------------------------------- accel/tcg/translator.c | 138 ++++++++++++----- accel/tcg/user-exec.c | 59 ++++++- 16 files changed, 1628 insertions(+), 1422 deletions(-) I'll have a look at $ git log -p v7.1.0..v7.2.0 accel/tcg/ > (Putting it all async work seems ‘excessive’ but maybe it’s needed now?) > > > Cheers > Mark
On Wed, Dec 28, 2022 at 02:04:39PM +0100, Philippe Mathieu-Daudé wrote: > When requesting the flatview output with 'info mtree -f', > the MemoryRegion priority is irrelevant and noise. Remove it. It is in many cases still helpful debugging information to me. Maybe it's because it can cause mis-alignment when the priority >= 10? If that's the case, how about tabbing it properly instead of removing? Thanks,
diff --git a/softmmu/memory.c b/softmmu/memory.c index e05332d07f..89713dd5ce 100644 --- a/softmmu/memory.c +++ b/softmmu/memory.c @@ -3330,22 +3330,20 @@ static void mtree_print_flatview(gpointer key, gpointer value, mr = range->mr; if (range->offset_in_region) { qemu_printf(MTREE_INDENT TARGET_FMT_plx "-" TARGET_FMT_plx - " (prio %d, %s%s): %s @" TARGET_FMT_plx, + " (%s%s): %s @" TARGET_FMT_plx, int128_get64(range->addr.start), int128_get64(range->addr.start) + MR_SIZE(range->addr.size), - mr->priority, range->nonvolatile ? "nv-" : "", range->readonly ? "rom" : memory_region_type(mr), memory_region_name(mr), range->offset_in_region); } else { qemu_printf(MTREE_INDENT TARGET_FMT_plx "-" TARGET_FMT_plx - " (prio %d, %s%s): %s", + " (%s%s): %s", int128_get64(range->addr.start), int128_get64(range->addr.start) + MR_SIZE(range->addr.size), - mr->priority, range->nonvolatile ? "nv-" : "", range->readonly ? "rom" : memory_region_type(mr), memory_region_name(mr));
When requesting the flatview output with 'info mtree -f', the MemoryRegion priority is irrelevant and noise. Remove it. Signed-off-by: Philippe Mathieu-Daudé <philmd@linaro.org> --- softmmu/memory.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-)