* [PATCH v2] softmmu/vl.c: fix too slow TCG regression
@ 2020-02-27 16:14 Igor Mammedov
2020-02-27 16:35 ` Philippe Mathieu-Daudé
` (2 more replies)
0 siblings, 3 replies; 4+ messages in thread
From: Igor Mammedov @ 2020-02-27 16:14 UTC (permalink / raw)
To: qemu-devel
Cc: peter.maydell, f4bug, nieklinnenbank, hsp.cat7, pbonzini,
alex.bennee, rth
Commit a1b18df9a4 moved -m option parsing after configure_accelerators()
that broke TCG accelerator initialization which accesses global ram_size
from size_code_gen_buffer() which is equal to 0 at that moment.
Partially revert a1b18df9a4, by returning set_memory_options() to its
original location and only keep 32-bit host VA check and 'memory-backend'
size check introduced by fe64d06afc at current place.
Signed-off-by: Igor Mammedov <imammedo@redhat.com>
---
v2:
* check if user used -m SZ before comparing ram_size with backend's size.
Fixes case where qemu exits with error when backend's size doesn't match
dafault ram_size (i.e. user haven't used -m option at all)
(Peter Maydell <peter.maydell@linaro.org>)
* keep order addr space size check for 32-bit host after
backend size check. So it would be affected by backend's size as well.
---
softmmu/vl.c | 54 ++++++++++++++++++++++++++++------------------------
1 file changed, 29 insertions(+), 25 deletions(-)
diff --git a/softmmu/vl.c b/softmmu/vl.c
index a9cce78f45..16ff5a16a3 100644
--- a/softmmu/vl.c
+++ b/softmmu/vl.c
@@ -2558,7 +2558,7 @@ static bool object_create_delayed(const char *type, QemuOpts *opts)
}
-static void set_memory_options(uint64_t *ram_slots, ram_addr_t *maxram_size,
+static bool set_memory_options(uint64_t *ram_slots, ram_addr_t *maxram_size,
MachineClass *mc)
{
uint64_t sz;
@@ -2634,30 +2634,8 @@ static void set_memory_options(uint64_t *ram_slots, ram_addr_t *maxram_size,
exit(EXIT_FAILURE);
}
- if (current_machine->ram_memdev_id) {
- Object *backend;
- ram_addr_t backend_size;
-
- backend = object_resolve_path_type(current_machine->ram_memdev_id,
- TYPE_MEMORY_BACKEND, NULL);
- backend_size = object_property_get_uint(backend, "size", &error_abort);
- if (mem_str && backend_size != ram_size) {
- error_report("Size specified by -m option must match size of "
- "explicitly specified 'memory-backend' property");
- exit(EXIT_FAILURE);
- }
- ram_size = backend_size;
- }
-
- if (!xen_enabled()) {
- /* On 32-bit hosts, QEMU is limited by virtual address space */
- if (ram_size > (2047 << 20) && HOST_LONG_BITS == 32) {
- error_report("at most 2047 MB RAM can be simulated");
- exit(1);
- }
- }
-
loc_pop(&loc);
+ return !!mem_str;
}
static int global_init_func(void *opaque, QemuOpts *opts, Error **errp)
@@ -2861,6 +2839,7 @@ void qemu_init(int argc, char **argv, char **envp)
bool list_data_dirs = false;
char *dir, **dirs;
const char *mem_path = NULL;
+ bool have_custom_ram_size;
BlockdevOptionsQueue bdo_queue = QSIMPLEQ_HEAD_INITIALIZER(bdo_queue);
QemuPluginList plugin_list = QTAILQ_HEAD_INITIALIZER(plugin_list);
int mem_prealloc = 0; /* force preallocation of physical target memory */
@@ -3821,6 +3800,9 @@ void qemu_init(int argc, char **argv, char **envp)
machine_class = select_machine();
object_set_machine_compat_props(machine_class->compat_props);
+ have_custom_ram_size = set_memory_options(&ram_slots, &maxram_size,
+ machine_class);
+
os_daemonize();
/*
@@ -4296,7 +4278,29 @@ void qemu_init(int argc, char **argv, char **envp)
current_machine->cpu_type = parse_cpu_option(cpu_option);
}
- set_memory_options(&ram_slots, &maxram_size, machine_class);
+ if (current_machine->ram_memdev_id) {
+ Object *backend;
+ ram_addr_t backend_size;
+
+ backend = object_resolve_path_type(current_machine->ram_memdev_id,
+ TYPE_MEMORY_BACKEND, NULL);
+ backend_size = object_property_get_uint(backend, "size", &error_abort);
+ if (have_custom_ram_size && backend_size != ram_size) {
+ error_report("Size specified by -m option must match size of "
+ "explicitly specified 'memory-backend' property");
+ exit(EXIT_FAILURE);
+ }
+ ram_size = backend_size;
+ }
+
+ if (!xen_enabled()) {
+ /* On 32-bit hosts, QEMU is limited by virtual address space */
+ if (ram_size > (2047 << 20) && HOST_LONG_BITS == 32) {
+ error_report("at most 2047 MB RAM can be simulated");
+ exit(1);
+ }
+ }
+
current_machine->ram_size = ram_size;
current_machine->maxram_size = maxram_size;
current_machine->ram_slots = ram_slots;
--
2.18.1
^ permalink raw reply related [flat|nested] 4+ messages in thread* Re: [PATCH v2] softmmu/vl.c: fix too slow TCG regression
2020-02-27 16:14 [PATCH v2] softmmu/vl.c: fix too slow TCG regression Igor Mammedov
@ 2020-02-27 16:35 ` Philippe Mathieu-Daudé
2020-02-27 16:37 ` Philippe Mathieu-Daudé
2020-02-27 17:12 ` Peter Maydell
2 siblings, 0 replies; 4+ messages in thread
From: Philippe Mathieu-Daudé @ 2020-02-27 16:35 UTC (permalink / raw)
To: Igor Mammedov, qemu-devel
Cc: peter.maydell, f4bug, nieklinnenbank, hsp.cat7, pbonzini,
alex.bennee, rth
Can we improve the patch subject? "vl: Fix slower TCG on softmmu" or better?
On 2/27/20 5:14 PM, Igor Mammedov wrote:
> Commit a1b18df9a4 moved -m option parsing after configure_accelerators()
> that broke TCG accelerator initialization which accesses global ram_size
> from size_code_gen_buffer() which is equal to 0 at that moment.
>
> Partially revert a1b18df9a4, by returning set_memory_options() to its
> original location and only keep 32-bit host VA check and 'memory-backend'
> size check introduced by fe64d06afc at current place.
>
> Signed-off-by: Igor Mammedov <imammedo@redhat.com>
Tested-by: Philippe Mathieu-Daudé <philmd@redhat.com>
> ---
> v2:
> * check if user used -m SZ before comparing ram_size with backend's size.
> Fixes case where qemu exits with error when backend's size doesn't match
> dafault ram_size (i.e. user haven't used -m option at all)
> (Peter Maydell <peter.maydell@linaro.org>)
> * keep order addr space size check for 32-bit host after
> backend size check. So it would be affected by backend's size as well.
> ---
> softmmu/vl.c | 54 ++++++++++++++++++++++++++++------------------------
> 1 file changed, 29 insertions(+), 25 deletions(-)
>
> diff --git a/softmmu/vl.c b/softmmu/vl.c
> index a9cce78f45..16ff5a16a3 100644
> --- a/softmmu/vl.c
> +++ b/softmmu/vl.c
> @@ -2558,7 +2558,7 @@ static bool object_create_delayed(const char *type, QemuOpts *opts)
> }
>
>
> -static void set_memory_options(uint64_t *ram_slots, ram_addr_t *maxram_size,
> +static bool set_memory_options(uint64_t *ram_slots, ram_addr_t *maxram_size,
> MachineClass *mc)
> {
> uint64_t sz;
> @@ -2634,30 +2634,8 @@ static void set_memory_options(uint64_t *ram_slots, ram_addr_t *maxram_size,
> exit(EXIT_FAILURE);
> }
>
> - if (current_machine->ram_memdev_id) {
> - Object *backend;
> - ram_addr_t backend_size;
> -
> - backend = object_resolve_path_type(current_machine->ram_memdev_id,
> - TYPE_MEMORY_BACKEND, NULL);
> - backend_size = object_property_get_uint(backend, "size", &error_abort);
> - if (mem_str && backend_size != ram_size) {
> - error_report("Size specified by -m option must match size of "
> - "explicitly specified 'memory-backend' property");
> - exit(EXIT_FAILURE);
> - }
> - ram_size = backend_size;
> - }
> -
> - if (!xen_enabled()) {
> - /* On 32-bit hosts, QEMU is limited by virtual address space */
> - if (ram_size > (2047 << 20) && HOST_LONG_BITS == 32) {
> - error_report("at most 2047 MB RAM can be simulated");
> - exit(1);
> - }
> - }
> -
> loc_pop(&loc);
> + return !!mem_str;
> }
>
> static int global_init_func(void *opaque, QemuOpts *opts, Error **errp)
> @@ -2861,6 +2839,7 @@ void qemu_init(int argc, char **argv, char **envp)
> bool list_data_dirs = false;
> char *dir, **dirs;
> const char *mem_path = NULL;
> + bool have_custom_ram_size;
> BlockdevOptionsQueue bdo_queue = QSIMPLEQ_HEAD_INITIALIZER(bdo_queue);
> QemuPluginList plugin_list = QTAILQ_HEAD_INITIALIZER(plugin_list);
> int mem_prealloc = 0; /* force preallocation of physical target memory */
> @@ -3821,6 +3800,9 @@ void qemu_init(int argc, char **argv, char **envp)
> machine_class = select_machine();
> object_set_machine_compat_props(machine_class->compat_props);
>
> + have_custom_ram_size = set_memory_options(&ram_slots, &maxram_size,
> + machine_class);
> +
> os_daemonize();
>
> /*
> @@ -4296,7 +4278,29 @@ void qemu_init(int argc, char **argv, char **envp)
> current_machine->cpu_type = parse_cpu_option(cpu_option);
> }
>
> - set_memory_options(&ram_slots, &maxram_size, machine_class);
> + if (current_machine->ram_memdev_id) {
> + Object *backend;
> + ram_addr_t backend_size;
> +
> + backend = object_resolve_path_type(current_machine->ram_memdev_id,
> + TYPE_MEMORY_BACKEND, NULL);
> + backend_size = object_property_get_uint(backend, "size", &error_abort);
> + if (have_custom_ram_size && backend_size != ram_size) {
> + error_report("Size specified by -m option must match size of "
> + "explicitly specified 'memory-backend' property");
> + exit(EXIT_FAILURE);
> + }
> + ram_size = backend_size;
> + }
> +
> + if (!xen_enabled()) {
> + /* On 32-bit hosts, QEMU is limited by virtual address space */
> + if (ram_size > (2047 << 20) && HOST_LONG_BITS == 32) {
> + error_report("at most 2047 MB RAM can be simulated");
> + exit(1);
> + }
> + }
> +
> current_machine->ram_size = ram_size;
> current_machine->maxram_size = maxram_size;
> current_machine->ram_slots = ram_slots;
>
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH v2] softmmu/vl.c: fix too slow TCG regression
2020-02-27 16:14 [PATCH v2] softmmu/vl.c: fix too slow TCG regression Igor Mammedov
2020-02-27 16:35 ` Philippe Mathieu-Daudé
@ 2020-02-27 16:37 ` Philippe Mathieu-Daudé
2020-02-27 17:12 ` Peter Maydell
2 siblings, 0 replies; 4+ messages in thread
From: Philippe Mathieu-Daudé @ 2020-02-27 16:37 UTC (permalink / raw)
To: Igor Mammedov, qemu-devel
Cc: peter.maydell, f4bug, nieklinnenbank, hsp.cat7, pbonzini,
alex.bennee, rth
On 2/27/20 5:14 PM, Igor Mammedov wrote:
> Commit a1b18df9a4 moved -m option parsing after configure_accelerators()
> that broke TCG accelerator initialization which accesses global ram_size
> from size_code_gen_buffer() which is equal to 0 at that moment.
>
> Partially revert a1b18df9a4, by returning set_memory_options() to its
> original location and only keep 32-bit host VA check and 'memory-backend'
> size check introduced by fe64d06afc at current place.
>
> Signed-off-by: Igor Mammedov <imammedo@redhat.com>
> ---
> v2:
> * check if user used -m SZ before comparing ram_size with backend's size.
> Fixes case where qemu exits with error when backend's size doesn't match
> dafault ram_size (i.e. user haven't used -m option at all)
> (Peter Maydell <peter.maydell@linaro.org>)
> * keep order addr space size check for 32-bit host after
> backend size check. So it would be affected by backend's size as well.
> ---
> softmmu/vl.c | 54 ++++++++++++++++++++++++++++------------------------
> 1 file changed, 29 insertions(+), 25 deletions(-)
>
> diff --git a/softmmu/vl.c b/softmmu/vl.c
> index a9cce78f45..16ff5a16a3 100644
> --- a/softmmu/vl.c
> +++ b/softmmu/vl.c
> @@ -2558,7 +2558,7 @@ static bool object_create_delayed(const char *type, QemuOpts *opts)
> }
>
>
> -static void set_memory_options(uint64_t *ram_slots, ram_addr_t *maxram_size,
> +static bool set_memory_options(uint64_t *ram_slots, ram_addr_t *maxram_size,
> MachineClass *mc)
> {
> uint64_t sz;
> @@ -2634,30 +2634,8 @@ static void set_memory_options(uint64_t *ram_slots, ram_addr_t *maxram_size,
> exit(EXIT_FAILURE);
> }
>
> - if (current_machine->ram_memdev_id) {
> - Object *backend;
> - ram_addr_t backend_size;
> -
> - backend = object_resolve_path_type(current_machine->ram_memdev_id,
> - TYPE_MEMORY_BACKEND, NULL);
> - backend_size = object_property_get_uint(backend, "size", &error_abort);
> - if (mem_str && backend_size != ram_size) {
> - error_report("Size specified by -m option must match size of "
> - "explicitly specified 'memory-backend' property");
> - exit(EXIT_FAILURE);
> - }
> - ram_size = backend_size;
> - }
> -
> - if (!xen_enabled()) {
> - /* On 32-bit hosts, QEMU is limited by virtual address space */
> - if (ram_size > (2047 << 20) && HOST_LONG_BITS == 32) {
> - error_report("at most 2047 MB RAM can be simulated");
> - exit(1);
> - }
> - }
> -
> loc_pop(&loc);
> + return !!mem_str;
> }
>
> static int global_init_func(void *opaque, QemuOpts *opts, Error **errp)
> @@ -2861,6 +2839,7 @@ void qemu_init(int argc, char **argv, char **envp)
> bool list_data_dirs = false;
> char *dir, **dirs;
> const char *mem_path = NULL;
> + bool have_custom_ram_size;
> BlockdevOptionsQueue bdo_queue = QSIMPLEQ_HEAD_INITIALIZER(bdo_queue);
> QemuPluginList plugin_list = QTAILQ_HEAD_INITIALIZER(plugin_list);
> int mem_prealloc = 0; /* force preallocation of physical target memory */
> @@ -3821,6 +3800,9 @@ void qemu_init(int argc, char **argv, char **envp)
> machine_class = select_machine();
> object_set_machine_compat_props(machine_class->compat_props);
>
> + have_custom_ram_size = set_memory_options(&ram_slots, &maxram_size,
> + machine_class);
> +
> os_daemonize();
>
> /*
> @@ -4296,7 +4278,29 @@ void qemu_init(int argc, char **argv, char **envp)
> current_machine->cpu_type = parse_cpu_option(cpu_option);
> }
>
> - set_memory_options(&ram_slots, &maxram_size, machine_class);
> + if (current_machine->ram_memdev_id) {
> + Object *backend;
> + ram_addr_t backend_size;
> +
> + backend = object_resolve_path_type(current_machine->ram_memdev_id,
> + TYPE_MEMORY_BACKEND, NULL);
> + backend_size = object_property_get_uint(backend, "size", &error_abort);
> + if (have_custom_ram_size && backend_size != ram_size) {
> + error_report("Size specified by -m option must match size of "
> + "explicitly specified 'memory-backend' property");
> + exit(EXIT_FAILURE);
> + }
> + ram_size = backend_size;
> + }
> +
> + if (!xen_enabled()) {
> + /* On 32-bit hosts, QEMU is limited by virtual address space */
> + if (ram_size > (2047 << 20) && HOST_LONG_BITS == 32) {
> + error_report("at most 2047 MB RAM can be simulated");
> + exit(1);
> + }
> + }
> +
> current_machine->ram_size = ram_size;
> current_machine->maxram_size = maxram_size;
> current_machine->ram_slots = ram_slots;
>
Reviewed-by: Philippe Mathieu-Daudé <philmd@redhat.com>
^ permalink raw reply [flat|nested] 4+ messages in thread* Re: [PATCH v2] softmmu/vl.c: fix too slow TCG regression
2020-02-27 16:14 [PATCH v2] softmmu/vl.c: fix too slow TCG regression Igor Mammedov
2020-02-27 16:35 ` Philippe Mathieu-Daudé
2020-02-27 16:37 ` Philippe Mathieu-Daudé
@ 2020-02-27 17:12 ` Peter Maydell
2 siblings, 0 replies; 4+ messages in thread
From: Peter Maydell @ 2020-02-27 17:12 UTC (permalink / raw)
To: Igor Mammedov
Cc: QEMU Developers, Philippe Mathieu-Daudé, Niek Linnenbank,
Howard Spoelstra, Paolo Bonzini, Alex Bennée,
Richard Henderson
On Thu, 27 Feb 2020 at 16:15, Igor Mammedov <imammedo@redhat.com> wrote:
>
> Commit a1b18df9a4 moved -m option parsing after configure_accelerators()
> that broke TCG accelerator initialization which accesses global ram_size
> from size_code_gen_buffer() which is equal to 0 at that moment.
>
> Partially revert a1b18df9a4, by returning set_memory_options() to its
> original location and only keep 32-bit host VA check and 'memory-backend'
> size check introduced by fe64d06afc at current place.
>
> Signed-off-by: Igor Mammedov <imammedo@redhat.com>
> ---
> v2:
> * check if user used -m SZ before comparing ram_size with backend's size.
> Fixes case where qemu exits with error when backend's size doesn't match
> dafault ram_size (i.e. user haven't used -m option at all)
> (Peter Maydell <peter.maydell@linaro.org>)
> * keep order addr space size check for 32-bit host after
> backend size check. So it would be affected by backend's size as well.
Applied to master as a fix for the tcg perf regression, thanks.
-- PMM
^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2020-02-27 17:13 UTC | newest]
Thread overview: 4+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2020-02-27 16:14 [PATCH v2] softmmu/vl.c: fix too slow TCG regression Igor Mammedov
2020-02-27 16:35 ` Philippe Mathieu-Daudé
2020-02-27 16:37 ` Philippe Mathieu-Daudé
2020-02-27 17:12 ` Peter Maydell
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).