Message ID | 20231107044952.5461-1-anisinha@redhat.com |
---|---|
State | New |
Headers | show |
Series | [v2] tests/acpi/bios-tables-test: do not write new blobs unless there are changes | expand |
> On 07-Nov-2023, at 10:19 AM, Ani Sinha <anisinha@redhat.com> wrote: > > When dumping table blobs using rebuild-expected-aml.sh, table blobs from all > test variants are dumped regardless of whether there are any actual changes to > the tables or not. This creates lot of new files for various test variants that > are not part of the git repository. This is because we do not check in all table > blobs for all test variants into the repository. Only those blobs for those > variants that are different from the generic test-variant agnostic blob are > checked in. > > This change makes the test smarter by checking if at all there are any changes > in the tables from the checked-in gold master blobs and take actions > accordingly. > > When there are no changes: > - No new table blobs would be written. > - Existing table blobs will be refreshed (git diff will show no changes). > When there are changes: > - New table blob files will be dumped. > - Existing table blobs will be refreshed (git diff will show that the files > changed, asl diff will show the actual changes). > When new tables are introduced: > - Zero byte empty file blobs for new tables as instructed in the header of > bios-tables-test.c will be regenerated to actual table blobs. > > This would make analyzing changes to tables less confusing and there would > be no need to clean useless untracked files when there are no table changes. > > CC: peter.maydell@linaro.org > Signed-off-by: Ani Sinha <anisinha@redhat.com> > --- > tests/qtest/bios-tables-test.c | 14 +++++++++++++- > 1 file changed, 13 insertions(+), 1 deletion(-) > > changelog: > v2: commit description updated to make things a little clearer. > No actual changes. Ping ... > > diff --git a/tests/qtest/bios-tables-test.c b/tests/qtest/bios-tables-test.c > index 9f4bc15aab..743b509e93 100644 > --- a/tests/qtest/bios-tables-test.c > +++ b/tests/qtest/bios-tables-test.c > @@ -109,6 +109,7 @@ static const char *iasl; > #endif > > static int verbosity_level; > +static GArray *load_expected_aml(test_data *data); > > static bool compare_signature(const AcpiSdtTable *sdt, const char *signature) > { > @@ -241,21 +242,32 @@ static void test_acpi_fadt_table(test_data *data) > > static void dump_aml_files(test_data *data, bool rebuild) > { > - AcpiSdtTable *sdt; > + AcpiSdtTable *sdt, *exp_sdt; > GError *error = NULL; > gchar *aml_file = NULL; > + test_data exp_data = {}; > gint fd; > ssize_t ret; > int i; > > + exp_data.tables = load_expected_aml(data); > for (i = 0; i < data->tables->len; ++i) { > const char *ext = data->variant ? data->variant : ""; > sdt = &g_array_index(data->tables, AcpiSdtTable, i); > + exp_sdt = &g_array_index(exp_data.tables, AcpiSdtTable, i); > g_assert(sdt->aml); > + g_assert(exp_sdt->aml); > > if (rebuild) { > aml_file = g_strdup_printf("%s/%s/%.4s%s", data_dir, data->machine, > sdt->aml, ext); > + if (!g_file_test(aml_file, G_FILE_TEST_EXISTS) && > + sdt->aml_len == exp_sdt->aml_len && > + !memcmp(sdt->aml, exp_sdt->aml, sdt->aml_len)) { > + /* identical tables, no need to write new files */ > + g_free(aml_file); > + continue; > + } > fd = g_open(aml_file, O_WRONLY|O_TRUNC|O_CREAT, > S_IRUSR|S_IWUSR|S_IRGRP|S_IWGRP|S_IROTH); > if (fd < 0) { > -- > 2.42.0 >
> On 16-Nov-2023, at 11:30 AM, Ani Sinha <anisinha@redhat.com> wrote: > > > >> On 07-Nov-2023, at 10:19 AM, Ani Sinha <anisinha@redhat.com> wrote: >> >> When dumping table blobs using rebuild-expected-aml.sh, table blobs from all >> test variants are dumped regardless of whether there are any actual changes to >> the tables or not. This creates lot of new files for various test variants that >> are not part of the git repository. This is because we do not check in all table >> blobs for all test variants into the repository. Only those blobs for those >> variants that are different from the generic test-variant agnostic blob are >> checked in. >> >> This change makes the test smarter by checking if at all there are any changes >> in the tables from the checked-in gold master blobs and take actions >> accordingly. >> >> When there are no changes: >> - No new table blobs would be written. >> - Existing table blobs will be refreshed (git diff will show no changes). >> When there are changes: >> - New table blob files will be dumped. >> - Existing table blobs will be refreshed (git diff will show that the files >> changed, asl diff will show the actual changes). >> When new tables are introduced: >> - Zero byte empty file blobs for new tables as instructed in the header of >> bios-tables-test.c will be regenerated to actual table blobs. >> >> This would make analyzing changes to tables less confusing and there would >> be no need to clean useless untracked files when there are no table changes. >> >> CC: peter.maydell@linaro.org >> Signed-off-by: Ani Sinha <anisinha@redhat.com> >> --- >> tests/qtest/bios-tables-test.c | 14 +++++++++++++- >> 1 file changed, 13 insertions(+), 1 deletion(-) >> >> changelog: >> v2: commit description updated to make things a little clearer. >> No actual changes. > > Ping ... Ping again ... > >> >> diff --git a/tests/qtest/bios-tables-test.c b/tests/qtest/bios-tables-test.c >> index 9f4bc15aab..743b509e93 100644 >> --- a/tests/qtest/bios-tables-test.c >> +++ b/tests/qtest/bios-tables-test.c >> @@ -109,6 +109,7 @@ static const char *iasl; >> #endif >> >> static int verbosity_level; >> +static GArray *load_expected_aml(test_data *data); >> >> static bool compare_signature(const AcpiSdtTable *sdt, const char *signature) >> { >> @@ -241,21 +242,32 @@ static void test_acpi_fadt_table(test_data *data) >> >> static void dump_aml_files(test_data *data, bool rebuild) >> { >> - AcpiSdtTable *sdt; >> + AcpiSdtTable *sdt, *exp_sdt; >> GError *error = NULL; >> gchar *aml_file = NULL; >> + test_data exp_data = {}; >> gint fd; >> ssize_t ret; >> int i; >> >> + exp_data.tables = load_expected_aml(data); >> for (i = 0; i < data->tables->len; ++i) { >> const char *ext = data->variant ? data->variant : ""; >> sdt = &g_array_index(data->tables, AcpiSdtTable, i); >> + exp_sdt = &g_array_index(exp_data.tables, AcpiSdtTable, i); >> g_assert(sdt->aml); >> + g_assert(exp_sdt->aml); >> >> if (rebuild) { >> aml_file = g_strdup_printf("%s/%s/%.4s%s", data_dir, data->machine, >> sdt->aml, ext); >> + if (!g_file_test(aml_file, G_FILE_TEST_EXISTS) && >> + sdt->aml_len == exp_sdt->aml_len && >> + !memcmp(sdt->aml, exp_sdt->aml, sdt->aml_len)) { >> + /* identical tables, no need to write new files */ >> + g_free(aml_file); >> + continue; >> + } >> fd = g_open(aml_file, O_WRONLY|O_TRUNC|O_CREAT, >> S_IRUSR|S_IWUSR|S_IRGRP|S_IWGRP|S_IROTH); >> if (fd < 0) { >> -- >> 2.42.0
> On 21-Nov-2023, at 12:39 PM, Ani Sinha <anisinha@redhat.com> wrote: > > > >> On 16-Nov-2023, at 11:30 AM, Ani Sinha <anisinha@redhat.com> wrote: >> >> >> >>> On 07-Nov-2023, at 10:19 AM, Ani Sinha <anisinha@redhat.com> wrote: >>> >>> When dumping table blobs using rebuild-expected-aml.sh, table blobs from all >>> test variants are dumped regardless of whether there are any actual changes to >>> the tables or not. This creates lot of new files for various test variants that >>> are not part of the git repository. This is because we do not check in all table >>> blobs for all test variants into the repository. Only those blobs for those >>> variants that are different from the generic test-variant agnostic blob are >>> checked in. >>> >>> This change makes the test smarter by checking if at all there are any changes >>> in the tables from the checked-in gold master blobs and take actions >>> accordingly. >>> >>> When there are no changes: >>> - No new table blobs would be written. >>> - Existing table blobs will be refreshed (git diff will show no changes). >>> When there are changes: >>> - New table blob files will be dumped. >>> - Existing table blobs will be refreshed (git diff will show that the files >>> changed, asl diff will show the actual changes). >>> When new tables are introduced: >>> - Zero byte empty file blobs for new tables as instructed in the header of >>> bios-tables-test.c will be regenerated to actual table blobs. >>> >>> This would make analyzing changes to tables less confusing and there would >>> be no need to clean useless untracked files when there are no table changes. >>> >>> CC: peter.maydell@linaro.org >>> Signed-off-by: Ani Sinha <anisinha@redhat.com> >>> --- >>> tests/qtest/bios-tables-test.c | 14 +++++++++++++- >>> 1 file changed, 13 insertions(+), 1 deletion(-) >>> >>> changelog: >>> v2: commit description updated to make things a little clearer. >>> No actual changes. >> >> Ping ... > > Ping again ... Ping … > >> >>> >>> diff --git a/tests/qtest/bios-tables-test.c b/tests/qtest/bios-tables-test.c >>> index 9f4bc15aab..743b509e93 100644 >>> --- a/tests/qtest/bios-tables-test.c >>> +++ b/tests/qtest/bios-tables-test.c >>> @@ -109,6 +109,7 @@ static const char *iasl; >>> #endif >>> >>> static int verbosity_level; >>> +static GArray *load_expected_aml(test_data *data); >>> >>> static bool compare_signature(const AcpiSdtTable *sdt, const char *signature) >>> { >>> @@ -241,21 +242,32 @@ static void test_acpi_fadt_table(test_data *data) >>> >>> static void dump_aml_files(test_data *data, bool rebuild) >>> { >>> - AcpiSdtTable *sdt; >>> + AcpiSdtTable *sdt, *exp_sdt; >>> GError *error = NULL; >>> gchar *aml_file = NULL; >>> + test_data exp_data = {}; >>> gint fd; >>> ssize_t ret; >>> int i; >>> >>> + exp_data.tables = load_expected_aml(data); >>> for (i = 0; i < data->tables->len; ++i) { >>> const char *ext = data->variant ? data->variant : ""; >>> sdt = &g_array_index(data->tables, AcpiSdtTable, i); >>> + exp_sdt = &g_array_index(exp_data.tables, AcpiSdtTable, i); >>> g_assert(sdt->aml); >>> + g_assert(exp_sdt->aml); >>> >>> if (rebuild) { >>> aml_file = g_strdup_printf("%s/%s/%.4s%s", data_dir, data->machine, >>> sdt->aml, ext); >>> + if (!g_file_test(aml_file, G_FILE_TEST_EXISTS) && >>> + sdt->aml_len == exp_sdt->aml_len && >>> + !memcmp(sdt->aml, exp_sdt->aml, sdt->aml_len)) { >>> + /* identical tables, no need to write new files */ >>> + g_free(aml_file); >>> + continue; >>> + } >>> fd = g_open(aml_file, O_WRONLY|O_TRUNC|O_CREAT, >>> S_IRUSR|S_IWUSR|S_IRGRP|S_IWGRP|S_IROTH); >>> if (fd < 0) { >>> -- >>> 2.42.0
On Tue, Nov 07, 2023 at 10:19:51AM +0530, Ani Sinha wrote: > When dumping table blobs using rebuild-expected-aml.sh, table blobs from all > test variants are dumped regardless of whether there are any actual changes to > the tables or not. This creates lot of new files for various test variants that > are not part of the git repository. This is because we do not check in all table > blobs for all test variants into the repository. Only those blobs for those > variants that are different from the generic test-variant agnostic blob are > checked in. > > This change makes the test smarter by checking if at all there are any changes > in the tables from the checked-in gold master blobs and take actions > accordingly. > > When there are no changes: > - No new table blobs would be written. > - Existing table blobs will be refreshed (git diff will show no changes). > When there are changes: > - New table blob files will be dumped. > - Existing table blobs will be refreshed (git diff will show that the files > changed, asl diff will show the actual changes). > When new tables are introduced: > - Zero byte empty file blobs for new tables as instructed in the header of > bios-tables-test.c will be regenerated to actual table blobs. > > This would make analyzing changes to tables less confusing and there would > be no need to clean useless untracked files when there are no table changes. > > CC: peter.maydell@linaro.org > Signed-off-by: Ani Sinha <anisinha@redhat.com> > --- > tests/qtest/bios-tables-test.c | 14 +++++++++++++- > 1 file changed, 13 insertions(+), 1 deletion(-) > > changelog: > v2: commit description updated to make things a little clearer. > No actual changes. Reviewed-by: Daniel P. Berrangé <berrange@redhat.com> With regards, Daniel
On Tue, 7 Nov 2023 10:19:51 +0530 Ani Sinha <anisinha@redhat.com> wrote: > When dumping table blobs using rebuild-expected-aml.sh, table blobs from all > test variants are dumped regardless of whether there are any actual changes to > the tables or not. This creates lot of new files for various test variants that > are not part of the git repository. This is because we do not check in all table > blobs for all test variants into the repository. Only those blobs for those > variants that are different from the generic test-variant agnostic blob are > checked in. > > This change makes the test smarter by checking if at all there are any changes > in the tables from the checked-in gold master blobs and take actions > accordingly. > > When there are no changes: > - No new table blobs would be written. > - Existing table blobs will be refreshed (git diff will show no changes). > When there are changes: > - New table blob files will be dumped. > - Existing table blobs will be refreshed (git diff will show that the files > changed, asl diff will show the actual changes). > When new tables are introduced: > - Zero byte empty file blobs for new tables as instructed in the header of > bios-tables-test.c will be regenerated to actual table blobs. > > This would make analyzing changes to tables less confusing and there would > be no need to clean useless untracked files when there are no table changes. > > CC: peter.maydell@linaro.org > Signed-off-by: Ani Sinha <anisinha@redhat.com> Acked-by: Igor Mammedov <imammedo@redhat.com> > --- > tests/qtest/bios-tables-test.c | 14 +++++++++++++- > 1 file changed, 13 insertions(+), 1 deletion(-) > > changelog: > v2: commit description updated to make things a little clearer. > No actual changes. > > diff --git a/tests/qtest/bios-tables-test.c b/tests/qtest/bios-tables-test.c > index 9f4bc15aab..743b509e93 100644 > --- a/tests/qtest/bios-tables-test.c > +++ b/tests/qtest/bios-tables-test.c > @@ -109,6 +109,7 @@ static const char *iasl; > #endif > > static int verbosity_level; > +static GArray *load_expected_aml(test_data *data); > > static bool compare_signature(const AcpiSdtTable *sdt, const char *signature) > { > @@ -241,21 +242,32 @@ static void test_acpi_fadt_table(test_data *data) > > static void dump_aml_files(test_data *data, bool rebuild) > { > - AcpiSdtTable *sdt; > + AcpiSdtTable *sdt, *exp_sdt; > GError *error = NULL; > gchar *aml_file = NULL; > + test_data exp_data = {}; > gint fd; > ssize_t ret; > int i; > > + exp_data.tables = load_expected_aml(data); > for (i = 0; i < data->tables->len; ++i) { > const char *ext = data->variant ? data->variant : ""; > sdt = &g_array_index(data->tables, AcpiSdtTable, i); > + exp_sdt = &g_array_index(exp_data.tables, AcpiSdtTable, i); > g_assert(sdt->aml); > + g_assert(exp_sdt->aml); > > if (rebuild) { > aml_file = g_strdup_printf("%s/%s/%.4s%s", data_dir, data->machine, > sdt->aml, ext); > + if (!g_file_test(aml_file, G_FILE_TEST_EXISTS) && > + sdt->aml_len == exp_sdt->aml_len && > + !memcmp(sdt->aml, exp_sdt->aml, sdt->aml_len)) { > + /* identical tables, no need to write new files */ > + g_free(aml_file); > + continue; > + } > fd = g_open(aml_file, O_WRONLY|O_TRUNC|O_CREAT, > S_IRUSR|S_IWUSR|S_IRGRP|S_IWGRP|S_IROTH); > if (fd < 0) {
diff --git a/tests/qtest/bios-tables-test.c b/tests/qtest/bios-tables-test.c index 9f4bc15aab..743b509e93 100644 --- a/tests/qtest/bios-tables-test.c +++ b/tests/qtest/bios-tables-test.c @@ -109,6 +109,7 @@ static const char *iasl; #endif static int verbosity_level; +static GArray *load_expected_aml(test_data *data); static bool compare_signature(const AcpiSdtTable *sdt, const char *signature) { @@ -241,21 +242,32 @@ static void test_acpi_fadt_table(test_data *data) static void dump_aml_files(test_data *data, bool rebuild) { - AcpiSdtTable *sdt; + AcpiSdtTable *sdt, *exp_sdt; GError *error = NULL; gchar *aml_file = NULL; + test_data exp_data = {}; gint fd; ssize_t ret; int i; + exp_data.tables = load_expected_aml(data); for (i = 0; i < data->tables->len; ++i) { const char *ext = data->variant ? data->variant : ""; sdt = &g_array_index(data->tables, AcpiSdtTable, i); + exp_sdt = &g_array_index(exp_data.tables, AcpiSdtTable, i); g_assert(sdt->aml); + g_assert(exp_sdt->aml); if (rebuild) { aml_file = g_strdup_printf("%s/%s/%.4s%s", data_dir, data->machine, sdt->aml, ext); + if (!g_file_test(aml_file, G_FILE_TEST_EXISTS) && + sdt->aml_len == exp_sdt->aml_len && + !memcmp(sdt->aml, exp_sdt->aml, sdt->aml_len)) { + /* identical tables, no need to write new files */ + g_free(aml_file); + continue; + } fd = g_open(aml_file, O_WRONLY|O_TRUNC|O_CREAT, S_IRUSR|S_IWUSR|S_IRGRP|S_IWGRP|S_IROTH); if (fd < 0) {
When dumping table blobs using rebuild-expected-aml.sh, table blobs from all test variants are dumped regardless of whether there are any actual changes to the tables or not. This creates lot of new files for various test variants that are not part of the git repository. This is because we do not check in all table blobs for all test variants into the repository. Only those blobs for those variants that are different from the generic test-variant agnostic blob are checked in. This change makes the test smarter by checking if at all there are any changes in the tables from the checked-in gold master blobs and take actions accordingly. When there are no changes: - No new table blobs would be written. - Existing table blobs will be refreshed (git diff will show no changes). When there are changes: - New table blob files will be dumped. - Existing table blobs will be refreshed (git diff will show that the files changed, asl diff will show the actual changes). When new tables are introduced: - Zero byte empty file blobs for new tables as instructed in the header of bios-tables-test.c will be regenerated to actual table blobs. This would make analyzing changes to tables less confusing and there would be no need to clean useless untracked files when there are no table changes. CC: peter.maydell@linaro.org Signed-off-by: Ani Sinha <anisinha@redhat.com> --- tests/qtest/bios-tables-test.c | 14 +++++++++++++- 1 file changed, 13 insertions(+), 1 deletion(-) changelog: v2: commit description updated to make things a little clearer. No actual changes.