Bug-fix:Segmentation fault with powertop --html
Alexandra Yates
PATCH 1/1 Bug-fix:Segmentation fault with powertop --html
This patch fixes the segmentation fault erros seen in powertop --html.
Reported by: Julian Wollrath" <jwollrath(a)web.de>
Julian can you test the patch please. thank you. AY.
Alexandra Yates (1):
Bug-fix: html & csv report core indexes
src/cpu/cpu.cpp | 30 +++++++++++++++++++++--------- 1 file changed, 21 insertions(+), 9 deletions(-)
Signature · 1 line
1.7.9.5
Alexandra Yates
PATCH 1/1 Bug-fix: html & csv report core indexes
This patch allows the supported no Intel CPUs to be included on html & CSV reports. It fixes the rows index needed for package data. Additionally, fixes the calculation of title indexes when cores are absent from the CPU. These changes are related to C-states and P-states reports
--- src/cpu/cpu.cpp | 30 +++++++++++++++++++++--------- 1 file changed, 21 insertions(+), 9 deletions(-)
src/cpu/cpu.cpp +21 -9
@@ -466,7 +466,7 @@ void report_display_cpu_cstates(void) continue; /* Tables for PKG, CORE, CPU */ pkg_tbl_size.cols=2; - pkg_tbl_size.rows= cstates_num+1; + pkg_tbl_size.rows= ((cstates_num+1)-LEVEL_HEADER)+1; string pkg_data[pkg_tbl_size.cols * pkg_tbl_size.rows]; core_tbl_size.cols=2;
@@ -608,12 +608,18 @@ void report_display_cpu_cstates(void) } /* Report Output */ - title=title/core_num; + if(core_num > 0) + title=title/core_num; + else if( _core->children.size() > 0) + title=title/_core->children.size(); + init_pkg_table_attr(&std_table_css, pkg_tbl_size.rows, pkg_tbl_size.cols); report.add_table(pkg_data, &std_table_css); - init_core_table_attr(&std_table_css, title+1, core_tbl_size.rows, + if (!_core->can_collapse()){ + init_core_table_attr(&std_table_css, title+1, core_tbl_size.rows, core_tbl_size.cols); - report.add_table(core_data, &std_table_css); + report.add_table(core_data, &std_table_css); + } init_cpu_table_attr(&std_table_css, title+1, cpu_tbl_size.rows, cpu_tbl_size.cols); report.add_table(cpu_data, &std_table_css);
@@ -625,7 +631,7 @@ void report_display_cpu_pstates(void) { char buffer[512], buffer2[512], tmp_num[50]; unsigned int package, core, cpu; - int line; + int line, title=0; class abstract_cpu *_package, * _core, * _cpu; unsigned int i, pstates_num; const char* core_type = NULL;
@@ -671,7 +677,7 @@ void report_display_cpu_pstates(void) /* Tables for PKG, CORE, CPU */ pkg_tbl_size.cols=2; - pkg_tbl_size.rows=pstates_num+2; + pkg_tbl_size.rows=((pstates_num+1)-LEVEL_HEADER)+2; string pkg_data[pkg_tbl_size.cols * pkg_tbl_size.rows]; core_tbl_size.cols=2;
@@ -790,10 +796,16 @@ void report_display_cpu_pstates(void) } init_pkg_table_attr(&std_table_css, pkg_tbl_size.rows, pkg_tbl_size.cols); report.add_table(pkg_data, &std_table_css); - init_core_table_attr(&std_table_css, (pstates_num+2), + if(!_core->can_collapse()){ + title=pstates_num+2; + init_core_table_attr(&std_table_css, title, core_tbl_size.rows, core_tbl_size.cols); - report.add_table(core_data, &std_table_css); - init_cpu_table_attr(&std_table_css, (pstates_num+2), + report.add_table(core_data, &std_table_css); + } else { + title=pstates_num+1; + } + + init_cpu_table_attr(&std_table_css, title, cpu_tbl_size.rows, cpu_tbl_size.cols); report.add_table(cpu_data, &std_table_css); }
Signature · 1 line
1.7.9.5
Alexandra Yates
Quoted 7 lines
This patch fixes the segmentation fault erros seen in powertop --html.
Reported by: Julian Wollrath" <jwollrath(a)web.de>
Julian can you test the patch please. thank you. AY.
Quoted 10 lines
Alexandra Yates (1):
Bug-fix: html & csv report core indexessrc/cpu/cpu.cpp | 30 +++++++++++++++++++++--------- 1 file changed, 21 insertions(+), 9 deletions(-)--
1.7.9.5
The patch name is: [PATCH] Bug-fix: html & csv report core indexes.
Thank you,
Alexandra.
Julian Wollrath
Quoted 1 line
Julian can you test the patch please. thank you. AY.
Thank you for the patch but unfortunately powertop still segfaults:
$ powertop --html
Loaded 750 prior measurements
RAPL device for cpu 0
RAPL device for cpu 0
Preparing to take measurements
unknown op '{'
*** Error in `powertop': free(): invalid pointer: 0x000000000187cdef ***
Taking 1 measurement(s) for a duration of 20 second(s) each.
[1] 22616 segmentation fault (core dumped) powertop --htmlBacktrace:
#0 0x00007f63d8c8e1af in std::string::assign(std::string const&) ()
from /usr/lib/x86_64-linux-gnu/libstdc++.so.6
#1 0x0000000000410f31 in operator= (__str="C1", this=0x7fff65b509a0)
at /usr/include/c++/4.8/bits/basic_string.h:547#2 report_display_cpu_cstates () at cpu/cpu.cpp:587
#3 0x00000000004493f4 in one_measurement (seconds=seconds(a)entry=20, workload=workload(a)entry=0x7fff65b521b0 "") at main.cpp:221 #4 0x00000000004496a6 in make_report (time=20, workload=workload(a)entry=0x7fff65b521b0 "", iterations=iterations(a)entry=1, file=file(a)entry=0x7fff65b511b0 "powertop.html") at main.cpp:264 #5 0x000000000040741d in main (argc=2, argv=0x7fff65b532d8) at main.cpp:441
Cheers,
Julian
Alexandra Yates
Quoted 28 lines
Quoted 2 lines
Julian can you test the patch please. thank you. AY.
Thank you for the patch but unfortunately powertop still segfaults:
$ powertop --html Loaded 750 prior measurements RAPL device for cpu 0 RAPL device for cpu 0 Preparing to take measurements unknown op '{' *** Error in `powertop': free(): invalid pointer: 0x000000000187cdef *** Taking 1 measurement(s) for a duration of 20 second(s) each. [1] 22616 segmentation fault (core dumped) powertop --htmlBacktrace: #0 0x00007f63d8c8e1af in std::string::assign(std::string const&) () from /usr/lib/x86_64-linux-gnu/libstdc++.so.6 #1 0x0000000000410f31 in operator= (__str="C1", this=0x7fff65b509a0) at /usr/include/c++/4.8/bits/basic_string.h:547#2
report_display_cpu_cstates() at cpu/cpu.cpp:587
#30x00000000004493f4inone_measurement(seconds=seconds(a)entry=20, workload=workload(a)entry=0x7fff65b521b0 "") at main.cpp:221 #40x00000000004496a6inmake_report(time=20, workload=workload(a)entry=0x7fff65b521b0 "", iterations=iterations(a)entry=1, file=file(a)entry=0x7fff65b511b0 "powertop.html") at main.cpp:264 #50x000000000040741din main (argc=2, argv=0x7fff65b532d8) at main.cpp:441
Quoted 3 lines
Cheers,
Julian
Julian, thank you for testing. The patches I sent you yesterday fixed the previous segfault. I'm currently looking into another solution.
Thank you,
Alexandra.