July 2014 - PowerTOP - Ml01.01.Org

A series of patches towards limiting memory corruption and foot print

by Kalowsky, Daniel

Hi Powertop, I've been working with the interactive mode and have run into cases where powertop will crash due to a series of memory corruptions. The following series of patches have helped to reduce the frequency of the issue, although not completely solved it. The issues arise much faster on platforms where there are constrained amounts of RAM to work within. Patch 1 - When shutting down the interactive display, the display bits of memory is not correctly released. This patch provides a method for correctly doing so. Patch 2 - Solving a documented memory leak with a non-elegant solution. The path either adds the bundle to the stack, or it forgets about it. If it is forgotten about, make sure to clear that memory before moving on. This is done with a simple flag variable being set. Patch 3 - When the tuning window is updated, the current pointer is just set adrift and not properly free'd. This patch catches that issue and removes the dangling pointer by holding a reference to the pointer until it is reset or specifically free'd. Patch 4 - Someone actually added in the code to create a onetime pretty-print array, this patch just puts it to use by setting the variable. Patch 5 - Limiting the buffer copy to the size of the allocated buffer with snprintf. Patch 6 - There exist some processes and entries that can and do extend beyond the length of these buffers. This limits those entries so as not to corrupt other memory on the system when in interactive mode. Patch 7 - This is an untested patch, but follows along the same lines of Patch 6. It applies the same principals only for the report method. Patch 8 - Creates a clean_shutdown function that can be used to cleanup the memory space at shutdown time. Calls upon parts of Patch 1 to make this happen. There will more than likely be some more patches in the future as time permits.

7 years, 11 months

6
40
0 / 0

Re: [Powertop] [PATCH POWERTOP] Remove temporary file created during calibration

by Amit Kucheria

Mohammad, This fix should go upstream. cc'ing the powertop list. Regards, Amit On Thu, Jun 26, 2014 at 12:27 PM, Mohammad Merajul Islam Molla <meraj.enigma(a)gmail.com> wrote: > powertop creates temporary file in /tmp folder during calibration. > Simply closing does not remove these files. > Instead use unlink to remove the file. > > diff --git a/src/calibrate/calibrate.cpp b/src/calibrate/calibrate.cpp > index db368e0..1336c3f 100644 > --- a/src/calibrate/calibrate.cpp > +++ b/src/calibrate/calibrate.cpp > @@ -242,7 +242,7 @@ static void *burn_disk(void *dummy) > write(fd, buffer, 64*1024); > fdatasync(fd); > } > - close(fd); > + unlink(filename); > return NULL; > } > > > > -- > Thanks, > -Meraj > > _______________________________________________ > linaro-dev mailing list > linaro-dev(a)lists.linaro.org > http://lists.linaro.org/mailman/listinfo/linaro-dev

8 years, 1 month

3
3
0 / 0

[PATCH] powertop: update build instructions

by Mohammad Merajul Islam Molla

update powertop build instuctions in README file. Signed-off-by: Mohammad Merajul Islam Molla <meraj.enigma(a)gmail.com> --- README | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/README b/README index cd38cc5..efe95c3 100644 --- a/README +++ b/README @@ -2,9 +2,10 @@ ------------------------------ To build and install PowerTOP type the following commands, - ./configure - ./make - ./make install + autoreconf -sfi + ./configure + make + sudo make install Note: For Android (running Intel Architecture ) there is a Android.mk that was provided by community members, and at this time is supported -- 1.9.1

8 years, 1 month

4
3
0 / 0

How to decrease PS/2 Touchpad / Keyboard / Mouse wakeups

by Marc MERLIN

On a thinkpad T540p, I see: 4.10 W 6.6 ms/s 196.3 Interrupt PS/2 Touchpad / Keyboard / Mouse Even if the 4W figure is wrong, 200 wakeups for a touchpad and keyboard seems totally ridiculous. What can be done to decrease those? Thanks, Marc -- "A mouse is a device used to point at the xterm you want to type in" - A.S.R. Microsoft is to operating systems .... .... what McDonalds is to gourmet cooking Home page: http://marc.merlins.org/ | PGP 1024R/763BE901

8 years, 1 month

1
2
0 / 0

not getting the power consumption numbers

by rahul shrivastava

Hi, While installing version 2.6.1, I got the following error - make[2]: Entering directory `/home/rahul/powertop/powertop-2.6.1/src' /bin/bash ./csstoh.sh powertop.css css.h /bin/bash: ./csstoh.sh: No such file or directory So I installed version 2.5 and it compiled correctly. But, I am not getting the power consumption numbers. What I am getting is this * Usage Events/s Category Description* 66.7 us/s 26.5 Process [rcu_sched] 152.8 us/s 18.7 Interrupt [3] net_rx(softirq) 21.7 us/s 18.7 kWork od_dbs_timer 367.6 us/s 11.8 Process avahi-daemon: running [Ramanujan.local] 33.6 us/s 9.8 Timer tick_sched_timer 7.0 ms/s 2.0 Process powertop 299.8 us/s 2.0 Process /usr/sbin/irqbalance 50.3 us/s 2.0 Process /usr/lib/accountsservice/accounts-daemon 39.7 us/s 2.0 Process nmbd -D I am not getting any "power estimation" field. I am using ubuntu 12.04 with linux kernel version 3.11.0-24-generic Could you please let me know what could be wrong. Regards Rahul

8 years, 2 months

3
4
0 / 0

Re: [Powertop] [PATCH POWERTOP] Fix various resource leaks

by Amit Kucheria

Mohammad, This fix should go upstream. cc'ing the powertop list. Regards, Amit On Thu, Jun 26, 2014 at 12:42 PM, Mohammad Merajul Islam Molla <meraj.enigma(a)gmail.com> wrote: > Fixes some resource leaks detected by valgrind and coverity scan. > > > diff --git a/src/devices/ahci.cpp b/src/devices/ahci.cpp > index ac06460..7f704b6 100644 > --- a/src/devices/ahci.cpp > +++ b/src/devices/ahci.cpp > @@ -64,8 +64,10 @@ static string disk_name(char *path, char *target, > char *shortname) > sprintf(line, "%s/%s/model", pathname, dirent->d_name); > file = fopen(line, "r"); > if (file) { > - if (fgets(line, 4096, file) == NULL) > + if (fgets(line, 4096, file) == NULL) { > + fclose(file); > break; > + } > fclose(file); > c = strchr(line, '\n'); > if (c) > diff --git a/src/devices/devfreq.cpp b/src/devices/devfreq.cpp > index e16951c..23c4b0c 100644 > --- a/src/devices/devfreq.cpp > +++ b/src/devices/devfreq.cpp > @@ -238,6 +238,7 @@ void create_all_devfreq_devices(void) > > callback fn = &devfreq_dev_callback; > process_directory(p.c_str(), fn); > + closedir(dir); > } > > void initialize_devfreq(void) > diff --git a/src/perf/perf_bundle.cpp b/src/perf/perf_bundle.cpp > index b0e982b..cf1ae11 100644 > --- a/src/perf/perf_bundle.cpp > +++ b/src/perf/perf_bundle.cpp > @@ -142,8 +142,10 @@ static void parse_event_format(const char *event_name) > > buf = read_file(file); > free(file); > - if (!buf) > + if (!buf) { > + free(name); > return; > + } > > pevent_parse_event(perf_event::pevent, buf, strlen(buf), sys); > free(name); > diff --git a/src/tuning/bluetooth.cpp b/src/tuning/bluetooth.cpp > index e0bdf12..5100a8a 100644 > --- a/src/tuning/bluetooth.cpp > +++ b/src/tuning/bluetooth.cpp > @@ -144,8 +144,10 @@ int bt_tunable::good_bad(void) > if (file) { > char line[2048]; > /* first line is standard header */ > - if (fgets(line, 2047, file) == NULL) > + if (fgets(line, 2047, file) == NULL) { > + pclose(file); > goto out; > + } > memset(line, 0, 2048); > if (fgets(line, 2047, file) == NULL) { > result = last_check_result = TUNE_GOOD; > > > > -- > Thanks, > -Meraj > > _______________________________________________ > linaro-dev mailing list > linaro-dev(a)lists.linaro.org > http://lists.linaro.org/mailman/listinfo/linaro-dev >

8 years, 2 months

3
2
0 / 0

Requesting help to get the power consumption of each thread or process id

by Christopher Mansour

Hi, I was wondering if somebody can help in doing the following:I want to get the power consumption of each thread running. I am doing an experiment. I built a program that executes on several threads and I need to know the power consumed by each thread or by each core of the eight CPU cores I have. I do know the process ID of the processes that start but I don't know how to get their power consumption using powertop. I appreciate somebody's help. Regards, Christopher Mansour

8 years, 2 months

2
1
0 / 0

[PATCH] Powertop: remove unnecessary code

by Salman Ahmed

nr_open is already initialized to NR_OPEN_DEF. Removing unnecessary checking and assignment. Signed-off-by: Salman Ahmed <salman0yam(a)gmail.com> --- src/main.cpp | 2 -- 1 file changed, 2 deletions(-) diff --git a/src/main.cpp b/src/main.cpp index cf4e547..990bc1f 100644 --- a/src/main.cpp +++ b/src/main.cpp @@ -293,8 +293,6 @@ static int get_nr_open(void) { file.open("/proc/sys/fs/nr_open", ios::in); if (file) { file >> nr_open; - if (!file) - nr_open = NR_OPEN_DEF; file.close(); } return nr_open; -- 1.8.1.2 -- Thanks, Salman Ahmed

8 years, 2 months

1
0
0 / 0

[PATCH UPDATED] powertop: update build instructions

by Mohammad Merajul Islam Molla

update powertop build instuctions in README file. Signed-off-by: Mohammad Merajul Islam Molla <meraj.enigma(a)gmail.com> --- README | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/README b/README index cd38cc5..bc48d3d 100644 --- a/README +++ b/README @@ -2,9 +2,10 @@ ------------------------------ To build and install PowerTOP type the following commands, + ./autogen.sh ./configure - ./make - ./make install + make + sudo make install Note: For Android (running Intel Architecture ) there is a Android.mk that was provided by community members, and at this time is supported -- 1.9.1

8 years, 2 months

4
3
0 / 0

Regarding powertop VFS ops/sec output

by Mohammad Merajul Islam Molla

Hello, It looks like powertop requires a kernel patch to display VFS ops/sec. The last patch was for kernel v3.3.0. The patch needs to be updated, otherwise powertop always displays 0 VFS ops/sec. Considering changes starting with kernel v3.9, the kernel side patch would look like as below. I have tested it with kernel v3.16.0-rc4 and with this patch powertop displays and updates VFS ops/sec. diff --git a/fs/fs-writeback.c b/fs/fs-writeback.c index be568b7..2bcd6d5 100644 --- a/fs/fs-writeback.c +++ b/fs/fs-writeback.c @@ -1156,6 +1156,9 @@ void __mark_inode_dirty(struct inode *inode, int flags) if ((inode->i_state & flags) == flags) return; + if (flags & (I_DIRTY_SYNC | I_DIRTY_DATASYNC | I_DIRTY_PAGES)) + trace_writeback_inode_dirty(inode, flags); + if (unlikely(block_dump)) block_dump___mark_inode_dirty(inode); diff --git a/include/trace/events/writeback.h b/include/trace/events/writeback.h index cee02d6..c1b1a95 100644 --- a/include/trace/events/writeback.h +++ b/include/trace/events/writeback.h @@ -102,6 +102,35 @@ DEFINE_EVENT(writeback_dirty_inode_template, writeback_dirty_inode, TP_ARGS(inode, flags) ); +/* + * Tracepoint for dirtying an inode; used by PowerTOP + */ +TRACE_EVENT(writeback_inode_dirty, + + TP_PROTO(struct inode *inode, int flags), + + TP_ARGS(inode, flags), + + TP_STRUCT__entry( + __field( __kernel_dev_t, dev ) + __field( ino_t, ino ) + __field( u32, flags ) + ), + + TP_fast_assign( + __entry->dev = inode->i_sb->s_dev; + __entry->ino = inode->i_ino; + __entry->flags = flags; + ), + + TP_printk("dev %d:%d ino %lu flags %d %s", MAJOR(__entry->dev), + MINOR(__entry->dev), + (unsigned long) __entry->ino, + __entry->flags, + show_inode_state(__entry->flags) + ) +); + DECLARE_EVENT_CLASS(writeback_write_inode_template, TP_PROTO(struct inode *inode, struct writeback_control *wbc), However, there is an alternative thought. Starting with kernel v3.9, kernel provides a built in trace event called - "writeback_dirty_inode" - DEFINE_EVENT(writeback_dirty_inode_template, writeback_dirty_inode, TP_PROTO(struct inode *inode, int flags), TP_ARGS(inode, flags) ); But since the traceevent powertop expects is "writeback_inode_dirty" and the names and output format don't match, it would require changes in powertop; some kernel side patch may be needed too to account for DIRTY_PAGES. Any comment or thought on this? -- Thanks, -Meraj

8 years, 2 months

1
0
0 / 0

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

PowerTOP July 2014