[DEVS & EXP-USERS ONLY] HTC One X - High I/O wait. Same issue as the N7. (confirmed)

Search This thread

show-p1984

Inactive Recognized Developer
Jul 30, 2011
3,700
11,805
[DEVS & EXP-USERS ONLY] HTC One X - High I/O wait. Same issue as the N7. (confirmed)

This little thread aims at getting some information about an issue I thought for months I am alone with.
Since I got my One X, stock or running custom stuff, my phone always gets sluggish far far beyond the point of being useable if I fill my /data partition with too much data (~1GB free triggers it usually), or move large files around in it (dd'ing some zeros onto it triggers it too).
No kind of rebooting, cache wiping, or anything else will fix it permanently. The lags will be back after a very short while once you experienced them.
Vodafone told me to RMA the phone (it was only software branded), but If you know me, then you know that I am lazy in these kinds of things and since I thought I was the only one and a wipe of /data fixed it every time for a short period, I didn't do it.
Yeah I know. Lazy b*tch. :p

A few hours ago 'Aux' who had complained about a very sluggish phone a few times already joined IRC and described his problem in more detail and I suddenly made the connection in my mind between my issue and his. He then allowed me to debug that stuff, and I came up with the exact same issue I am facing continuously:
High I/O waits while writing/reading from /data.

This thread aims at gathering intel on this issue. Do not answer here with pointless questions or howto...? stuff.
Ah btw: No, one lag after unlocking the phone is not the issue I am talking about. I mean: ~1-10 fps on the launcher, app opening takes minutes (games) & still lags like hell and downloading apps in the market is close to impossible. (~10byte p second)
So if you don't experience this, don't post logs. I am serious, I will report anyone who wastes my time on that.


To check if your phone suffers from this, run 'top'
if it shows something like this:
Code:
User 2%, System 1%, [COLOR="Red"][B]IOW 74%[/B][/COLOR], IRQ 0%
User 34 + Nice 0 + Sys 18 + Idle 263 + IOW 902 + IRQ 0 + SIRQ 1 = 1218

  PID PR CPU% S  #THR     VSS     RSS PCY UID      Name
 2061  0   1% S    13 471140K  32956K  fg u0_a75   com.google.android.apps.reader
  519  1   0% S    73 559264K  58024K  fg system   system_server
 1035  1   0% S    20 512864K  63100K  fg u0_a86   com.anddoes.launcher
 2265  1   0% R     1   1132K    500K     root     top
[COLOR="red"][B]  105  0   0% D     1      0K      0K     root     mmcqd/0[/B][/COLOR]
  898  1   0% S    14 483120K  60344K  fg u0_a39   com.android.systemui
  882  0   0% S     1   2768K   1488K     wifi     /system/bin/wpa_supplicant
  435  0   0% S     1      0K      0K     root     kworker/0:2
  173  1   0% S    11  29644K  10104K  fg system   /system/bin/surfaceflinger
[B][COLOR="red"]  123  1   0% D     1      0K      0K     root     jbd2/mmcblk0p15[/COLOR][/B]
  879  0   0% S     1      0K      0K     root     irq/340-wl12xx

run 'vmstat', if it looks like this
Code:
procs  memory                       system          cpu
 r  b    free mapped   anon   slab    in   cs  flt  us ni sy id wa ir
 0  [COLOR="Red"]3[/COLOR]   29488  64232 261024  45956  1788 2723    0  22  0 29 99 [COLOR="red"]96[/COLOR]  0
 0  [COLOR="red"]2[/COLOR]   29488  64284 261056  45952   620  905    0   1  0  4 99 [COLOR="red"]99[/COLOR]  0
 1  [COLOR="red"]6[/COLOR]   29556  64284 261056  45908   472  755    0   1  0  4 99 [COLOR="red"]99[/COLOR]  0
 3  [COLOR="red"]3[/COLOR]   29496  64284 261056  45908   454  680    0   0  0  1 99 [COLOR="red"]99[/COLOR]  0
 0  [COLOR="red"]8[/COLOR]   29436  64284 261056  45904   462  767    0   4  0  4 99 [COLOR="red"]98[/COLOR]  0
 3  [COLOR="red"]4[/COLOR]   31300  64328 261328  45840  2177 4151    0  50  0 15 99 [COLOR="red"]92[/COLOR]  0

and if 'cat /proc/fs/jbd2/mmcblk0p15-8/info' looks like this too:
Code:
cat /proc/fs/jbd2/mmcblk0p15-8/info
649 transaction, each up to 2196 blocks
average:
  0ms waiting for transaction
  1460ms running transaction
[B][COLOR="Red"]  16544590ms transaction was being locked[/COLOR][/B]
  0ms flushing data (in ordered mode)
  710ms logging transaction
  603582us average transaction commit time
  100 handles per transaction
  7 blocks per transaction
  8 logged blocks per transaction
Then you are most likely affected.


What I need from you:
Everything the above commands throw out as well as the following:
Code:
echo 1 > /proc/sys/vm/block_dump ; cat /proc/kmsg ; echo 0 > /proc/sys/vm/block_dump
(this is very spammy, you might want to add '> file.log' to the kmsg part)
-----
iostat
-----
iostat -kd 5
(keep it running for about 30 seconds)
-----
cat /proc/meminfo
-----
cat /proc/diskstats

This issue is the same I am facing on my N7. ~2.5Gb free will trigger a very poor performance on the N7 as well, so we have found a general bug in one of the used components (hardware if it's the same, the jbd journaling system, ... )

To figure out what this is exactly I need as much intel as you can gather if your device is in such a state.
I will attach my findings, this thread and all logs gathered on the already existing bug report for the N7 @google. (since it is more likely to get an answer out of them than it is to get one from HTC, sadly)

If you have a N7 facing the same issue:
Attach the logs here too, but please specify that they are from a N7.
 
Last edited:

leventccc

Senior Member
Jan 31, 2011
797
430
24
Istanbul
It didn't happened to me while using my htc one x (probably because I don't fill the data partition). But I'm sure that low storage makes my nexus 7 slower and laggier(tested it). I hope you can find a solution to that. I will post logs tomorrow because I don't have time right now.
 

AuxLV

Senior Member
Apr 30, 2009
435
729
38
Riga
Temporary fix found:

1. Reboot to CWM.
2. Nandroid backup.
3. Factory reset/format data.
4. Nandroid restore.

And HOX back to normal! You can write lots more of data! Incredible!

It would be nice if guys with N7 tested.
 
  • Like
Reactions: borekon

Tabtoub

Senior Member
Aug 3, 2011
1,001
520
Tunis
So here is my commands output, hope it helps
EDIT: oups, I thought you need only these 3 commands, I didn't read your post to the end, will post the other commands tomorrow

TOP
Code:
User 5%, System 6%, IOW 0%, IRQ 0%
User 17 + Nice 0 + Sys 20 + Idle 279 + IOW 0 + IRQ 0 + SIRQ 0 = 316

  PID PR CPU% S  #THR     VSS     RSS PCY UID      Name
 4480  0   5% R     1   1100K    468K  fg app_27   top
  571  0   3% S    86 409676K  74528K  fg system   system_server
 4200  0   2% S    14 302320K  49804K  fg app_27   jackpal.androidterm
 4460  0   1% S     1      0K      0K  fg root     kworker/0:2
  897  0   0% S    13 308828K  65552K  fg system   com.android.systemui
   39  0   0% S     1      0K      0K  fg root     kworker/u:1
  145  0   0% S    33 192540K   9800K  fg media    /system/bin/mediaserver
  142  0   0% S    10  33188K  13436K  fg system   /system/bin/surfaceflinger
  885  0   0% S     1   2716K   1224K  fg wifi     /system/bin/wpa_supplicant
   23  0   0% S     1      0K      0K  fg root     sync_supers
   24  0   0% S     1      0K      0K  fg root     bdi-default
   25  1   0% S     1      0K      0K  fg root     kblockd
   26  0   0% S     1      0K      0K  fg root     irq/114-tegra_s
   27  1   0% S     1      0K      0K  fg root     spi_tegra-1
   28  1   0% S     1      0K      0K  fg root     irq/125-tegra_s
   29  1   0% S     1      0K      0K  fg root     spi_tegra-3
   30  0   0% S     1      0K      0K  fg root     khubd
   31  1   0% S     1      0K      0K  fg root     irq/118-tps8003
   32  1   0% S     1      0K      0K  fg root     tps65200
   33  1   0% S     1      0K      0K  fg root     rpciod
   34  0   0% S     1      0K      0K  fg root     kworker/0:1
   35  1   0% S     1      0K      0K  fg root     cpu-tegra
   36  1   0% S     1      0K      0K  fg root     cpu-tegra3
   37  1   0% S     1      0K      0K  fg root     cpu-tegra3-plug
   38  1   0% S     1      0K      0K  fg root     baseband_xmm_po
   40  1   0% S     1      0K      0K  fg root     htc_simhotswap
   41  1   0% S     1      0K      0K  fg root     charger_ctrl_ti
   42  1   0% S     1      0K      0K  fg root     batt_timer
   43  1   0% S     1      0K      0K  fg root     khungtaskd
   44  1   0% S     1      0K      0K  fg root     kswapd0
   45  1   0% S     1      0K      0K  fg root     fsnotify_mark
   46  1   0% S     1      0K      0K  fg root     nfsiod
   47  1   0% S     1      0K      0K  fg root     crypto
   63  1   0% S     1      0K      0K  fg root     tegradc.0/a
   64  1   0% S     1      0K      0K  fg root     tegradc.0/b
   65  1   0% S     1      0K      0K  fg root     tegradc.0/c
   66  0   0% S     1      0K      0K  fg root     irq/223-host_sp
   67  1   0% S     1      0K      0K  fg root     nvhdcp1
   68  1   0% S     1      0K      0K  fg root     tegradc.1/a
   69  1   0% S     1      0K      0K  fg root     tegradc.1/b
   70  1   0% S     1      0K      0K  fg root     tegradc.1/c
   71  1   0% S     1      0K      0K  fg root     mhl_sii9234_wq
   72  1   0% S     1      0K      0K  fg root     nct1008
   73  1   0% S     1      0K      0K  fg root     vib
   74  1   0% S     1      0K      0K  fg root     cable_detect
   82  1   0% S     1      0K      0K  fg root     fsl_tegra_udc
   83  1   0% S     1      0K      0K  fg root     f_mtp
   84  1   0% S     1      0K      0K  fg root     file-storage
   86  0   0% S     1      0K      0K  fg root     irq/393-synapti
   87  1   0% S     1      0K      0K  fg root     cm3629_wq
   88  1   0% D     1      0K      0K  fg root     kinteractiveup
   89  1   0% S     1      0K      0K  fg root     kn3ocold
   90  1   0% S     1      0K      0K  fg root     led
   91  1   0% S     1      0K      0K  fg root     led_powerkey
  103  1   0% S     1      0K      0K  fg root     binder
  104  1   0% S     1      0K      0K  fg root     hd-audio0
  105  0   0% S     1      0K      0K  fg root     mmcqd/0
  106  0   0% S     1      0K      0K  fg root     irq/77-tegra_ac
  107  0   0% S     1      0K      0K  fg root     irq/77-tegra_ac
  108  1   0% S     1      0K      0K  fg root     detect
  109  1   0% S     1      0K      0K  fg root     button
  110  1   0% S     1      0K      0K  fg root     HS_PMIC_DETECT
  111  1   0% S     1      0K      0K  fg root     HS_PMIC_BUTTON
  112  1   0% S     1      0K      0K  fg root     HS_GPIO_DETECT
  113  1   0% S     1      0K      0K  fg root     HS_GPIO_BUTTON
  114  1   0% S     1      0K      0K  fg root     poke_queue
  115  1   0% S     1      0K      0K  fg root     rq_stats
  116  0   0% S     1    324K    184K  fg root     /sbin/ueventd
  119  0   0% S     1      0K      0K  fg root     kworker/u:2
  120  0   0% S     1      0K      0K  fg root     jbd2/mmcblk0p12
  121  1   0% S     1      0K      0K  fg root     ext4-dio-unwrit
  122  0   0% S     1      0K      0K  fg root     jbd2/mmcblk0p15
  123  1   0% S     1      0K      0K  fg root     ext4-dio-unwrit
  124  0   0% S     1      0K      0K  fg root     jbd2/mmcblk0p13
  125  1   0% S     1      0K      0K  fg root     ext4-dio-unwrit
  126  1   0% S     1      0K      0K  fg root     jbd2/mmcblk0p3-
  127  1   0% S     1      0K      0K  fg root     ext4-dio-unwrit
  128  0   0% S     1      0K      0K  fg root     jbd2/mmcblk0p18
  129  1   0% S     1      0K      0K  fg root     ext4-dio-unwrit
  130  1   0% S     1      0K      0K  fg root     jbd2/mmcblk0p19
  131  1   0% S     1      0K      0K  fg root     ext4-dio-unwrit

Vmstat
Code:
procs  memory                       system          cpu
 r  b    free mapped   anon   slab    in   cs  flt  us ni sy id wa ir
 2  0   67328 119288 326872  37496   201  448    0  13  0  8 80  0  0
 0  0   67328 119252 326876  37496   224  398    0  19  0  5 77  0  0
 0  0   66820 119164 327452  37396   680 1076    0  21  0 15 99  0  0
 0  0   66820 119180 327452  37388   188  267    0  10  0  6 86  0  0
 0  0   66820 119252 327560  37388   264  517    0  14  0 16 73  0  0
 0  0   66820 119252 327576  37388   185  376    0  14  0  6 81  0  0
 0  0   66820 119252 327504  37388   233  400    0  15  0  6 82  0  0
 1  0   68928 119252 325308  37388   205  399    0  16  0  6 80  0  0
 0  0   68928 119252 325312  37388   252  614    0  12  0 11 80  0  0
 0  0   68928 119252 325340  37380   245  385    0  10  0  9 81  0  0
 0  0   68928 119252 325344  37376   245  408    0   8  0  7 88  0  0
 3  0   68928 119252 325344  37376   256  340    0  11  0 14 78  0  0
 2  0   68928 119252 325344  37376   283  411    0  10  0 13 81  0  0
 0  0   68928 119252 325344  37376   227  421    0  10  0  7 85  0  0
 0  0   68928 119252 325348  37376   230  399    0  12  0 11 79  0  0
 0  0   68928 119252 325348  37376   201  405    0  11  0  7 85  0  0
 1  0   68928 119252 325348  37376   215  392    0  10  0 10 83  0  0
 0  0   68928 119252 325352  37376   221  399    0  11  0  8 83  0  0
 0  0   68928 119252 325352  37376   202  378    0  12  0  7 84  0  0
 0  0   68928 119252 325404  37376   253  793    1  16  0 11 75  1  0
procs  memory                       system          cpu
 r  b    free mapped   anon   slab    in   cs  flt  us ni sy id wa ir
 0  0   68928 119264 325372  37380   216  409    0   7  0 11 84  0  0
 2  0   68928 119264 325376  37380   195  424    0  18  0  6 79  0  0
 0  0   68928 119264 325376  37380   207  406    0  11  0  6 86  0  0
 0  0   68928 119264 325376  37380   211  400    0  15  0  2 86  0  0
 0  0   69672 118380 325484  37380   618 1261    0  38  0 16 49  0  0
 2  0   69176 119304 325500  37380   328  787    0  14  0  9 77  0  0

cat /proc/fs/jbd2/mmcblk0p15-8/info
Code:
971 transaction, each up to 2196 blocks
average:
  0ms waiting for transaction
  370ms running transaction
  0ms transaction was being locked
  0ms flushing data (in ordered mode)
  10ms logging transaction
  6714us average transaction commit time
  211 handles per transaction
  5 blocks per transaction
  6 logged blocks per transaction
 
Last edited:

fizzlington

Senior Member
May 9, 2011
122
30
Brisbane
I'll start by saying I don't have any logs to share. But I experienced the same thing after filling my phone with music. I was overseas and had a useless phone for a week.

Although you said it only seemed to affect your phone, my serial starts with HT23MW, could be from the same batch of phones?
 

HebeGuess

Senior Member
May 15, 2012
766
357
Samsung Galaxy S20 FE
While reading op I recall just saw someone asking JBQ on a Android Building's 4.1.2 in AOSP discussion..

any changes to address this problem?

"Nexus 7 slow when less than 3GB free "
http://productforums.google.com/forum/#!msg/mobile/loqbCbKVMWE/veH_7NAk-YgJ
- show quoted text -

I then went back there and follow up the link above..

By reading into the discussion thread at least we can confirmed that Google is working on it and the issue still left un-fix on JRO03S...

:angel:
 

show-p1984

Inactive Recognized Developer
Jul 30, 2011
3,700
11,805
So here is my commands output, hope it helps
EDIT: oups, I thought you need only these 3 commands, I didn't read your post to the end, will post the other commands tomorrow

TOP
Vmstat
cat /proc/fs/jbd2/mmcblk0p15-8/info
Now look into the op, do you see the part where I say you shouldn't post here if you don't have those lags?
Now look again at the op, and check your logs with the one posted, especially the red parts. You notice something?
Thanks for posting logs with no issue present.
I'll start by saying I don't have any logs to share. But I experienced the same thing after filling my phone with music. I was overseas and had a useless phone for a week.

Although you said it only seemed to affect your phone, my serial starts with HT23MW, could be from the same batch of phones?
Mine is a HT23JW, though I really doubt that it is a hardware issue, or at least not solely a hardware issue.
 
  • Like
Reactions: VictoriousShooter

zombiefly

Senior Member
Jan 8, 2006
291
47
i am not able to post logs at the moment, but wish to add my experiences to this topic.

Since getting my HOX, i had issues when transferring large amounts of data (music) to the phone. As has been said above, the phone becomes unresponsive, heats up and basically doesn't work until you either: a) wait 24-36hrs with it on charge b) refresh a rom from a nandroid

I moved to custom roms on day 2 of owning the HOX and was most disappointed to see it happen again. I have come up with this workaround to copy music to the phone:

1. mount as usb drive
2. copy as much data as you like over usb
3. after transfer, right click on the usb icon near the clock (in windows) and eject the usb drive from the computer
4. Without ending usb transfer mode on the phone, reboot it.

bizarrely, this stops the lag phase after music transfer. I've been able to do this 4 times successively now without running into the issue. I have no idea why this is??? o_O
 

JohnCorleone

Senior Member
Dec 19, 2010
16,188
5,864
Whittier,CA
i am not able to post logs at the moment, but wish to add my experiences to this topic.

Since getting my HOX, i had issues when transferring large amounts of data (music) to the phone. As has been said above, the phone becomes unresponsive, heats up and basically doesn't work until you either: a) wait 24-36hrs with it on charge b) refresh a rom from a nandroid

I moved to custom roms on day 2 of owning the HOX and was most disappointed to see it happen again. I have come up with this workaround to copy music to the phone:

1. mount as usb drive
2. copy as much data as you like over usb
3. after transfer, right click on the usb icon near the clock (in windows) and eject the usb drive from the computer
4. Without ending usb transfer mode on the phone, reboot it.

bizarrely, this stops the lag phase after music transfer. I've been able to do this 4 times successively now without running into the issue. I have no idea why this is??? o_O

I believe that's what I do when I add my music. And haven't had any lag but also.may not have enough on there.

Also, I have a Nexus 7 and TeamViewer if you want to check anything out with the N7

Sent from my EVO using Xparent SkyBlue Tapatalk 2
 

Xmister

Senior Member
Jul 4, 2010
1,037
1,946
From the symptoms it looks like TRIM is not working.

With ext4 filesystems we need a mount option "discard" to enable this. Of course it's only really affective if we mount using this every time since a format.
 

show-p1984

Inactive Recognized Developer
Jul 30, 2011
3,700
11,805
Confirmed, it's exactly the same issue as the "Nexus 7 - issue".

n7_iowait.png
 
Last edited:

zimphishmonger

Senior Member
Jun 4, 2009
3,890
729
I can confirm this issue, when you have <1GB of free space available. I didnt pull up terminal to see how bad the IO was, but the UI was crawling. Almost didnt need to, knew it was this issue.

Ill try and pull a log later...
 
Last edited:

zvieratko

Senior Member
Dec 9, 2007
421
110
Prague
I had the same problem on occasion - 100% IOWAIT and the fix was to factory reset and restore from backup. Usually happened after numerous upgrades of the same ROM (so no full wipe for a while but a lot of IO on /system). I also suspected trim/discard was not working, but I thought our phones didn't have anything like that and suspected filesystem corruption.
I also experienced FS corruption that triggered the same behaviour - fsync and journaling is there for a reason and a popular "tweak" is to disable it - bad, bad, bad for consistency...

Btw I just ran LagFix - worked fine, but I was surprised that a second pass trimmed a lot on /data again - about 50% what the first run did... subsequent runs did nothing as expected. (and it worked on a Sense ROM for me :)) EDIT: looks like it didn't work after all? After reboot it just trims again the ~same amount of bytes... (nonsensical on /system)
 
Last edited:

Top Liked Posts

  • There are no posts matching your filters.
  • 13
    [DEVS & EXP-USERS ONLY] HTC One X - High I/O wait. Same issue as the N7. (confirmed)

    This little thread aims at getting some information about an issue I thought for months I am alone with.
    Since I got my One X, stock or running custom stuff, my phone always gets sluggish far far beyond the point of being useable if I fill my /data partition with too much data (~1GB free triggers it usually), or move large files around in it (dd'ing some zeros onto it triggers it too).
    No kind of rebooting, cache wiping, or anything else will fix it permanently. The lags will be back after a very short while once you experienced them.
    Vodafone told me to RMA the phone (it was only software branded), but If you know me, then you know that I am lazy in these kinds of things and since I thought I was the only one and a wipe of /data fixed it every time for a short period, I didn't do it.
    Yeah I know. Lazy b*tch. :p

    A few hours ago 'Aux' who had complained about a very sluggish phone a few times already joined IRC and described his problem in more detail and I suddenly made the connection in my mind between my issue and his. He then allowed me to debug that stuff, and I came up with the exact same issue I am facing continuously:
    High I/O waits while writing/reading from /data.

    This thread aims at gathering intel on this issue. Do not answer here with pointless questions or howto...? stuff.
    Ah btw: No, one lag after unlocking the phone is not the issue I am talking about. I mean: ~1-10 fps on the launcher, app opening takes minutes (games) & still lags like hell and downloading apps in the market is close to impossible. (~10byte p second)
    So if you don't experience this, don't post logs. I am serious, I will report anyone who wastes my time on that.


    To check if your phone suffers from this, run 'top'
    if it shows something like this:
    Code:
    User 2%, System 1%, [COLOR="Red"][B]IOW 74%[/B][/COLOR], IRQ 0%
    User 34 + Nice 0 + Sys 18 + Idle 263 + IOW 902 + IRQ 0 + SIRQ 1 = 1218
    
      PID PR CPU% S  #THR     VSS     RSS PCY UID      Name
     2061  0   1% S    13 471140K  32956K  fg u0_a75   com.google.android.apps.reader
      519  1   0% S    73 559264K  58024K  fg system   system_server
     1035  1   0% S    20 512864K  63100K  fg u0_a86   com.anddoes.launcher
     2265  1   0% R     1   1132K    500K     root     top
    [COLOR="red"][B]  105  0   0% D     1      0K      0K     root     mmcqd/0[/B][/COLOR]
      898  1   0% S    14 483120K  60344K  fg u0_a39   com.android.systemui
      882  0   0% S     1   2768K   1488K     wifi     /system/bin/wpa_supplicant
      435  0   0% S     1      0K      0K     root     kworker/0:2
      173  1   0% S    11  29644K  10104K  fg system   /system/bin/surfaceflinger
    [B][COLOR="red"]  123  1   0% D     1      0K      0K     root     jbd2/mmcblk0p15[/COLOR][/B]
      879  0   0% S     1      0K      0K     root     irq/340-wl12xx

    run 'vmstat', if it looks like this
    Code:
    procs  memory                       system          cpu
     r  b    free mapped   anon   slab    in   cs  flt  us ni sy id wa ir
     0  [COLOR="Red"]3[/COLOR]   29488  64232 261024  45956  1788 2723    0  22  0 29 99 [COLOR="red"]96[/COLOR]  0
     0  [COLOR="red"]2[/COLOR]   29488  64284 261056  45952   620  905    0   1  0  4 99 [COLOR="red"]99[/COLOR]  0
     1  [COLOR="red"]6[/COLOR]   29556  64284 261056  45908   472  755    0   1  0  4 99 [COLOR="red"]99[/COLOR]  0
     3  [COLOR="red"]3[/COLOR]   29496  64284 261056  45908   454  680    0   0  0  1 99 [COLOR="red"]99[/COLOR]  0
     0  [COLOR="red"]8[/COLOR]   29436  64284 261056  45904   462  767    0   4  0  4 99 [COLOR="red"]98[/COLOR]  0
     3  [COLOR="red"]4[/COLOR]   31300  64328 261328  45840  2177 4151    0  50  0 15 99 [COLOR="red"]92[/COLOR]  0

    and if 'cat /proc/fs/jbd2/mmcblk0p15-8/info' looks like this too:
    Code:
    cat /proc/fs/jbd2/mmcblk0p15-8/info
    649 transaction, each up to 2196 blocks
    average:
      0ms waiting for transaction
      1460ms running transaction
    [B][COLOR="Red"]  16544590ms transaction was being locked[/COLOR][/B]
      0ms flushing data (in ordered mode)
      710ms logging transaction
      603582us average transaction commit time
      100 handles per transaction
      7 blocks per transaction
      8 logged blocks per transaction
    Then you are most likely affected.


    What I need from you:
    Everything the above commands throw out as well as the following:
    Code:
    echo 1 > /proc/sys/vm/block_dump ; cat /proc/kmsg ; echo 0 > /proc/sys/vm/block_dump
    (this is very spammy, you might want to add '> file.log' to the kmsg part)
    -----
    iostat
    -----
    iostat -kd 5
    (keep it running for about 30 seconds)
    -----
    cat /proc/meminfo
    -----
    cat /proc/diskstats

    This issue is the same I am facing on my N7. ~2.5Gb free will trigger a very poor performance on the N7 as well, so we have found a general bug in one of the used components (hardware if it's the same, the jbd journaling system, ... )

    To figure out what this is exactly I need as much intel as you can gather if your device is in such a state.
    I will attach my findings, this thread and all logs gathered on the already existing bug report for the N7 @google. (since it is more likely to get an answer out of them than it is to get one from HTC, sadly)

    If you have a N7 facing the same issue:
    Attach the logs here too, but please specify that they are from a N7.
    4
    Confirmed, it's exactly the same issue as the "Nexus 7 - issue".

    n7_iowait.png
    4
    System trim - https://www.dropbox.com/s/91site2hr31kuf3/Fstrim_sys_CWM.zip

    Data trim - https://www.dropbox.com/s/053a73xcznmxv60/Fstrimmer_data_CWM.zip

    As requested, fstrim zip to trim data, cache AND system from recovery in first zip.

    I would recommend just trimming data and cache, as system isnt written to on a continual basis during normal use of rom, so even if it required trimming it wouldnt be noticable. Plus of course its trimmed when its formatted when roms are flashed.

    Again, check amount trimmed in recovery.log on sd.
    3
    Exactly....what are we doing here than except for resurrecting a topic which wasn't posted in over a year !
    3
    I've compiled fstrim for Android and packaged it inside APK with nice GUI. You can find LagFix tool here. It is much better then backing up and formatting /data. And A LOT faster too!