Hello,
Anyone has any idea how to safely increase dmesg buffer size to 32MB on Nexus 9? I use Android 5.0.1.
I set CONFIG_LOG_BUF_SHIFT=22 in flounder_defconfig at compile time in order to increase buffer to 4MB (also edit Kconfig to allow increasing it over 2 MB).
When i set CONFIG_LOG_BUF_SHIFT to 23 (8 MB) and beyond i will get "unhandled level 2 translation fault" exceptions logged into dmesg log or even worse device wont boot.
I assume this exception is because of CMA. It make sens as dmesg buffer memory is aligned to 4bytes so 4MBx4=16 MB. The CMA is set to 16MB increasing it to 32 MB will make device not boot.
Thank you!
Related
I notice SD cards are rated 60x 133x or greater. Will I notice a difference on my Blue Angel?
Is the BA USB interface version 2.0?
I'm torn between getting the Sandisk 1GB (with built in USB) or the new 2GB, a third option is the new Sandisk Extremee III (1GB) with superfast 20Mbps xfer.
I am curious t this as well.
I recently purchased a no-frills 2GB SD card from newegg that claimed to be 150x. When benchmark testing with PocketMechanic, I saw that my storage card only did 1.8x read / 0.8x write. However, the onboard storage benchmarked to be 2.9x read / 4.2x write. [Note: Both onboard Storage and SD Storage card were formatted before testing.]
I haven't done the research to find out if BA is USB 1.1 or 2.0, but I've noticed that a USB SD card reader is significantly faster than transfering files through the BA. It took 4 minutes to transfer 73MB for 193 files and 39 folders via windows explorer, or roughly 300K/s, where as I was able to transfer 1.5GB in 401 files and 59 folders in roughly the same time via USB SD card reader.
If the speed of the no-name card is in question, I confirmed the speed to be at least 5.5MB/s, by copying back 1.5GB in 401 files / 59 folders in 270 seconds via USB to SD card reader. Had this been one single contiguous file, the transfer rates would have been alot faster... I could have confirmed by copying a like-sized file to the media, but you'll have to take my word on this, as I've already spent quite some time satisfying my own curiousity to answer your question
See below for Pocket Mechanic details:
Details:
Card tested "Storage Card"
Min. read time: 0.25ms
Max read time: 18.00 ms
Avg. read time 1.87ms
Sector/block: variable 4-64
Min write time: 0.50ms
Max write time: 16.00 ms
Avg. write time: 4.02ms
Read/write ratio: 0.46
Sector size: 512 bytes
Start Sector: 0
End Sector: 3910656
Total Sectors read: 2504
Total data read: 1.00MB
Total read time: 4673 ms
Total sectors written: 2504
Total data written: 1.00MB
Total write time: 10057ms
Details:
Card tested "Storage"
Min. read time: 0.25ms
Max read time: 16.00 ms
Avg. read time 1.14ms
Sector/block: variable 4-64
Min write time: 0.50ms
Max write time: 9.25 ms
Avg. write time: 0.80ms
Read/write ratio: 1.43
Sector size: 512 bytes
Start Sector: 0
End Sector: 125488
Total Sectors read: 2504
Total data read: 1.00MB
Total read time: 2855 ms
Total sectors written: 2504
Total data written: 1.00MB
Total write time: 2001ms
Thanks, pity you ont have any other cards to compare too.
FYI, my curiousity got the best of me again, so I tested my no frill card's speed more thoroughly. According to SiSoft Sandra, my card is at least 55x to 65x at 64MB chunks, via USB reader. See below for full details.
I also performed my own test, where I copied a single contiguous 1.9GB file (1944973970 bytes) to and from my SD card via USB reader.
Read: 262.50 seconds, 7.4MB/sec, 40x (@ 176K/x) or 50x (@ 150K/x)
Write: 204.84 seconds, 9.4MB/sec, 53x (@ 176K/x) or 63x (@ 150K/x)
This is no where close to the 150X claimed speed, but this may be a limitation of my 2.33Ghz Centrino laptop's 5400RPM hdd? Really can't tell since I don't have another SD card to test with, nor do I have current plans to test on other hardware...
Hope this helps!
---------
SiSoftware Sandra
Benchmark Results
Combined Index : 1296 operation(s)/min
Endurance Factor : 9.2
512B Files Test : 1382 operation(s)/min
32kB Files Test : 1617 operation(s)/min
256kB Files Test : 1012 operation(s)/min
2MB Files Test : 124 operation(s)/min
64MB Files Test : 8 operation(s)/min
Results Interpretation : Higher index values are better.
Performance Test Status
Run ID : ALEX2 on Saturday, August 27, 2005 at 11:51:07 PM
SMP Test : No
Total Test Threads : 1
SMT Test : No
Dynamic MP/MT Load Balance : No
Processor Affinity : No
512B Files Test
Read Performance : 13831 operation(s)/min (115 kB/sec, 0x)
Write Performance : 494 operation(s)/min (4 kB/sec, 0x)
Delete Performance : 1358 operation(s)/min
File Fragments : 1.0
Combined Index : 1382 operation(s)/min
32kB Files Test
Read Performance : 7001 operation(s)/min (3734 kB/sec, 21x)
Write Performance : 653 operation(s)/min (348 kB/sec, 1x)
Delete Performance : 1366 operation(s)/min
File Fragments : 1.0
Combined Index : 1617 operation(s)/min
256kB Files Test
Read Performance : 1865 operation(s)/min (7957 kB/sec, 45x)
Write Performance : 507 operation(s)/min (2163 kB/sec, 12x)
Delete Performance : 1340 operation(s)/min
File Fragments : 1.0
Combined Index : 1012 operation(s)/min
2MB Files Test
Read Performance : 267 operation(s)/min (9114 kB/sec, 51x)
Write Performance : 52 operation(s)/min (1775 kB/sec, 10x)
Delete Performance : 937 operation(s)/min
File Fragments : 1.0
Combined Index : 124 operation(s)/min
64MB Files Test
Read Performance : 8 operation(s)/min (8738 kB/sec, 49x)
Write Performance : 6 operation(s)/min (6554 kB/sec, 37x)
Delete Performance : 555 operation(s)/min
File Fragments : 1.0
Combined Index : 8 operation(s)/min
Endurance Test Status
Operating System Disk Cache Used : No
Use Overlapped I/O : No
Test File Size : 32MB
Block Size : 512 byte(s)
File Fragments : 1
Endurance Benchmark Breakdown
Repeated Sector ReWrite : 333 kB/s
Sequential Sector Write : 332 kB/s
Random Sector Write : 25 kB/s
Drive
Total Size : 1.9GB
Free Space : 1.9GB, 100%
Cluster Size : 4kB
Performance Tips
Notice 5901 : 1x=176kB/s; As some device makers use 1x=150kB/s exercise caution when comparing measured vs. published ratings.
Notice 5008 : To change benchmarks, click Options.
Notice 5004 : Synthetic benchmark. May not tally with 'real-life' performance.
Notice 5006 : Only compare the results with ones obtained using the same version!
Notice 5207 : Consider using the File System Benchmark for non-Flash devices.
Notice 5900 : Endurance factor can only be used on the same type of device (SLC or MLC).
Tip 11 : Use the 'Switch Chart Type' button to switch between Detailed and Combined charts.
Tip 2 : Double-click tip or press Enter while a tip is selected for more information about the tip.
Link: https://source.android.com/compatibility/downloads.html#android-403
According to CDD document, it is stated below
"Device implementations MUST have at least 340MB of memory available to the kernel and userspace. The 340MB MUST be in addition to any memory dedicated to hardware components such as radio, video, and so on that is not under the kernel's control."
MY Understanding is:
340 MB is required for both user and kernel space. 340 MB is for hardware components ,
So, in total minimum 680 MB
NOTE: Please correct me if my understanding is wrong ?
But, if i look at mobiles that are available in the market, they are not compliance with the requirement.
For example:
HTC ONE V , has 512 MB of RAM, but still runs android 4.0.3 , in which the requirement is 680MB.
Considering above, how the OEM vendor passes compatibility test?
keestupillo said:
Link: https://source.android.com/compatibility/downloads.html#android-403
According to CDD document, it is stated below
"Device implementations MUST have at least 340MB of memory available to the kernel and userspace. The 340MB MUST be in addition to any memory dedicated to hardware components such as radio, video, and so on that is not under the kernel's control."
MY Understanding is:
340 MB is required for both user and kernel space. 340 MB is for hardware components ,
So, in total minimum 680 MB
NOTE: Please correct me if my understanding is wrong ?
But, if i look at mobiles that are available in the market, they are not compliance with the requirement.
For example:
HTC ONE V , has 512 MB of RAM, but still runs android 4.0.3 , in which the requirement is 680MB.
Considering above, how the OEM vendor passes compatibility test?
Click to expand...
Click to collapse
i think the requirement should be 380 mb for memory and kernel(its my point of view)
When i lauch a AVD in the emulator i get... "the memory needed by this avd exceeds the max specified in your HAXM configuration". When i edit/adjust the ram in the AVD to be eual or less than the HAXM config(dropping it to 768, making them eual) then relaunch the avd, the ram size remains 1024 and the same error messege is given. In haxm set up, my ram range is 512 to 949mb(default 512). I manually set haxm ram size to 768. No matter what value i assign for the AVD ram it stays 1024 when launched. Is there something i am missing as to how to save the AVD ram changes. I hope there is a answwer, I would like to have it running correctly to use with some developement tutorials. Any help would be greatly appreciated.
My phone: S2 with CM 12.1 (latest nightly)
Hi,
since 2 days my phone has an odd behaviour.
In the first place I'd tried to take a screenshot. But in the notification only a note appear, that the "screenshot couldn't be captured". It doesn't matter in which app I try to take a screenshot.
The second problem is, when I try to take a picture with the default camera app. It appears, that the app it take the photo, but it doesn't save any.
The third odd behaviour is, that within the gallery app, it show all the picture folders and images within, but not a single thumbnail nor the picture itself in the detail view.
I can't save any (type of) file to the phone memory.
My first thought was, maybe the memory on the internal SD doesn't have enough space. But here it goes really weird. Following statistic by DiskInfo app:
- System (FACTORYFS) 540 MB free (of 1 GB) - sidenote - I had repit(ed) the partions with lanchons script.
- Data (DATAFS) 3,9 GB free (of 6 GB)
- internal SD (UMS) 2,2 GB free (of 7,5 GB)
Things which worked well:
- Text messaging (but I can't receive picture or videos)
- Reading Twitter and see all images (but can't save any of them)
- Browsing with Firefox (but can't download any document or file)
Does anyone have any idea what's going wrong?
I solved the problem.
By using the latest version of lanchons repit script, which seems to repair something.
Here some parts of the recovery.log:
Code:
Cluster 34558 out of range (42174573 > 982078). Setting to EOF.
[...]
Cluster 35069 out of range (101237010 > 982078). Setting to EOF.
[...]
/DCIM/Camera/IMG_20160213_103344.jpg
Contains a free cluster (34432). Assuming EOF.
/DCIM/Camera/IMG_20160213_103344.jpg
[...]
/gReader/.cache/-2010898466/img_10
File size is 72812 bytes, cluster chain length is 8192 bytes.
Truncating file to 8192 bytes.
Reclaimed 723 unused clusters (5922816 bytes).
Free cluster summary wrong (285140 vs. really 285863)
Auto-correcting.
Starting verification pass.
Performing changes.
/dev/block/mmcblk0p11: 23293 files, 696214/982077 clusters
info: errors detected, retrying the file system check
dosfsck 3.0.12, 29 Oct 2011, FAT32, LFN
Starting check/repair pass.
Starting verification pass.
[...]
Hello I own d802, get it 1 week ago, i done some work on optimising cm12.1 kernel. I used stock google compiler to build(4.8 its included in cyanogenmod). I tested it on d802 need tests for other variants. if you interested contact me. List of patches are: (GITHUB LINK)
writeback: fix occasional slow sync(1)
Fix uninitialized div_s64 for gcc 4.9
selinux: fix sel_write_enforce broken return value
gpu: ion: fix corruption of ion_import_dma_buf
mnt: Fail collect_mounts when applied to unmounted mounts
Revert "msm: kgsl: Wake up GPU only for touch events"
arch/arm: boot: dts: Fix vidc ddr table and set max freq as MSM8974AC
Input: aiptek - fix crash on detecting device without endpoints
Input: aiptek - adjust error-handling code label
slub: fix incorrect return type of get_any_partial()
mm: slub: Default slub_max_order to 0
ipv4, fib: pass LOOPBACK_IFINDEX instead of 0 to flowi4_iif
Revert "msm: use of swp{b} is deprecated for ARMv6+"
Revert "decompressor: Actually enable unaligned memory access for ARM…
arm: kernel: Drop warning about return_address not using unwind tables
msm: use of swp{b} is deprecated for ARMv6+
lz4: fix another possible overrun
lz4: ensure length does not wrap
lz4: fix compression/decompression signedness mismatch
lib/lz4: correct the LZ4 license
decompressor: Actually enable unaligned memory access for ARMv6+
mm: fix prctl_set_vma_anon_name
mm/slub: don't wait for high-order page allocation
I wont relase before test other variants
why cm12.1?
Why 12.1 ? Maybe cm 13 better than 12.1 and more use