Does Full Disk Encryption affect TRIM command? - Google Pixel Questions & Answers

Recently i noticed that write performance looks really bad on my phone (~10MB/s seq write, ~3MB/s random write), so i am suspecting if there's conflict between trim and encryption.
Is there a way to find out if TRIM is working correctly? (ideally without root)
Thanks!

Related

Three Very Important Technical Questions

My phone was acting weird (programs that were installed were not being found) so I soft reset. THe phone became infinitely paused at the touch pro flash screen. So i tried reflashing with my SD card. I flashed 2 different roms, both with filesizes greater than 110 megs. They would get stuck infinitely flashing through the SD card. Finally I flashed a rom (Mighty 6) which was under 100 megs and it worked.
1) It almost sounds like I had a device filesystem corruption or something. How would I determine if a portion of my device's filesystem is corrupt?
2) If part of the file system is corrupt, is there a way to mark the corrupt portions?
3) If it's not a corrupt filesystem, what else could it be?
Thanks guys, I'd appreciate any help!
Have you tried flashing using activesync? or a different SD card?
If so and you still have the same problem, then its quite probable you have some corrupted memory on the phone (sry, i dont know how to block use of the corrupted parts) - if not i'd say you have a corrupted SD card, and everytime you use it, it screws the flash.
Did you not try a hard reset between a softreset and a new flash??
Beeble said:
Have you tried flashing using activesync? or a different SD card?
If so and you still have the same problem, then its quite probable you have some corrupted memory on the phone (sry, i dont know how to block use of the corrupted parts) - if not i'd say you have a corrupted SD card, and everytime you use it, it screws the flash.
Did you not try a hard reset between a softreset and a new flash??
Click to expand...
Click to collapse
Hey, thanks for the reply...I could not get an activesync connection since the OS wouldn't load. I tried two different cards same result. I formatted both cards. I tried hard reset, at the time the phone would not even hard reset. Fortunately I have yet to encounter any issues with since I flashed Mighty's Rom.
I have a thread about this on ppcgeeks as well. I mentioned there it would be nice to have a utility such as chkdsk on windows mobile. I guess I'll take it as it goes, since it seems my phone is functioning correctly right now.
A few notes about Flash (the hardware type) and flashing (the 'firmware' type).
Flash memory isn't permanent. Writing to the flash memory itself is a (slightly) destructive process. Older chips used to have only a few hundred re-writes; newer flash devices support millions (and some of the newer ones, billions) of rewrites. This is important to your question because the hardware that controls the flash itself is designed to deal with the possibility of bad blocks.
Just like a normal HDD can develop bad sectors, Flash memory can develop bad 'blocks'. In fact, most flash chips actually ship with several bad blocks from the factory. Knowing the manufacturing process is imperfect, manufacturers which would design a flash chip meant to store 600 blocks might actually develop one with 650 blocks; factory testing is then performed to ensure that the amount of valid blocks is greater than or equal to 600.
The hardware/firmware controller that controls read/writes to the flash chip is designed to catalog all invalid bytes- bytes which, for example, after being written retain their previous value. It then automatically reroutes data writing around the bad block.
More succinctly, you shouldn't have to mark bad portions of the NAND flash- the device should be doing it by itself.
As a second note, you don't actually need a real ActiveSync connection to flash 'via ActiveSync.' A connection to the bootloader using the activesync drivers works fine most of the time- this is the device's "recovery mode."
ktemkin said:
A few notes about Flash (the hardware type) and flashing (the 'firmware' type).
Flash memory isn't permanent. Writing to the flash memory itself is a (slightly) destructive process. Older chips used to have only a few hundred re-writes; newer flash devices support millions (and some of the newer ones, billions) of rewrites. This is important to your question because the hardware that controls the flash itself is designed to deal with the possibility of bad blocks.
Just like a normal HDD can develop bad sectors, Flash memory can develop bad 'blocks'. In fact, most flash chips actually ship with several bad blocks from the factory. Knowing the manufacturing process is imperfect, manufacturers which would design a flash chip meant to store 600 blocks might actually develop one with 650 blocks; factory testing is then performed to ensure that the amount of valid blocks is greater than or equal to 600.
The hardware/firmware controller that controls read/writes to the flash chip is designed to catalog all invalid bytes- bytes which, for example, after being written retain their previous value. It then automatically reroutes data writing around the bad block.
More succinctly, you shouldn't have to mark bad portions of the NAND flash- the device should be doing it by itself.
As a second note, you don't actually need a real ActiveSync connection to flash 'via ActiveSync.' A connection to the bootloader using the activesync drivers works fine most of the time- this is the device's "recovery mode."
Click to expand...
Click to collapse
Excellent information, much of what I was too lazy to look up . Considering the information about bad sectors on flash-based memory that you have given me, do you have any ideas as to what the issue may(is) have been?
The only other explanation (albeit in my limited knowledge lol) that I can come up with is that the data within the roms I used somehow became corrupted, either during the download process, or the process of transferring the data to the card. I was in linux when I transferred the data to the card. Although I find this highly unlikely, I guess it could be a driver implementation issue with my cardreader and linux. I have not had a problem thus far though.

[Q] Understanding damaged blocks and partitons affected.

Hi there!
Since I have 3 bad blocks on my polaris, it's barely usable at Windows Mobile, and not so stable at Android.
I thought I can reinstall Android, but placing system/data on those unaffected sections. Maybe It would work fine.
The 3 blocks involved are:
BLOCK 589 (0x24D)
BLOCK 717 (0x2CD)
BLOCK 1446 (0x5A6)
As JockyW said, Radio, OS and a Fat partition is affected.
I know that if Radio is affected, maybe there's nothing much to do. But it works fine, at least for calling, texting and using wireless lan.
I would try to install again system/data to SD card, but I feel it a bit slower than NAND setup. Maybe System on unaffected NAND and Data on SD would be fine.
Would someone like to help me?
If we figure out how to do this, maybe a couple of bricked phones would get back to live
Regards!

[Q] corruption of filesys possible for Eee pad ?

I am wondering about this:
at the moment my Eee pad (1st gen) hangs during a large file copy (unresponsive OS) but its still copying the file from USB to internal memory. What would happen if I forcefully switch off the device ? Would the filesys be filled with blocks of data that aren't assigned to any file and use up free space ? Like with Fat and NTFS filesystems that then can be repaired with chkdsk on Windows.
What is the Eee pads /data filesys called ? I saw EXT4 when doing busybox's mount command to see the mounted partitions but when I google I find information that says it uses Tuxera Filesystem. Never heard of that one
When its EXT4 it is a journalling filesys huh so does that mean there can never be data blocks that take up space and are not assigned to any file ?
Can't help much, but I can say it is an SSD, which usually means that once the power is disconnected to the drive it simply STOPS. Unlike a hard disk, which has to spin down, while the head is still connected to the disk (which may be why it writes blocks to the disk and corrupts files), I would say these devices are very much designed with the thought of 'instant power loss' in mind, and that the drives do not corrupt data that easily (non-volatile etc).
luna_c666 said:
Can't help much, but I can say it is an SSD, which usually means that once the power is disconnected to the drive it simply STOPS.
Click to expand...
Click to collapse
well I'm not talking about random corruption caused by the spinning down but caused by aborting a file being written so the file is left 'unclosed' and possibly the registering of its used blocks in the 'File Allocation Table' that then isnt up to date . The immediate problem I had has been solved but I still wonder if the Filesystem of the Eee pad can get corrupted even when its FS would be a journalling FS and what tool I can use to test/repair its file integrity/used blocks list etc. I am rooted ;-)
Also if its a journalling FS, does the Eee pad replay the journal (to make the FS consistent again) when booting up ?
well from what I can tell having disconnected a few times myself during transfers, the protocol seems to be something like this;
Check to see that free space exists, allocate the free space needed for transfer, transfer the files and name the files accordingly..if the files aren't finalized, they don't get a name, this basically means the data, while some of it did transfer, it is left as 'free space' if it isn't tied to a filename in the system. But this is all based on presumption so take it FWIW I don't think it will corrupt data, I could be wrong, but with the amount of tinkering I have done on this thing I would imagine I would have encountered data corruption already.. making an educated guess here though..

[Q] EXT vs FAT

Is it possible to use EXT on the main partition instead of fat? It may improve stability and CM7 already supports EXT
it is theoretically possible to make tho whole sd card to ext, but if you use windows it wouldnt be clever to do that bacause windows can only read data from Fat or NTFS as far as i know
I know about the incompatibility issues, but I prefer a better phone and airdroid-like software may solve the problem of transfering files from and to the phone. Today I tried this but the phone froze on the lockscreen after boot. Is there anything to set?
I succesfully mounted a folder from my EXT partition and is really faster than the FAT partition, but the media process keeps force closing because of the permissions.
Is there anyway to solve this? The problem is visible with the camera, sometimes it works sometimes not, but when it does is really snappy and when you change the permissions of the photos the problem is solved; as far as I've been testing I haven't seen any other issue

Internal storage is vfat? (CM 10.1)

So I was wondering for a while why I couldn't do symbolic links and how in the world the fs was case insensitive. I looked in the mount list for something else and noticed that the "sdcard" was being mounted as vfat, not EXT4. I thought that even stock was supposed to have switched over amd that this was part of the whole point of the swtich to the MTP USB connection rather than UMS? Is there any particular reason it is on vfat? Is there any way I can switch without reinstalling from scratch? Even if it's not any faster, there are times when I want symbolic links, journalized writes, and etc that EXT4 does. Or am I just completely missing something?

Categories

Resources