rt/snapraid
1
0
Fork 0
mirror of https://github.com/amadvance/snapraid.git synced 2026-10-26 19:01:57 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions
1706 commits 14 branches 78 tags 69 MiB
Commit graph

565 commits

Author SHA1 Message Date
Andrea Mazzoleni
84c089aae4 Use 'data' in the configuration file instead of 'disk' 2016-01-11 18:54:53 +01:00
Andrea Mazzoleni
7cb33cc046 Add the new --test-io-cache option to control the cache size 2016-01-11 18:54:53 +01:00
Andrea Mazzoleni
984ea5a329 Log info about the filesystem used 2016-01-09 17:19:43 +01:00
Andrea Mazzoleni
9430106c71 Allow to disable inline assembly from configure 
This is required for the -fsanitize options that don't
support inline assembly statements.
 2016-01-09 11:19:52 +01:00
Andrea Mazzoleni
2e12b10590 Signal condition variables outside the lock 
This reduces the number of context switches, because the
signaled thread can start immediately.

Some interesting discussion is at:

Condvars: signal with mutex locked or not?
http://www.domaigne.com/blog/computing/condvars-signal-with-mutex-locked-or-not/

Calling pthread_cond_signal without locking mutex
4544494 (4544494)
 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
8e47e2b024 Signal the 'done' condition only at the waiting index 
This reduces the number of context switches.
 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
22d806a04f Minor beatification at the IO module 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
d71b1b21b6 Always print the CPU value, even if 0% 
In dry, it's really 0.
 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
c870b04b66 Refactor locks use in support 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
ec5b383c1d Add support for multithread data and parity reading in dry 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
e15b5b122c Remove the unnecessary parity_ptr variable in dry 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
da5be1c5a7 Reduce memory occupation for MT 
The copy buffer doesn't need a copy for each thread.
 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
fb404ef149 Print the amount of memory used for MT 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
4191989fa7 Restore error reporting for parity write 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
fa635f72b8 In the check log write status:correct only in verbose mode 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
a3121c31c5 Add support for multithread parity writing in sync 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
e69e8640ce Use pread also in import and search 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
18cc80bd7e Now pread and pwrite use is required for multithread 
We need to support concurrent access at the same file handle
from multiple threads, and they cannot share the same file pointer.
 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
62b7587c3a IO functions renaming to prepare for parity write support 
No functional change.
 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
d9f7657107 Remove the unnecessary parity_ptr variable in sync 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
0849bfc261 Use a thread-safe pread/pwrite implementation for Windows 
This may be useful in future for concurrent calls to the same file.
 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
b619b9bee1 Use a faster memory test 
We have now many buffers to test, and we need a faster test.
 2016-01-09 10:25:56 +01:00
Andrea Mazzoleni
55ac2b0d25 Use a lot more read-ahead buffers 
This improves the performance even more.
 2016-01-09 10:25:47 +01:00
Andrea Mazzoleni
afd29347bb Restore the io error limit feature for multithread support 2016-01-09 10:18:14 +01:00
Andrea Mazzoleni
cbf2040b89 Add mutex protection at the filesystem structure 
In sync we need to modify the filesystem, like to call
fs_deallocate(). We need then proper protection.
This also removes the need of multiple fs_last cached value.
 2016-01-09 10:18:14 +01:00
Andrea Mazzoleni
74204a42d7 Add support for multithread data reading in sync 2016-01-09 10:18:14 +01:00
Andrea Mazzoleni
51af8931b7 Add support for multithread data and parity reading in scrub 2016-01-09 10:18:14 +01:00
Andrea Mazzoleni
3c00248aa6 Add level knownledge to parity handle 2016-01-09 10:18:14 +01:00
Andrea Mazzoleni
a279333c0b Alternate scrub next block computation 2016-01-09 10:18:14 +01:00
Andrea Mazzoleni
2f5c6a39a9 Fix unaligned memory access in import 
After extent conversion, the hash is not not aligned anymore,
and we cannot cast anymore to uint32.
 2016-01-09 10:06:37 +01:00
Andrea Mazzoleni
9d4fd25774 Avoid a float division by zero in testing 
In the regression testing the fake probability goes
to 0, resulting in an impossible division.
 2016-01-09 10:05:57 +01:00
Andrea Mazzoleni
a45641dcc9 Add explicit conversion to unsigned from byte 
Implicit conversion is to signed int, that may overlow,
resulting in what C considers undefined behaviour.
Better to avoid it.
 2016-01-09 10:01:33 +01:00
Andrea Mazzoleni
9b918c3ca0 Some more coverage commenting 2016-01-06 00:26:24 +01:00
Andrea Mazzoleni
84f7a623ab Be more verbose in log and messages about UUID 2016-01-05 23:35:55 +01:00
Andrea Mazzoleni
cc8ca88ecf Add some more coverage exclusions 2016-01-05 12:23:05 +01:00
Andrea Mazzoleni
c879140663 Print the exit status of addr2line 2016-01-05 11:16:03 +01:00
Andrea Mazzoleni
fe20160958 Fix the precision field in the addr2line call 
It must be a int type in x64.
 2016-01-05 11:04:04 +01:00
Andrea Mazzoleni
62aba9228a Add a new test option to force parity update 2016-01-04 10:43:43 +01:00
Andrea Mazzoleni
5249f9cb04 Fix various typos in comments 2016-01-02 16:52:20 +01:00
Andrea Mazzoleni
b853b4c406 More explicative StackTrace message 2016-01-02 12:23:48 +01:00
Andrea Mazzoleni
7ad70c26f8 Remove the need of helgrind/drd suppression 
Using a mutex for protectign the global memory counter
should not affect performance.
 2016-01-01 17:57:04 +01:00
Andrea Mazzoleni
b059457190 Disable multi thread for content write 
It's not necessarely faster, so better to be safe
and avoid possible slowdown.
Also, the original reason of this feature was to detect
potential memory errors, but the original issue triggering this need
was instead not related to it. It was instead the problem to empty links.
 2016-01-01 17:56:42 +01:00
Andrea Mazzoleni
354d51725b Enable again multi write for content files for monothread environments 
If no thread support is present, keep the old way to write content files.
 2016-01-01 17:37:00 +01:00
Andrea Mazzoleni
f4289a9d7e More explicative rename message 
Some commands don't have UUID, and cannot detect
the rename.
 2015-12-30 22:21:48 +01:00
Andrea Mazzoleni
a8e0214a9a Clarify the message about text content file not supported 2015-12-29 08:05:32 +01:00
Andrea Mazzoleni
776c6a48ef Add the new --no-warnings option for release 2015-12-28 14:49:10 +01:00
Andrea Mazzoleni
cacc1cb0ce Autorename disks using UUID for matching 2015-12-28 13:55:22 +01:00
Andrea Mazzoleni
801d01c55b Store UUIDs in the disk element 2015-12-28 12:09:13 +01:00
Andrea Mazzoleni
c80fd07058 Remove warnings when building with low optimizations 2015-12-20 17:23:58 +01:00
Andrea Mazzoleni
584c362ba4 Handle correctly files without offset 
This avoid the warnings about duplicate offsets.
 2015-12-20 17:19:57 +01:00