Dirk Engling
6604d65779
Some clients have added the /stats URL as /announce URL and cause a lot of expensive operations. Prevent that
9 months ago
Dirk Engling
9f08041585
Fix tasklist iterators
9 months ago
Dirk Engling
384799b5b3
Merge branch 'master' into erdgeist-fix-top10-stats
9 months ago
Dirk Engling
4baec2592c
handle -1 length reads properly
9 months ago
Dirk Engling
9c98e1e775
Remember hash and not torrent pointer for use after releasing the lock on torrent vector
9 months ago
Dirk Engling
cc24e1ab2d
Add error if WANT_DYNAMIC_ACCESSLIST is enabled without any accesslist
2 years ago
Dirk Engling
6e591d7437
Add atomicity qualifier
2 years ago
Dirk Engling
f62398c748
Clean accesslists every five minutes
2 years ago
Dirk Engling
be825f5759
Add support for dynamic accesslists
2 years ago
Dirk Engling
110868ec4e
Fix return code check inversion introduced in last commit
3 years ago
Dirk Engling
e89905166c
Mark return code of write() as intentionally unused
3 years ago
Romain Porte
019d58d154
fix implicit fallthrough spelling
...
This commit fixes the syntax of the implicit fallthrough comments, in
order to be matched by GCC (and probably other compilers as well) with
the following regular expression:
[ \t.!]*([Ee]lse,? |[Ii]ntentional(ly)? )? fall(s | |-)?thr(ough|u)[
\t.!]*(-[^\n\r]*)?
See: https://gcc.gnu.org/onlinedocs/gcc/Warning-Options.html#index-Wimplicit-fallthrough=
3 years ago
Romain Porte
c4fc41a831
opentracker.c: check set*id return values
...
This commit fix the following similar warnings:
opentracker.c:562:7: warning: ignoring return value of ‘setegid’
declared with attribute ‘warn_unused_result’ [-Wunused-result]
[…]
The man page of these functions ask users to explicitly check the return
value in case of any error happening.
3 years ago
Dirk Engling
9a20ebe3f2
Fix typo. Thanks to Romain Porte
3 years ago
Dirk Engling
68a17b3ae4
Obey DESTDIR macro. Thanks to Romain Porte
3 years ago
Dirk Engling
80faf1c452
Using strip from enviroment to allow different strippers
4 years ago
Dirk Engling
298fe52f52
Make accesslists work again by testing the actual result of bsearch. Thanks to Hanno
4 years ago
Dirk Engling
3168071175
Fix two issues when splitting an iovec to large iobatches
4 years ago
Dirk Engling
99544a1030
Handle eagain case in trywrite
4 years ago
Dirk Engling
e33efb5086
Add newline to debug string
4 years ago
Dirk Engling
0178c27662
Merge branch 'split-iobatches' of erdgeist.org:opentracker into split-iobatches
4 years ago
Dirk Engling
dfe687aa81
Remove old mman header
4 years ago
Dirk Engling
397cd035a4
Turn mmaps into malloc
4 years ago
Dirk Engling
2f228c27c2
Retry successful writes immediately
4 years ago
Dirk Engling
95f1780f0b
Split huge iovecs over multiple io_batches
4 years ago
Dirk Engling
e87978860b
gzip iovecs always end on the boundary and don't need to be fixed
4 years ago
Dirk Engling
1a349bfa43
Add a handler for the EAGAIN reason code to iob_send
4 years ago
Dirk Engling
bfc398182f
Rework fullscrape worker, unifying non-gzip and gzip code was a bad idea
4 years ago
Dirk Engling
27f8189d84
accesslist checker should not operate on an empty list
4 years ago
Dirk Engling
58dedd001d
srandomdev is not available on linuxes by default
4 years ago
Dirk Engling
b73b3b17cc
Use arc4random whereever we need strong entropy
4 years ago
Dirk Engling
ccef1d0ccd
Use dev random by default
4 years ago
Dirk Engling
616119ee22
Turn random() to nrand48 in inner loop where it is not cryptographically relevant to avoid futex storm. Thanks and credits to Bart Smienk
4 years ago
Dirk Engling
0695b48870
White space
4 years ago
Dirk Engling
00bd3b2009
unmap file if a memory allocation error occurs
4 years ago
Dirk Engling
102ba9075b
Make accesslist reload thread safe. The last commit actually would make a free possible while another thread was bsearching that memory
4 years ago
Dirk Engling
fde79836e6
Try accessing the access lists without locks by making the replacement process as atomic as possible.
4 years ago
Dirk Engling
7c905ba729
De-bottleneck mutex access code
4 years ago
Dirk Engling
6411f1567f
Remove debug code for self pipes
7 years ago
Dirk Engling
d6d4fb0668
Tell libowfat about EWOULDBLOCK conditions
7 years ago
Dirk Engling
48c9c6172b
register both ends of self pipe to libowfat in order to use them
7 years ago
Dirk Engling
9c7be324f5
break belongs inside the guard
8 years ago
Dirk Engling
d1e6e4486c
incorporate a more verbose ascii dump, provided by Tom <tom@foscore.com>
8 years ago
Dirk Engling
d4598cc930
Get rid of some warnings regarding header guards
8 years ago
Dirk Engling
0ebc0ed6a3
Avoid double free when iob_addbuf_free fails, thanks to Sami Farin
8 years ago
Dirk Engling
ba25d2b2a8
Forgot to include one error code. Thanks to Lars Seipel
9 years ago
Dirk Engling
75c216c82d
dont count scrapes as full scrapes, fall through fail
9 years ago
Dirk Engling
a215479d1a
Accept was counted twice
9 years ago
Dirk Engling
a06c3cfeb7
Add missing enum value. Thanks to zhangdongmao
9 years ago
Dirk Engling
fca2b993a9
Increase number of locks to accomodate for more cores on contemporary CPUs
10 years ago