Version-Release number of selected component: wget-1.14-3.fc18 Additional info: libreport version: 2.0.18 abrt_version: 2.0.18 backtrace_rating: 4 cmdline: /usr/bin/wget '--header=Content-Type: multipart/form-data; boundary=--------------------------4a1c55cf63ecfea6' --post-file=/tmp/chromium-upload-22235e9a6125103d https://clients2.google.com/cr/report --timeout=10 --tries=1 -O /dev/fd/3 crash_function: sk_value kernel: 3.6.7-5.fc18.x86_64 truncated backtrace: :Thread no. 1 (10 frames) : #0 sk_value at stack.c:303 : #1 x509_name_encode at x_name.c:266 : #2 x509_name_ex_i2d at x_name.c:229 : #3 ASN1_item_ex_i2d at tasn_enc.c:181 : #4 asn1_item_flags_i2d at tasn_enc.c:122 : #5 ASN1_item_i2d at tasn_enc.c:91 : #6 i2d_X509_NAME at x_name.c:129 : #7 X509_NAME_cmp at x509_cmp.c:199 : #8 OBJ_bsearch_ex_ at obj_dat.c:700 : #9 internal_find at stack.c:229
Created attachment 650980 [details] File: core_backtrace
Created attachment 650981 [details] File: environ
Created attachment 650982 [details] File: limits
Created attachment 650983 [details] File: backtrace
Created attachment 650984 [details] File: cgroup
Created attachment 650985 [details] File: executable
Created attachment 650986 [details] File: maps
Created attachment 650987 [details] File: dso_list
Created attachment 650988 [details] File: proc_pid_status
Created attachment 650989 [details] File: var_log_messages
Created attachment 650990 [details] File: open_fds
To me it looks more like a bug in OpenSSL library. Changing component to openssl.
Can you reproduce this bug with regular wget call to https:// URL? I think this is some kind of memory corruption before the OpenSSL is called. The reason is the backtrace is from OpenSSL initialization and if the bug was really there basically anything that uses OpenSSL for SSL calls would be broken.
OpenSSL works fine and I couldn't reproduce the bug with wget https URLs This bug occurred when the google chrome browser tried to upload a crash report to google.
*** Bug 892993 has been marked as a duplicate of this bug. ***
Back to wget as I believe this is not a bug in OpenSSL but rather some incorrect call of OpenSSL from wget or memory corruption from wget.
I noticed more strange memory corruption in other programs recently and as it turned out I had a defective RAM module. Effects were very random and rare but could be the cause of most memory corruption I lately reported. I assumed it is related to the beta state of F18 so I didnt suspect the RAM at first but MemTest found a small address range with errors. I am extremely sorry for taking up your time hunting down bugs which are most likely not there. I decided to close my latest bug reports as I can not assure that the memory corruption was actually caused by the program rather than the defect RAM module.