http://nvd.nist.gov/nvd.cfm?cvename=CVE-2007-3023 "unsp.c in ClamAV before 0.90.3 and 0.91 before 0.91rc1 does not properly calculate the end of a certain buffer, with unknown impact and remote attack vectors." http://nvd.nist.gov/nvd.cfm?cvename=CVE-2007-3024 "libclamav/others.c in ClamAV before 0.90.3 and 0.91 before 0.91rc1 uses insecure permissions for temporary files that are created by the cli_gentempstream function in clamd/clamdscan, which might allow local users to read sensitive files." http://nvd.nist.gov/nvd.cfm?cvename=CVE-2007-3122 "The parsing engine in ClamAV before 0.90.3 and 0.91 before 0.91rc1 allows remote attackers to bypass scanning via a RAR file with a header flag value of 10, which can be processed by WinRAR." http://nvd.nist.gov/nvd.cfm?cvename=CVE-2007-3123 "unrar.c in libclamav in ClamAV before 0.90.3 and 0.91 before 0.91rc1 allows remote attackers to cause a denial of service (core dump) via a crafted RAR file with a modified vm_codesize value, which triggers a heap-based buffer overflow." Not checked whether 0.88.x in FC-6 and earlier are affected.
One more that I didn't find in Bugzilla and apparently affects 0.90.x series earlier than 0.90.3: http://nvd.nist.gov/nvd.cfm?cvename=CVE-2007-2029 "File descriptor leak in the PDF handler in Clam AntiVirus (ClamAV) allows remote attackers to cause a denial of service via a crafted PDF file."
Another problem is, that freshclam is writing to logs this: Reading CVD header (daily.cvd): Ignoring mirror 212.7.0.71 (too often connection s with outdated version) I think updating is not working properly. This does not happen always and I am unable to reproduce this on other machine. And last problem, libclamav is loading it's database a long time (aprox. 2-5 minutes). This bug can be also fixed with upgrade to clamav 0.91. What is the problem with releasing clamav-0.91.2 built in Koji? How can I help to release clamav updates with less time? My Fedora system is vulnerable to some vulnerabilities with old clamav. :-(