<html> <head> <base href="https://bugs.freedesktop.org/"> </head> <body> <div> <a class="bz_bug_link bz_status_ASSIGNED " title="ASSIGNED - FcCacheOffsetsValid() couldn't detect broken cache file" href="https://bugs.freedesktop.org/show_bug.cgi?id=103237#c10">Comment # 10</a> on <a class="bz_bug_link bz_status_ASSIGNED " title="ASSIGNED - FcCacheOffsetsValid() couldn't detect broken cache file" href="https://bugs.freedesktop.org/show_bug.cgi?id=103237">bug 103237</a> from <a class="email" href="mailto:youngb.shin@samsung.com" title="Youngbok Shin <youngb.shin@samsung.com>"> Youngbok Shin</a> <pre>> Hmm, does it guarantee the file size? Do you mean fsync() guarantees synchronization without considering its file size? Then, as I know, yes. It writes a whole buffer to the file descriptor at once. I'll be bad for performance. I think even if we can see non-zero values at head of invalid cache file, we still need to consider the file corruption by power outage. If a file system couldn't create a whole file or write anything on a file by power outage, we wouldn't call it "corruption". Here is a Q&A about these scenarios. <a href="https://serverfault.com/questions/403891/journaled-filesystems-and-power-failure">https://serverfault.com/questions/403891/journaled-filesystems-and-power-failure</a> If we can't allow bad performance by calling fsync() for writing the big cache files (to prevent very very rare file corruption issues), I think we need to add more validation code. Just like your temporary patch for the issue.</pre> </div> <hr> You are receiving this mail because: <ul> <li>You are on the CC list for the bug.</li> </ul> </body> </html>