Message-Id: <5.0.2.1.0.20010108231043.03469460@pop5.banet.net> X-Sender: usbanet DOT farley3 AT pop5 DOT banet DOT net X-Mailer: QUALCOMM Windows Eudora Version 5.0.2 Date: Mon, 08 Jan 2001 23:21:43 -0500 To: Eli Zaretskii From: "Peter J. Farley III" Subject: Re: Fw: Patch for statfs.c Cc: djgpp-workers AT delorie DOT com, Martin Str|mberg , ceo AT nbensacomputers DOT com In-Reply-To: References: <5 DOT 0 DOT 2 DOT 1 DOT 0 DOT 20010107121357 DOT 00aa6580 AT pop5 DOT banet DOT net> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Reply-To: djgpp-workers AT delorie DOT com Errors-To: nobody AT delorie DOT com X-Mailing-List: djgpp-workers AT delorie DOT com X-Unsubscribes-To: listserv AT delorie DOT com Precedence: bulk At 09:00 AM 1/8/01 +0200, Eli Zaretskii wrote: > >On Sun, 7 Jan 2001, Peter J. Farley III wrote: > >> Note that the rebuilt df reports too many 1024 blocks, 326552*1024 = >> 334389248 while 325888*1024 = 333709312. The difference is 664 >blocks, >> which I cannot explain easily. > >Since two versions of statfs which are supposed to be identical return >different values, I think the only way to find out why is to step inside > >both versions with a debugger and see where does the difference come >from. Perhaps there's some bug in how the transfer buffer is layed out >when passing requests to the CDROM driver vie Int 2Fh. Could it be that fil316b.zip was built with the older version of statfs that is packaged in the fil316s.zip archive (which might explain the difference)? Can objdump be used to extract just one function's code from an executable? If so, what are the switches to use? I've reviewed the objdump info and there doesn't seem to be a way to specify the *function* name to be extracted (unless I'm blanking out and not seeing it), just the "section", which apparently means things like ".text", ".bss", etc. If we could extract just the statfs function, we could diff the output of objdump of the statfs code from the distributed binary and the current libc version of statfs to see how they are different. I can and will also test that old version of statfs from fil316s.zip by itself, to see if it returns the values I see from the distributed df.exe. >> Filesystem 1024-blocks Used Available Capacity Mounted on >> EXCALIBUR_16X9_FF_NA 6729942 6729942 0 100% y:/ >[snip] >> 0 file(s) 0 bytes >> 1 dir(s) 0.00 MB free >> 6,572.21 MB total disk space, 100% in use > >Please make a point of showing the numbers in the same units. It is >very >annoying to grab a calculator or invoke Calc each time to compare the >reported sizes. > >According to my calculations, 6,572.21 MB is exactly 6729943 bytes, so >these two numbers are consistent. Is that what you wanted to say? Yes, that is what I wanted to say. I will recalculate the numbers as you request in future posts. --------------------------------------------------------- Peter J. Farley III (pjfarley AT dorsai DOT org OR pjfarley AT banet DOT net)