Hi,
since there have been several reports of the ROS CDs failing to boot with the "Invalid partition table magic" message I though if the line 85 in freeldr/fs/fs.c can be the cause. If the CD-ROM drive would get assigned the 0x80 BIOS number (can this everr happen?) then we would end up reading partition table from it and crash later. Any thoughts?
Regards, Filip
Index: freeldr/fs/fs.c =================================================================== RCS file: /CVS/ReactOS/freeldr/freeldr/fs/fs.c,v retrieving revision 1.14 diff -u -r1.14 fs.c --- freeldr/fs/fs.c 25 May 2004 21:31:27 -0000 1.14 +++ freeldr/fs/fs.c 30 Oct 2004 15:01:41 -0000 @@ -82,7 +82,7 @@ }
// Check for ISO9660 file system type - if (DriveNumber > 0x80 && FsRecIsIso9660(DriveNumber)) + if (DriveNumber >= 0x80 && FsRecIsIso9660(DriveNumber)) { DbgPrint((DPRINT_FILESYSTEM, "Drive is a cdrom drive. Assuming ISO-9660 file system.\n"));
"Filip Navara" xnavara@volny.cz wrote:
Hi,
since there have been several reports of the ROS CDs failing to boot with the "Invalid partition table magic" message I though if the line 85 in freeldr/fs/fs.c can be the cause. If the CD-ROM drive would get assigned the 0x80 BIOS number (can this everr happen?) then we would end up reading partition table from it and crash later. Any thoughts?
This shouldn't happen because the specifiation says thea the BIOS number of a bootable CD-ROM drive must be 0x81 and above. I have seen BIOS numbers like 0xA1 but OTOH there are lots of broken BIOSes out there.
Regards, Eric