Sigh. Turns out the "unwind" cleanup in ufs_mkdir() and ufs_makeinode() has changed, but the MAC code was using the old unwinding rather than the new 'bad' label. The EA code had an error case I forgot about, and so I bumped into the bad unwinding code, causing the panic. Now fixed in the trustedbsd_mac tree -- sorry for the false alarm.
Robert N M Watson FreeBSD Core Team, TrustedBSD Project [EMAIL PROTECTED] NAI Labs, Safeport Network Services On Fri, 28 Dec 2001, Robert Watson wrote: > > Recently got this: > > ./account missinmg (directory noto created: Deviced not configured)e > = 040700, inum = 6, fs = /var > panic: ffs_valloc: dup alloc > Debugger("panic") > Stopped at Debugger+0x44: pushl %ebx > db> trace > Debugger(c03a3b9b) at Debugger+0x44 > panic(c03bb9e1,c03bb9c0,41c0,6,c22ab0d4) at panic+0x70 > ffs_valloc(cb18c920,41c0,c22c3d00,cb14caa0,cb14cc18) at ffs_valloc+0x129 > ufs_mkdir(cb14cc18,cb14ccd4,c0274066,cb14cc18,c9fe0014) at ufs_mkdir+0x85 > ufs_vnoperate(cb14cc18,c9fe0014,c9fdfe00,c9fdff04,c2420200) at > ufs_vnoperate+0x15 > vn_mkdir(804fd80,1c0,0,c9fdff04,cb14cd40) at vn_mkdir+0x19e > mkdir(c9fdff04,cb14cd20,804fd82,8058100,0) at mkdir+0x15 > syscall(2f,2f,2f,0,8058100) at syscall+0x24c > syscall_with_err_pushed() at syscall_with_err_pushed+0x1b > --- syscall (136, FreeBSD ELF, mkdir), eip = 0x280d1617, esp = 0xbfbffbb4, > ebp = 0xbfbffc50 --- > db> > > This is in a fairly modified MAC-based host I'm using, so it could be my > own fault, but I was under the impression that the dup alloc bugs had been > fixed? I'm not sure how far integ'd the MAC tree is, but should be > withing the last week. > > Robert N M Watson FreeBSD Core Team, TrustedBSD Project > [EMAIL PROTECTED] NAI Labs, Safeport Network Services > > > To Unsubscribe: send mail to [EMAIL PROTECTED] > with "unsubscribe freebsd-current" in the body of the message > To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-current" in the body of the message