Changes from 1.1.94 to 1.1.95

These changes are archived on ftp.crynwr.com:pub/kchanges and http://www.crynwr.com/kchanges. I cannot answer questions about Linux -- I merely summarize the kernel patches after reading them. I try to make them useful to everyone, but kernel hackers should take them with a grain of salt and read the patches themselves. The patch file is patch95.gz

• More Alpha changes.
• Make sure that the arch/i386 subdirectories are made correctly.
• Register the standard I/O devices with /proc/ioports.
• Dump the real eax when doing a i386 register dump.
• Comment out the thread cloning code.
• Oops, the ptrace page middle code had a bug.
• Sparc SUN4M support and Sparc Reference MMU added to sparc/config.in.
• IDE driver adds support for non-standard 3rd/4th drive interface on Promise cards.
• Remove excess "multiple mode disabled" messages from IDE driver.
• Floppy driver now has different select delays for each floppy drive type.
• Yeeks! Bad parenthesization in console driver causes screen inversion on monochrome systems to fail.
• Shift-NumLock now lets you work in application keypad mode and still be able to change the NumLock flag.
• ARCNET driver should probe better. Modularized. Handles more protocol ID's correctly. Does both transmit and receive IRQ's.
• Fix to depca driver for conflict with VESA SCSI.
• Free the SLIP buffers when the channel is closed down, not when the device is closed.
• Qlogic driver now supports the Control Concepts ISA board (see source).
• Qlogic driver now defaults to slower I/O for more reliability.
• Seagate driver cleaned up.
• IN_CLASSC (detect a class C network) define was wrong.
• IP Firewalling updated to FreeBSD-Current (but FreeBSD version isn't really fully baked).
• More PCI device and vendor IDs added.
• Default maximum number of tasks changed from 128 to 512.
• Make sure the shared memory handler return ENOMEM if it should.
• Eliminate a few race conditions in the forking code.
• Send INT signal to init, and HUP signals to all other processes.
• Implement max-length ("precision") printing in kprintf.
• Bug fixed in memory manager, which pointed out bug in swapper.
• bind() shouldn't abort existing but dead sockets. Stops FTP netin:.. I hope.
• bind() works correctly for RAW sockets. Note that FreeBSD at least is broken in this respect so be careful with compatibility tests...
• Don't do ARP proxy across hardware types!
• Update the device on a real multicast group delete rather than any time but...
• Take out some root-only checks from IPX code.

[Index][Next]