256 bytes) caused it to break on many devices.
The SCSI spec says that for commands with 8-bit length fields, a value of 0
means 256 bytes. As it turns out, many devices don't deal with that
properly. Some interpret the 0 as 0, and return no data. Others return
more than 256 bytes of data, and cause an overrun.
The fix is to tell the device we've only allocated SHORT_INQUIRY_LENGTH (36
bytes) of inquiry data, instead of sizeof(struct scsi_inquiry_data).
camcontrol.c: Change inq_len in the call to scsi_inquiry() to
SHORT_INQUIRY_LENGTH, and add a long comment
explaining the reason for the change.
scsi_all.h: Add a comment above the definitinon of
SHORT_INQUIRY_LENGTH alerting people that it is
both the initial probe inquiry length, and the
minimum amount of data needed for scsi_print_inquiry()
to function.
scsi_all.c: Add a comment about SHORT_INQUIRY_LENGTH being the
minimum amount of data needed for
scsi_print_inquiry() to function.
Reviewed by: gibbs
Approved by: jkh
Reported by: "John W. DeBoskey" <jwd@unx.sas.com>
a mess in securelevel environments. Results in one warning during
/etc/rc as it attempts to remove file flags, but this is harmless.
Approved by: High Lord Hubbard
from useful drivers such as the 3D DRI drivers I will be porting for
hardware accelerated OpenGL. The hardware will still be reported during
boot using the nomatch system.
Approved by: jkh
was needed to make attach/detach of devices work, which is
needed for the PCCARD support.
(PCCARD support is still not working though, more to come on that)
Support the CMD646 chip which is used on many alphas, sadly only
in WDMA2 mode, as the silicon is broken beyond belief for UDMA modes.
Lots of cosmetic fixes here and there.
Sorry for the size of this megapatchfromhell but it was not
possible otherwise...
newbus patches based on work from: dfr (Doug Rabson)
accept a new command; in high load cases it may be too busy for the old
value.
This loop needs something to tie it to real time, rather than just the CPU's
ability to fetch from the L1 data cache, but this hack works for now.
Approved by: jkh
firmware prompt. Several sleepy folk mistook the '>>>' for the SRM
prompt, which was never the desired idea.
Submitted by: Andrew Gallatin <gallatin@cs.duke.edu>
Approved by: jkh
fixes some namespace pollution in general and breakage of modules that
aren't in the sys tree in particular (<machine/ipl.h> includes further
headers that aren't installed under /usr/include).
Reimplemented SPLASSERT() so that it is more machine independent and
less bloated and doesn't require the <machine/ipl.h> include spam.
In particular, don't assume that `cpl' can be printed using %08x
format. The alpha arch doesn't even have `cpl'. SPLASSERT() was
harmless on alphas because it isn't actually used.
run out of KVM through a mmap()/fork() bomb that allocates hundreds
of thousands of vm_map_entry structures.
Add panic to make null-pointer dereference crash a little more verbose.
Add a new sysctl, vm.max_proc_mmap, which specifies the maximum number
of mmap()'d spaces (discrete vm_map_entry's in the process). The value
defaults to around 9000 for a 128MB machine. The test is scaled for the
number of processes sharing a vmspace (aka linux threads). Setting
the value to 0 disables the feature.
PR: kern/16573
Approved by: jkh
1) Fix a bug in the int15 function 87 emulation where we only copied half
of what the BIOS asked for. This caused the Mylex RAID adapter to go
haywire and start trashing memory when you tried to boot from it.
2) Don't use interrupt 19 to reboot. Instead, set the reboot flag to a warm
boot and jump to the BIOS's reboot handler. int 19 doesn't clear memory
or restore the interrupt vector table, and thus really isn't safe. For
example, when booting off of PXE, the PXE BIOS eats up a chunk of memory
for its internal data and structures. Since we rebooted via int 19,
using the 'reboot' command in the loader resulted in that memory not
being reclaimed by the BIOS. Thus, after a few PXE boots, the system
was out of lower memory.
3) Catch any int 19 calls made by a BTX client or a user pressing
Ctrl-Alt-Delete and shutdown BTX and reboot the machine cleanly. This
fixes Ctrl-Alt-Delete in the loader and in boot2 instead of presenting
the user with a BTX fault.
Approved by: jkh
Found by: 1) by msmith
This unspams the boot messages, concentrating on the drivers that have
actually been probed.
This basically resurrects revision 1.106 from old /sys/i386/isa/isa.c.
Reviewed by: jkh, dfr
code & it survives a buildworld. So remove the dire warnings about
Noritake support being untested.
o Remove a disconserting printf() left over from NetBSD
Approved by: jkh
1) Non-AST4 multiport cards were broken by bypassing the code that changes
`idev' to the multiport master device.
2) AST4 multiport cards apparently were broken by inverting the test for
the master device having an irq.
3) Error handling for nonexistent master devices was broken by removing a
check for a null pointer.
4) `int' error codes returned by bus_get_resource() were assigned directly
to the boolean variable com->no_irq. Probably harmless, since the
boolean is implemented as a u_char.
Submitted by: part 1) by Chris Radek <cradek@in221.inetnebr.com>
part 2) by yokota
Approved by: jkh
curproc. This only makes any difference on SMP, where we used a
(potentially very bogus) switchtime from our own CPU to calculate
resource usage on another CPU.
This should remove some if not all calcru() related warnings on SMP.
Approved by: jkh
#! /bin/sh # -*- perl -*-
This is simply "delete everything after the next '#', not counting the
first char in the line". No effort has been made to allow quoting,
backslash escaping or '#' in interpreter names.
The complies to POSIX 1003.2 in that Posix says the implementation is
free to choose whatever it likes.
PR: bin/16393
Andrew's problems with SCSI on some alphas- do not call isp_update
directly to update parameters- just mark them as being ready to
update for the next command- the system would just hang on a READ
CAPACITY for a drive. Really annoying because it wouldn't even timeout
(and it has a timeout) so either the SET PARAMETERS call was nuking
things or the f/w was really dropping the ball.
approved: jkh
Reviewed by: gallatin@freebsd.org
is gone as a define. We just don't support fast posting for anything less
than the 1240/1080/1280/12160 or Fibre Channel cards.
Put in support for CDB's larger than 12 bytes for parallel SCSI (up to 44
bytes are allowed).
Approved: jkh
- Remove all the code intended to deal with experimental
C1010 revisions. This code got useless due to commercial
chip revisions having been fixed.
Fixes:
- Rewrite/rework the WSR condition handling.
Previous drivers snooped on the BUS through the SBDL IO
register and this has been discovered to trigger a spurious
SCSI parity error when WSR had been set by chip and cleared
by SCRIPTS prior to reading SBDL bit [0...7].
On the other hand, the C1010 does not use the SWIDE register
when synchronous data transfers are taking place and
requires a CHMOV (1) WHEN DATA_IN to be performed in order
to move to memory the residual byte when WSR is set and
the residual byte is useful data.
BTW, the new WSR handling by the driver is simpler.
- No longer attempt to read from SCRIPTS the SBDL register.
This is intended to avoid to be victimized again by any other
issue regarding the handling of this register by 8xx chips.
Miscellaneous:
- The driver is now able to handle the WSR + IGN RESIDUE
condition at the end of a DATA IN I/O without need of a
programmed interrupt. It is a minor? optimization.
- A few other minor cosmetic changes.
This driver version fixes notably a permanent SCSI parity
error condition at boot that can be triggerred due to recent
changes in cam_xpt.c between 1.79 and 1.80.
Changes in CAM/XPT are fine, but the new handling of the full
INQUIRY may trigger the driver problem when a target returned
an odd value in the `additionnal length' field of the INQUIRY
response.
The diff against previous driver version is large, but it
consists approximatively in:
- 350 lignes removed and not compiled in previous drivers
(They addressed experimental C1010 revisions)
- 250 lignes added or changed, half being comments or empty
lines.
So, in fact, the real changes are about 120 lines of source.
About 80 lines address SCRIPTS changes and about 40 lines
address C code changes.
Approved by: jkh
New WSR handling reviewed by Pamela Delaney <pam.delaney@lsil.com>
(For back-porting to Linux sym53c8xx driver 1.6x series)
``jail'', and move the set_hostname_allowed sysctl there, as well as
fixing a bug in the sysctl that resulted in jails being over-limited
(preventing them from reading as well as writing the hostname). Also,
correct some formatting issues, courtesy bde :-).
Reviewed by: phk
Approved by: jkh
for 1020/1X80/12160/2X00- for readability. Add in 12160 (Ultra3)
support- but not with PPR just yet. Fix and clarify fetching of
return parameter for getting firmware rev which for the 2200 contains
the connection topology (Private Loop (NL-port), N-port, FL-port,
F-port). Synthesize the connection topology for the 2100 which can
only be Private Loop or FL-port. Handle a couple of new async
mailbox commands which signify connection in Point-to-Point mode
(N-port or F-port) or indicate various toe stubbing getting to same.
Approved: jkh@freebsd.org
- trying for a fit for a PnP configuration from a device
- soaking up resources from a configuration that were not allocated by
the driver
do not attempt to activate them. Only a device driver that is aware of
the nature of the resource and its suitability can be certain that
activating a resource, particularly a memory resource, is a safe thing
to do.
This was prompted by the discovery that many systems report all physical
memory through a PNP0c02 device; activating this resource maps all
physical memory into the kernel's virtual space, either blowing out the
kernel pagetable or in the worst case causing a panic in pmap_mapdev()
if the system has too much physical memory.
Authorised by: jkh
Reviewed by: dfr
won't appear on the screen, and "blinking" and "destructive" cursor
won't appear in the vty for which the text cursor is currently hidden.
Approved by: jkh
Some of IPsec authentication related functions should have
'const' for its 2nd argument, but not now.
But if someone try to use them, and passed const data for
those functions, then much bogus compile warnings will be
generated.
So those funcs prototype should be modified.
Requested by: archie
Approved by: jkh
KAME put INET6 related stuff into sys/netinet6 dir, but IPv6
standard API(RFC2553) require following files to be under sys/netinet.
netinet/ip6.h
netinet/icmp6.h
Now those header files just include each following files.
netinet6/ip6.h
netinet6/icmp6.h
Also KAME has netinet6/in6.h for easy INET6 common defs
sharing between different BSDs, but RFC2553 requires only
netinet/in.h should be included from userland.
So netinet/in.h also includes netinet6/in6.h inside.
To keep apps portability, apps should not directly include
above files from netinet6 dir.
Ideally, all contents of,
netinet6/ip6.h
netinet6/icmp6.h
netinet6/in6.h
should be moved into
netinet/ip6.h
netinet/icmp6.h
netinet/in.h
but to avoid big changes in this stage, add some hack, that
-Put some special macro define into those files under neitnet
-Let files under netinet6 cause error if it is included
from some apps, and, if the specifal macro define is not
defined.
(which should have been defined if files under netinet is
included)
-And let them print an error message which tells the
correct name of the include file to be included.
Also fix apps which includes invalid header files.
Approved by: jkh
Obtained from: KAME project
problem was basically (for offset > 4096):
vtophys(addr) + offset != vtophys(addr + offset)
Also, use TD's with a maximum size of 4k instead of 8kb for OHCI
controllers.
This problem occurs in drivers that use large transfer sizes:
umass, host2host and ethernet with jumbo frames.
or not a process in a jail, with privilege, may set the jail's hostname.
Defaults to 1, which permits this. May be set to 0 by a process with
appropriate privilege outside of jail. Preventing hostname renaming
from within a jail is currently required to make jails manageable, as they
a currently identifiable only by hostname using /proc, which may be
modified without this sysctl being set to 0. This will be documented
in upcoming man commits.
Authorized by: jkh, the ever-patient
for optimizing the unpause operation no-longer exist, and this is much
safer.
When restarting the sequencer, reconstitute the free SCB list on the card.
This deals with a single instruction gap between marking the SCB as free
and actually getting it onto the free list.
Reduce the number of transfer negotiations that occur. In the past, we
renegotiated after every reported check condition status. This ensures
that we catch devices that have unexpectidly reset. In this situation,
the target will always report the check condition before performing a
data-phase. The new behavior is to renegotiate for any check-condition where
the residual matches the orginal data-length of the command (including
0 length transffers). This avoids renegotiations during things like
variable tape block reads, where the check condition is reported only
to indicate the residual of the read.
Revamp the parity error detection logic. We now properly report and
handle injected parity errors in all phases. The old code used to hang
on message-in parity errors.
Correct the reporting of selection timeout errors to the XPT. When
a selection timeout occurs, only the currently selecting command
is flagged with SELTO status instead of aborting all currently active
commands to that target.
Fix flipped arguments in ahc_match_scb and in some of the callers of this
routine. I wish that gcc allowed you to request warnings for enums passed
as ints.
Make ahc_find_msg generically handle all message types.
Work around the target mode data-in wideodd bug in all non-U2 chips.
We can now do sync-wide target mode transfers in target mode across the
hole product line.
Use lastphase exclusively for handling timeouts. The current phase
doesn't take the bus free state into account.
Fix a bug in the timeout handler that could cause corruption of the
disconnected list.
When sending an embedded cdb to a target, ensure that we start on a
quad word boundary in the data-fifo. It seems that unaligned stores
do not work correctly.
chipset to be probed & attached on newer Dell PowerEdge servers, such as
the 2400 and 4400.
Reviewed by: dfr, msmith, jlemon
Tested by: hnokubi@yyy.or.jp (in a previous incantation)
Approved by: jkh
ptys in ways that might be unethical, especially towards processes not in
jail, or in other jails.
Submitted by: phk
Reviewed by: rwatson
Approved by: jkh
at the same time.
When rfc1323 and rfc1644 option are enabled by sysctl,
and tcp over IPv6 is tried, kernel panic happens by the
following check in tcp_output(), because now hdrlen is bigger
in such case than before.
/*#ifdef DIAGNOSTIC*/
if (max_linkhdr + hdrlen > MHLEN)
panic("tcphdr too big");
/*#endif*/
So change the above check to compare with MCLBYTES in #ifdef INET6 case.
Also, allocate a mbuf cluster for the header mbuf, in that case.
Bug reported at KAME environment.
Approved by: jkh
Reviewed by: sumikawa
Obtained from: KAME project
(1)When all related IPv6 addresses are removed,
then remove the associated IPv6 prefix.
(2)When multiple IPv6 link local addrs exist for a same
interface , then let its IPv6 prefix have multiple
interface id, and create multiple IPv6 global addrs with same
interface id.
(3)When a new IPv6 link local addr is assigned for an
interface, then let its IPv6 prefix also have the
interface id of the new IPv6 link local addr, and
create new IPv6 global addrs with same interface id.
Approved by: jkh
KAME source addr selection rule had a problem to treat IPv6 site
local addr.
The rule is completely rewritten recently and the above problem
is also fixed, but rewriting same code part in freebsd4.0 is too
dangerous in this stage, so just add workaround to avoid
the problem. Just add code for IPv6 site local addresses into IPv6
source addr selection algorythm part.
is initialised by usb_init.
This might solve problems with some controllers not being initiliased
properly, because a delay was effectively a tsleep that returned
immediately.
Approved by: jhk
Up to now, errors from DEVICE_SUSPEND(root_bus) were ignored. The fix for
this problem (the introduction of defaults for device methods) has been
committed months ago by Doug Rabson.
Second, the suspended devices were not always properly resumed on error.
Third, swapped the order for calling restore hooks and restore methods, to
be in line with the cases above.
Reviewed by: Doug Rabson
Approved by: jhk
the PIRQD bit.
This fixes the problem of uhub0 hanging forever during boot when USB
keyboard support is switched on in the BIOS on motherboards with Intel
chipsets (UHCI).
Approved by: The Sheep
fix support for multiple HPT & Promise controllers.
support mixed 33/66 devices on the Promise 66 controllers.
fix the refcount stuff in the atapi drivers.
misc cleanups.
support is also included for the ASC38C1600 Ultra160 chipset, but
as firmware is not yet available for this chip, it is disabled.
Approved by: jkh@FreeBSD.org
attempt to copy the ethernet header forward and otherwise encapsulate
a packet for output.
This fixes the panic when using VLAN devices on hardware that doesn't
do 802.1Q tagging onboard. (That is to say, all drivers except the Tigon.)
My tests consisted of telnet, ttcp, and a pingflood of packets
between 1 and 1600 (plus headers) bytes.
MFC to follow in 1 week.
Approved by: jkh
Also update wicontrol to enable/disable encryption, set WEP keys and set the
TX key index. Silver cards only have 40-bit keys. This is something of a quick
hack, but it works well enough for me to commit this from the LinuxWorld
exhibit floor.
The WEP support only shows up if you have a card that supports it.
Would have been approved by: jkh, if he hadn't wandered off somewhere
Approved in his place by: msmith, who's standing right here
conversion to eliminate the compatability shims without making any
significant changes. This eliminates the shim warnings.
Obtained from: n_himba (tweaked by me, don't blame him for this)
Approved by: jkh
Some LAN card chip for fxp is known to cause problem at
interface initialization with IPv6 enabled. It happens at
some delicate timing.
And also, just adding some DELAY before IPv6 address
autoconfiguration is known to avoid the problem.
Complete fix is changing the driver not to use interrupt at
multicast filter initialization, but trying such change in
this stage will be dangerous.
So I add some DELAY() only inside #ifdef INET6 part,
as temporal workaround only for 4.0.
Approbed by: jkh
Noticed by: Mattias Pantzare <pantzer@ludd.luth.se>
Obtained from: openbsd-tech mailing list
bus_delete_resource.
Fixes a problem when the probe succeeded, but the attach failed. The
release of the resources was done inproperly.
Approved by: jkh
have a write in progress. Otherwise one can get in an infinite loop
trying to get them all flushed.
Submitted by: Matthew Dillon <dillon@apollo.backplane.com>
be safe enough to recommend for working around the problem with CHS mode
normally being limited to 65536*16*63 sectors.
Fixed bitrot in related debugging code.
Fixed related style bugs.
Removed another vestige of bad144 support.
actually used it since the only device that specified it (vga0) was marked
as "FLG_INVISIBLE" in userconfig and therefore never shown.
Suggested by: bde
