Trond Myklebust [Tue, 7 Oct 2003 16:51:33 +0000 (12:51 -0400)]
NFSv4 state model update
- Hierarchy of state attached to nfs4_client in order to
simplify state recovery.
state_owners hang off nfs4_client, whereas state hangs
off both open_owners and the nfs_inode.
- Model tries to minimize the number of open_owners on
the server by recycling unused open_owners into a pool.
- NFSv4 state attached to file->private_data. Previously
this was used by credentials (and still is for NFSv2/v3)
Abstract out setup/release of struct file and nfs_page
structure initialization in order to cope with these
conflicting uses of private_data.
Trond Myklebust [Tue, 7 Oct 2003 11:59:42 +0000 (07:59 -0400)]
Simplify the synchronous NFS read call interface by
passing a pointer to a filled nfs_read_data structure
(the same struct used by the asynchronous function calls)
Trond Myklebust [Tue, 7 Oct 2003 11:50:48 +0000 (07:50 -0400)]
The NFSv4 state model assumes that the client machine identifies
itself to the server once and once only.
This patch ensures that we do this by sharing the local identifier
struct nfs4_client among all mountpoints that talk to the same
server/ip address.
Trond Myklebust [Tue, 7 Oct 2003 11:35:49 +0000 (07:35 -0400)]
Fix up hangs with the upcall mechanism for RPCSEC_GSS and the
NFSv4 idmapper.
- Correct bugs in idmapper wait code.
- Correct layering bugs in RPCSEC_GSS and idmapper
- Ensure we advance the file pointer in case of partial
reads/writes of a message on the pipes
Trond Myklebust [Tue, 7 Oct 2003 11:30:00 +0000 (07:30 -0400)]
UDP round trip timer fix. Modify Karn's algorithm so that
we inherit timeouts from previous requests.
This means that we lengthen the window of time during which
we accept updates to the RTO estimate if we see an update.
Scheme proposed by Brian Mancuso, but it is standard for TCP
congestion control implementations.
Chen Yang [Tue, 7 Oct 2003 02:15:15 +0000 (19:15 -0700)]
[PATCH] InterMezzo maintainence patch
I have tested these patches and can atteste that they work.
- rwhron@earthlink.net:
include cleanup of fs/intermezzo
- kdw@neowiz.com:
memory leakage problem,race condition report
- Domen Puncer:
Use list_for_each() where applicable
- Chen Yang:
include file modification for more platforms to compile intermezzo
fix an oops problem
Gerd Knorr [Tue, 7 Oct 2003 00:53:45 +0000 (17:53 -0700)]
[PATCH] v4l: bttv driver update
This updates the bttv driver. It depends on the videobuf patch.
* usual tv card list updates.
* sysfs adaptions.
* new, experimental i2c adapter code for bt878 chips
(not used by default yet).
* various minor fixes.
Gerd Knorr [Tue, 7 Oct 2003 00:53:35 +0000 (17:53 -0700)]
[PATCH] v4l: videobuf update
This updates the videobuf module used by the bttv, saa7134 and saa7146
drivers.
Changes:
* adds support for overlays (i.e. PCI-PCI DMA transfers).
* adds support for userspace DMA.
* fixes bug in the read() helper function (catch I/O errors).
Len Brown [Mon, 6 Oct 2003 12:48:05 +0000 (08:48 -0400)]
[ACPI] Summary of changes for ACPICA version 20031002:
Fixed a problem with Index Fields where the index was not incremented
for fields that require multiple writes to the index/data registers
(Fields that are wider than the data register.)
Fixed a problem with all Field objects where a write could go beyond the
end-of-field if the field was larger than the access granularity and
therefore required multiple writes to complete the request. An extra
write beyond the end of the field could happen inadvertently.
Fixed a problem with Index Fields where a BUFFER_OVERFLOW error would
incorrectly be returned if the width of the Data Register was larger
than the specified field access width.
Completed fixes for LoadTable() and Unload() and verified their
operation. Implemented full support for the "DdbHandle" object
throughout the ACPI CA subsystem.
Implemented full support for the MADT and ECDT tables in the ACPI CA
header files. Even though these tables are not directly consumed by
ACPI CA, the header definitions are useful for ACPI device drivers.
Integrated resource descriptor fixes posted to the Linux ACPI list.
This included checks for minimum descriptor length, and support for
trailing NULL strings within descriptors that have optional string
elements.
Fixed a problem where the SMI_CMD register could be written even if it
was not supported on the platform (when FADT.SMI_CMD is zero)
Andi Kleen [Mon, 6 Oct 2003 08:15:45 +0000 (01:15 -0700)]
[PATCH] cpufreq for x86-64
Reuses the i386 cpufreq code and add an notifier to the x86-64
timer code to track frequency changes.
All previous criticism of earlier patches should be addressed:
- it doesn't touch i386 now
- Doesn't include the ACPI P states driver anymore following brodo's request.
Andi Kleen [Mon, 6 Oct 2003 08:15:36 +0000 (01:15 -0700)]
[PATCH] UID16 fixes
This fixes CONFIG_UID16 problems on x86-64 as discussed earlier.
CONFIG_UID16 now only selects the inclusion of kernel/uid16.c, all
conversions are triggered dynamically based on type sizes. This allows
x86-64 to both include uid16.c for emulation purposes, but not truncate
uids to 16bit in sys_newstat.
- Replace the old macros from linux/highuid.h with new SET_UID/SET_GID
macros that do type checking. Based on Linus' proposal.
- Fix everybody to use them.
- Clean up some cruft in the x86-64 32bit emulation allowed by this
(other 32bit emulations could be cleaned too, but I'm too lazy for
that right now)
- Add one missing EOVERFLOW check in x86-64 32bit sys_newstat while
I was at it.
there's one call to strncpy() in vfat that really should be a memcpy().
The source, msdos_name, is a space-filled char array, not a
NUL-terminated string.
The rest of the patch eliminates duplicate occurencies of the
CURRENT_TIME macro inside functions. This shrinks vfat.o by a few
bytes.
Russell King [Mon, 6 Oct 2003 01:18:43 +0000 (18:18 -0700)]
[PATCH] Fix sysrq-t free stack output
It seems that we're attempting to work out the free stack size using two
unrelated pointers for the lowest address of the stack. Fix this to use
the correct pointer.