The patchset contains mainly scalability and NUMA stuff, and anything
else that stops things from irritating me. It's meant to be pretty stable,
not so much a testing ground for new stuff.
I'd be very interested in feedback from anyone willing to test on any
platform, however large or small.
ftp://ftp.kernel.org/pub/linux/kernel/people/mbligh/2.5.73/patch-2.5.73-mjb2.bz2
additional patches that can be applied if desired:
(these three form the qlogic feral driver)
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.72/2.5.72-mm1/broken-out/linux-isp.patch
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.72/2.5.72-mm1/broken-out/isp-update-1.patch
ftp://ftp.kernel.org/pub/linux/kernel/people/akpm/patches/2.5/2.5.72/2.5.72-mm1/broken-out/isp-remove-pci_detect.patch
Since 2.5.73-mjb1 (~ = changed, + = added, - = dropped)
Notes:
Small release, but upside_down may have large impact.
On the other hand, it's very cool, and solves various awkward problems.
See if you can break it ;-)
Now in Linus' tree:
New:
+ upside_down William Lee Irwin
Allocate memory from the top down, shove the stack where it belongs.
+ node_spanned_pages Dave Hansen
Fix up NUMA beancounting
Dropped:
- numameminfo fix Martin J. Bligh
(merged with node_spanned_pages).
Pending:
Hyperthreaded scheduler (Ingo Molnar)
scheduler callers profiling (Anton or Bill Hartner)
Child runs first (akpm)
Kexec
e1000 fixes
Update the lost timer ticks code
pidmaps_nodepages (Dave Hansen)
Present in this patch:
early_printk Dave Hansen / Keith Mannthey
Allow printk before console_init
confighz Andrew Morton / Dave Hansen
Make HZ a config option of 100 Hz or 1000 Hz
config_page_offset Dave Hansen / Andrea
Make PAGE_OFFSET a config option
numameminfo Martin Bligh / Keith Mannthey
Expose NUMA meminfo information under /proc/meminfo.numa
schedstat Rick Lindsley
Provide stats about the scheduler under /proc/schedstat
schedstat2 Rick Lindsley
Provide more stats about the scheduler under /proc/schedstat
schedstat-scripts Rick Lindsley
Provide some scripts for schedstat analysis under scripts/
sched_tunables Robert Love
Provide tunable parameters for the scheduler (+ NUMA scheduler)
irq_affinity Martin J. Bligh
Workaround for irq_affinity on clustered apic mode systems (eg x440)
partial_objrmap Dave McCracken
Object based rmap for filebacked pages.
kgdb Andrew Morton
The older version of kgdb, synched with 2.5.54-mm1
thread_info_cleanup (4K stacks pt 1) Dave Hansen / Ben LaHaise
Prep work to reduce kernel stacks to 4K
interrupt_stacks (4K stacks pt 2) Dave Hansen / Ben LaHaise
Create a per-cpu interrupt stack.
stack_usage_check (4K stacks pt 3) Dave Hansen / Ben LaHaise
Check for kernel stack overflows.
4k_stack (4K stacks pt 4) Dave Hansen
Config option to reduce kernel stacks to 4K
4k_stacks_vs_kgdb Dave Hansen
Fix interaction between kgdb and 4K stacks
stacks_from_slab William Lee Irwin
Take kernel stacks from the slab cache, not page allocation.
thread_under_page William Lee Irwin
Fix THREAD_SIZE < PAGE_SIZE case
spinlock_inlining Andrew Morton & Martin J. Bligh
Inline spinlocks for profiling. Made into a ugly config option by me.
lockmeter John Hawkes / Hanna Linder
Locking stats.
reiserfs_dio Mingming Cao
DIO for Reiserfs
sched_interactive Ingo Molnar
Bugfix for interactive scheduler
kgdb_cleanup Martin J. Bligh
Stop kgdb renaming schedule to do_schedule when it's not even enabled
acenic_fix Martin J. Bligh
Fix warning in acenic driver
local_balance_exec Martin J. Bligh
Modify balance_exec to use node-local queues when idle
tcp_speedup Martin J. Bligh
Speedup TCP (avoid double copy) as suggested by Linus
disable preempt Martin J. Bligh
I broke preempt somehow, temporarily disable it to stop accidents
ppc64 fixes Anton Blanchard
Various PPC64 fixes / updates
config_debug Dave Hansen
Make '-g' for the kernel a config option
akpm_bear_pit Andrew Morton
Add a printk for some buffer error I was hitting
32bit_dev_t Andries Brouwer
Make dev_t 32 bit
dynamic_hd_struct Badari Pulavarty
Allocate hd_structs dynamically
lotsa_sds Badari Pulavarty
Create some insane number of sds
iosched_hashes Badari Pulavarty
Twiddle with the iosched hash tables for fun & profit
per_node_idt Zwane Mwaikambo
Per node IDT so we can do silly numbers of IO-APICs on NUMA-Q
config_numasched Dave Hansen
Turn NUMA scheduler into a config option
lockmeter_tytso Ted Tso
Fix lockmeter
aiofix2 Mingming Cao
fixed a bug in ioctx_alloc()
config_irqbal Keith Mannthey
Make irqbalance a config option
fs_aio_1_retry Suparna Bhattacharya
Filesystem aio. Chapter 1
fs_aio_2_read Suparna Bhattacharya
Filesystem aio. Chapter 2
fs_aio_3_write Suparna Bhattacharya
Filesystem aio. Chapter 3
fs_aio_4_down_wq Suparna Bhattacharya
Filesystem aio. Chapter 4
fs_aio_5_wrdown_wq Suparna Bhattacharya
Filesystem aio. Chapter 5
fs_aio_6_bread_wq Suparna Bhattacharya
Filesystem aio. Chapter 6
fs_aio_7_ext2getblk_wq Suparna Bhattacharya
Filesystem aio. Chapter 7
fs_aio_8_down_wq-ppc64 Suparna Bhattacharya
Filesystem aio. Chapter 8
fs_aio_9_down_wq-x86_64 Suparna Bhattacharya
Filesystem aio. Chapter 9
reslabify-pmd-pgd William Lee Irwin
Stick things back in the slab. Or something.
separate_pmd Dave Hansen
Separate kernel pmd per task.
banana_split Dave Hansen
Make PAGE_OFFSET play twister and limbo.
percpu_real_loadavg Dave Hansen / Martin J. Bligh
Tell me what the real load average is, and tell me per cpu.
nolock Dave McCracken
Nah, we don't like locks.
proc_pid_readdir Manfred Spraul
Make proc_pid_readdir more efficent. Allegedly.
mbind_part1 Matt Dobson
Bind some memory for NUMA.
mbind_part2 Matt Dobson
Bind some more memory for NUMA.
per_node_rss Matt Dobson
Track which nodes tasks mem is on, so sched can be sensible.
uninitialised timer Mikael Pettersson
Fix bogus cleanup so that gcc 2.95.4 works.
sysrq_t_fix Andrew Morton
Fix sysrq+t breakage where it showed the same stack for everyone
swsusp_state_check Matt Dobson
Fix a check in s/w suspend code
pci_hotplug Ivan K.
Fix up pci hotplug w/o hotplug enabled.
pfn_to_nid Martin J. Bligh
Dance around the twisted rats nest of crap in i386 include.
upside_down William Lee Irwin
Allocate memory from the top down, shove the stack where it belongs.
node_spanned_pages Dave Hansen
Fix up NUMA beancounting
-mjb Martin J. Bligh
Add a tag to the makefile
Martin J. Bligh said:
> The patchset contains mainly scalability and NUMA stuff, and anything
> else that stops things from irritating me. It's meant to be pretty
> stable, not so much a testing ground for new stuff.
>
> I'd be very interested in feedback from anyone willing to test on any
> platform, however large or small.
>
> ftp://ftp.kernel.org/pub/linux/kernel/people/mbligh/2.5.73/patch-2.5.73-mjb2.bz2
>
Are you interested in behaviour of this kernel on uniprocessor machines ?
I tested 2.5.72-mjb2 but it was full of oopses and crashes on my Duron
so I thought this patch is only for NUMA stuff.
Bye
Calin
--
# fortune
fortune: write error on /dev/null --- please empty the bit bucket
-----------------------------------------
This email was sent using SquirrelMail.
"Webmail for nuts!"
http://squirrelmail.org/
> Martin J. Bligh said:
>> The patchset contains mainly scalability and NUMA stuff, and anything
>> else that stops things from irritating me. It's meant to be pretty
>> stable, not so much a testing ground for new stuff.
>>
>> I'd be very interested in feedback from anyone willing to test on any
>> platform, however large or small.
>>
>> ftp://ftp.kernel.org/pub/linux/kernel/people/mbligh/2.5.73/patch-2.5.73-mjb2.bz2
>>
>
> Are you interested in behaviour of this kernel on uniprocessor machines ?
Yup.
> I tested 2.5.72-mjb2 but it was full of oopses and crashes on my Duron
> so I thought this patch is only for NUMA stuff.
Nope, it should work with any machine - you got the oopses?
If you have an old distro with glibc < 2.3.1, Bill thinks the upside_down
trick doesn't work because of some invalid assumptions glibc is making.
If that's the case, could you check that 2.5.73-mjb1 works OK?
Thanks,
M.
At some point in the past, Szonyi Calin <[email protected]> wrote:
>> I tested 2.5.72-mjb2 but it was full of oopses and crashes on my Duron
>> so I thought this patch is only for NUMA stuff.
On Sat, Jun 28, 2003 at 07:10:26AM -0700, Martin J. Bligh wrote:
> Nope, it should work with any machine - you got the oopses?
> If you have an old distro with glibc < 2.3.1, Bill thinks the upside_down
> trick doesn't work because of some invalid assumptions glibc is making.
> If that's the case, could you check that 2.5.73-mjb1 works OK?
If this is causing too much confusion and/or other anguish I can live
with it getting withdrawn and keep it rolling in the ultra-experimental
section (-wli).
Alternatively, it should be trivial to convert to a config option that's
off by default.
-- wli
On Sat, 28 Jun 2003, William Lee Irwin III wrote:
> At some point in the past, Szonyi Calin <[email protected]> wrote:
> >> I tested 2.5.72-mjb2 but it was full of oopses and crashes on my Duron
> >> so I thought this patch is only for NUMA stuff.
>
> On Sat, Jun 28, 2003 at 07:10:26AM -0700, Martin J. Bligh wrote:
> > Nope, it should work with any machine - you got the oopses?
> > If you have an old distro with glibc < 2.3.1, Bill thinks the upside_down
> > trick doesn't work because of some invalid assumptions glibc is making.
> > If that's the case, could you check that 2.5.73-mjb1 works OK?
>
> If this is causing too much confusion and/or other anguish I can live
> with it getting withdrawn and keep it rolling in the ultra-experimental
> section (-wli).
>
> Alternatively, it should be trivial to convert to a config option that's
> off by default.
Haven't had a chance to try this yet, so I don't have a feel for the
benefit (other than good karma). But having as an option is good,
depending on how out of date the lib has to be to have troubles should
determine default, this is not the stock kernel and can be defaulted to
enable new features IMHO.
The reason I haven't tried it is because I'm trying to find time to shake
out the 73-wli kernel, which will find a home on my slower boxen.
--
bill davidsen <[email protected]>
CTO, TMR Associates, Inc
Doing interesting things with little computers since 1979.