blktrace.git
18 years ago[PATCH] blktrace: note that the -b option is in KiB
Jens Axboe [Tue, 4 Oct 2005 17:54:09 +0000 (19:54 +0200)]
[PATCH] blktrace: note that the -b option is in KiB

18 years ago[PATCH] blktrace: optimize the get_event() normal case
Jens Axboe [Tue, 4 Oct 2005 09:13:58 +0000 (11:13 +0200)]
[PATCH] blktrace: optimize the get_event() normal case

We usually only see partial traces in very few numbers (like
a few in a million), so optimize the normal case by getting
a full trace instead of splitting it into a 4-byte and remainder
read.

18 years ago[PATCH] blkparse: only break out of last loop if not dumping to stdout
Jens Axboe [Mon, 3 Oct 2005 12:27:05 +0000 (14:27 +0200)]
[PATCH] blkparse: only break out of last loop if not dumping to stdout

Otherwise sending a ctrl-c when running btrace will quit, leaving
entries never dumped.

18 years ago[PATCH] blkparse: stdin sequence and aliases fixes
Jens Axboe [Mon, 3 Oct 2005 12:06:33 +0000 (14:06 +0200)]
[PATCH] blkparse: stdin sequence and aliases fixes

18 years ago[PATCH] blktrace: allow for unaligned events
Jens Axboe [Mon, 3 Oct 2005 12:06:07 +0000 (14:06 +0200)]
[PATCH] blktrace: allow for unaligned events

Unfortunately it looks like we have to support this for now,
as it appears relayfs will still miss events on buffer wraps
for the read cases.

18 years ago[PATCH verify_blkparse: dump number of sequence aliases as well
Jens Axboe [Mon, 3 Oct 2005 12:05:12 +0000 (14:05 +0200)]
[PATCH verify_blkparse: dump number of sequence aliases as well

18 years ago[PATCH] btrace: need large rb batch count for stdin sorting
Jens Axboe [Mon, 3 Oct 2005 12:04:48 +0000 (14:04 +0200)]
[PATCH] btrace: need large rb batch count for stdin sorting

18 years ago[PATCH] blktrace: various fixes
Jens Axboe [Mon, 3 Oct 2005 09:33:59 +0000 (11:33 +0200)]
[PATCH] blktrace: various fixes

- Better exit_trace() backoff
- Don't assume that write() always writes the full chunk
- kill mmap(), we will never use it for packet driven workloads
- Bug in get_data_read/read_data() for incremental reads
- Optimize tid_fd_lock() holding, drop before freeing memory

18 years ago[PATCH] blkparse: kill old force debug print
Jens Axboe [Mon, 3 Oct 2005 08:06:01 +0000 (10:06 +0200)]
[PATCH] blkparse: kill old force debug print

18 years ago[PATCH] kernel: update patch to 2.6.14-rc3
Jens Axboe [Mon, 3 Oct 2005 07:05:01 +0000 (09:05 +0200)]
[PATCH] kernel: update patch to 2.6.14-rc3

18 years ago[PATCH] blkparse: Add option to print missing entries
Jens Axboe [Sun, 2 Oct 2005 19:34:10 +0000 (21:34 +0200)]
[PATCH] blkparse: Add option to print missing entries

And default to off, it can be fairly verbose.

18 years ago[PATCH] blkparse: check smallest sequence read
Jens Axboe [Sun, 2 Oct 2005 13:39:14 +0000 (15:39 +0200)]
[PATCH] blkparse: check smallest sequence read

This allows check_sequence() to know when a skip happened and
when it is ok to proceed. Should be the final fix, now we now
when a sequence is missing and we nicely limit the memory
consumed in presence of large traces and missing entries.
This change brings down memory consumption of a skippy 1.3
million trace here down from 130MiB to 15MiB.

18 years ago[PATCH] blkparse: invert alloc checks, more intuitive
Jens Axboe [Fri, 30 Sep 2005 10:34:19 +0000 (12:34 +0200)]
[PATCH] blkparse: invert alloc checks, more intuitive

18 years ago[PATCH] blkparse: reduce memory consumption
Jens Axboe [Fri, 30 Sep 2005 09:52:24 +0000 (11:52 +0200)]
[PATCH] blkparse: reduce memory consumption

- Better skip detection

- Limit maximum number of t/bit allocation cache entries

18 years ago[PATCH] blkparse: fix ->sector printing warnings
Nathan Scott [Fri, 30 Sep 2005 05:33:30 +0000 (07:33 +0200)]
[PATCH] blkparse: fix ->sector printing warnings

18 years ago[PATCH] README: add reference to gitweb on brick
Jens Axboe [Thu, 29 Sep 2005 11:08:50 +0000 (13:08 +0200)]
[PATCH] README: add reference to gitweb on brick

18 years ago[PATCH] Assorted compiler warnings
Jens Axboe [Thu, 29 Sep 2005 10:01:37 +0000 (12:01 +0200)]
[PATCH] Assorted compiler warnings

18 years ago[PATCH] verify_blkparse: print events not ordered
Jens Axboe [Wed, 28 Sep 2005 15:57:03 +0000 (17:57 +0200)]
[PATCH] verify_blkparse: print events not ordered

And fix exit code.

18 years ago[PATCH] Add verify_blkparse description to README
Jens Axboe [Wed, 28 Sep 2005 15:50:56 +0000 (17:50 +0200)]
[PATCH] Add verify_blkparse description to README

18 years ago[PATCH] verify_blkparse: implement in C instead
Jens Axboe [Wed, 28 Sep 2005 15:45:51 +0000 (17:45 +0200)]
[PATCH] verify_blkparse: implement in C instead

The C version checks a 92MB file in 2 seconds where the php
version takes 23 seconds.

18 years ago[PATCH] verify_blkparse: missing line break
Jens Axboe [Wed, 28 Sep 2005 14:27:52 +0000 (16:27 +0200)]
[PATCH] verify_blkparse: missing line break

18 years ago[PATCH] verify_blkparse: fix parsing
Thomas Kenne [Wed, 28 Sep 2005 13:51:16 +0000 (15:51 +0200)]
[PATCH] verify_blkparse: fix parsing

Current version would break, if the device number didn't have
a leading space.

18 years ago[PATCH] blkparse: prune old found entries in rb_last
Jens Axboe [Wed, 28 Sep 2005 13:41:22 +0000 (15:41 +0200)]
[PATCH] blkparse: prune old found entries in rb_last

18 years ago[PATCH] verify_blkparse: make sure file is readable
Jens Axboe [Wed, 28 Sep 2005 13:40:16 +0000 (15:40 +0200)]
[PATCH] verify_blkparse: make sure file is readable

18 years ago[PATCH] verify_blkparse: proper check of file given
Jens Axboe [Wed, 28 Sep 2005 12:52:06 +0000 (14:52 +0200)]
[PATCH] verify_blkparse: proper check of file given

18 years ago[PATCH] blkparse: don't check time in last rb lookup
Jens Axboe [Wed, 28 Sep 2005 12:17:39 +0000 (14:17 +0200)]
[PATCH] blkparse: don't check time in last rb lookup

18 years ago[PATCH] Add script to verify correctness of blkparse output
Thomas Kenne [Wed, 28 Sep 2005 12:12:55 +0000 (14:12 +0200)]
[PATCH] Add script to verify correctness of blkparse output

Only checks the time.

18 years ago[PATCH] blkparse: make the rb track root per-device
Jens Axboe [Wed, 28 Sep 2005 09:34:29 +0000 (11:34 +0200)]
[PATCH] blkparse: make the rb track root per-device

18 years ago[PATCH] blkparse: misc assorted fixes
Jens Axboe [Wed, 28 Sep 2005 08:56:59 +0000 (10:56 +0200)]
[PATCH] blkparse: misc assorted fixes

- Move the trace verification earlier, so that traces on
  trace_list are always known valid

- Start with -1 sequence so when using the stopwatch the
  skip count is correct (first entry might not be 1)

- stdin get_cpu_info() fix

18 years ago[PATCH] kernel: move the plug/unplug logging around a little
Jens Axboe [Wed, 28 Sep 2005 06:55:52 +0000 (08:55 +0200)]
[PATCH] kernel: move the plug/unplug logging around a little

18 years ago[PATCH] blkparse: sequence fixes
Jens Axboe [Tue, 27 Sep 2005 22:03:59 +0000 (00:03 +0200)]
[PATCH] blkparse: sequence fixes

This should finally fix all the remaining sequence bugs. We
keep a backlog of already processed entries so we can lookup
any out-of-sequence events there. That takes care of one
direction of sequence and time mismatches. The other direction
is handled by fixing the trace_rb_find() lookup to work well
enough without having a precise time (which is the primary
key for the rb sort).

18 years ago[PATCH] blkparse: improve skip detection
Jens Axboe [Tue, 27 Sep 2005 16:18:04 +0000 (18:18 +0200)]
[PATCH] blkparse: improve skip detection

18 years agoForgot to kill old patch
Jens Axboe [Tue, 27 Sep 2005 13:38:08 +0000 (15:38 +0200)]
Forgot to kill old patch

18 years ago[PATCH] kernel: improve serialization and unplug logging
Jens Axboe [Tue, 27 Sep 2005 10:18:14 +0000 (12:18 +0200)]
[PATCH] kernel: improve serialization and unplug logging

18 years agoForgot to include the change for the last stopwatch fix...
Jens Axboe [Tue, 27 Sep 2005 09:04:43 +0000 (11:04 +0200)]
Forgot to include the change for the last stopwatch fix...

18 years ago[PATCH] blkparse: another stab at stopwatch_end fixing
Jens Axboe [Tue, 27 Sep 2005 08:22:33 +0000 (10:22 +0200)]
[PATCH] blkparse: another stab at stopwatch_end fixing

This should work reliably, the check is moved after we
have definitely called find_genesis().

18 years agoRevert "[PATCH] blkparse: final stopwatch check"
Jens Axboe [Tue, 27 Sep 2005 08:15:24 +0000 (10:15 +0200)]
Revert "[PATCH] blkparse: final stopwatch check"

This reverts a795e5fae97e14d77c3c3c7d0f6543e6c5f8c850 commit.

18 years ago[PATCH] blktrace: fix sparse warning
Jens Axboe [Tue, 27 Sep 2005 08:15:17 +0000 (10:15 +0200)]
[PATCH] blktrace: fix sparse warning

18 years ago[PATCH] blkparse: add %N format for bytes
Jens Axboe [Mon, 26 Sep 2005 17:05:40 +0000 (19:05 +0200)]
[PATCH] blkparse: add %N format for bytes

We need this for PC requests, they are not multiples of 512
like the FS requests. At the same time change the default for
PC requests to dump in bytes.

18 years ago[PATCH] blkparse: final stopwatch check
Jens Axboe [Mon, 26 Sep 2005 13:42:05 +0000 (15:42 +0200)]
[PATCH] blkparse: final stopwatch check

Delayed quit when all files have been read and there are no
further entries beyond that point, then it is safe to quit
out early there as well.

18 years ago[PATCH] blkparse: more stopwatch cleanups
Jens Axboe [Mon, 26 Sep 2005 13:36:19 +0000 (15:36 +0200)]
[PATCH] blkparse: more stopwatch cleanups

18 years ago[PATCH] blkparse: improve stopwatch
Jens Axboe [Mon, 26 Sep 2005 13:34:37 +0000 (15:34 +0200)]
[PATCH] blkparse: improve stopwatch

Move the checks much earlier, no point in doing a lot of
work for stuff we will toss out anyways.

18 years ago[PATCH] blkparse: hash-by-name fixes
Jens Axboe [Mon, 26 Sep 2005 12:38:23 +0000 (14:38 +0200)]
[PATCH] blkparse: hash-by-name fixes

- Be more precise in dumping (name, ...) in the stats, only do it
  when there really was more than one.

- The -n option needs to clear the hash-by-pid variable.

18 years ago[PATCH] blkparse: oops, missed removing a hash debug statement
Jens Axboe [Mon, 26 Sep 2005 12:32:59 +0000 (14:32 +0200)]
[PATCH] blkparse: oops, missed removing a hash debug statement

18 years ago[PATCH] TODO: add entry about fixing show_entries_rb() OOO logic
Jens Axboe [Mon, 26 Sep 2005 12:30:12 +0000 (14:30 +0200)]
[PATCH] TODO: add entry about fixing show_entries_rb() OOO logic

18 years ago[PATCH] blkparse: Add option to hash process by name
Jens Axboe [Mon, 26 Sep 2005 12:26:24 +0000 (14:26 +0200)]
[PATCH] blkparse: Add option to hash process by name

For some traces, you really don't want to see thousands of cc1
processes in the stats, you want to collect the stats by name
instead. This adds the -n option to blkparse, enabling that.

Switched to Jenkins hash at the same time, as it allows for
easy hash-by-u32 or hash-by-name as we see fit.

18 years ago[PATCH] blkparse: more sequence fixes
Jens Axboe [Mon, 26 Sep 2005 11:12:33 +0000 (13:12 +0200)]
[PATCH] blkparse: more sequence fixes

18 years ago[PATCH] blkparse: improve skipping and genesis_time lookup
Jens Axboe [Mon, 26 Sep 2005 10:26:29 +0000 (12:26 +0200)]
[PATCH] blkparse: improve skipping and genesis_time lookup

18 years ago[PATCH] blkparse: hardcode the default formats instead of parsing them
Jens Axboe [Mon, 26 Sep 2005 09:51:58 +0000 (11:51 +0200)]
[PATCH] blkparse: hardcode the default formats instead of parsing them

It takes a lot of time to manually parse each individual trace,
and there's really no point in doing that work if the user did
not specificy a custom format. We now only parse the given custom
formats.

Without any custom formats (eg the default), parsing a trace of
1.3 million entries now takes 4.3 seconds where it took 11.25 before.

18 years ago[PATCH] blkparse: force dump of remaining rb entries on reading done
Jens Axboe [Mon, 26 Sep 2005 08:50:04 +0000 (10:50 +0200)]
[PATCH] blkparse: force dump of remaining rb entries on reading done

Otherwise we may not display a significant chunk of entries.

18 years ago[PATCH] kernel: put back old syncronization
Jens Axboe [Sun, 25 Sep 2005 12:31:24 +0000 (14:31 +0200)]
[PATCH] kernel: put back old syncronization

It's actually the best - keep the trace on the stack, so multiple
invocations from the same CPU (ie from process and irq context at
the same time) get their own 't', and keep the local irq disable
to prevent interleaving of trace and PDU with other contexts.

18 years ago[PATCH] blkparse: split format parsing into seperate file
Jens Axboe [Sun, 25 Sep 2005 05:37:11 +0000 (07:37 +0200)]
[PATCH] blkparse: split format parsing into seperate file

18 years agoRevert "[PATCH] kernel: kill irq save/restore"
Jens Axboe [Sat, 24 Sep 2005 19:54:44 +0000 (21:54 +0200)]
Revert "[PATCH] kernel: kill irq save/restore"

This reverts 68bf440bf36d239ff3cb27b34b95ea7911c7c5a6 commit.

18 years ago[PATCH] kernel: kill irq save/restore
Jens Axboe [Fri, 23 Sep 2005 15:24:04 +0000 (17:24 +0200)]
[PATCH] kernel: kill irq save/restore

It should not be needed anymore, the writes a per-CPU and the
get_cpu()/put_cpu() will prevent preemption.

18 years ago[PATCH] kernel: avoid trace corruption on SMP
Jens Axboe [Fri, 23 Sep 2005 15:21:11 +0000 (17:21 +0200)]
[PATCH] kernel: avoid trace corruption on SMP

Maintain a static trace structure per-CPU.

18 years ago[PATCH] blkparse: sequence fixes
Jens Axboe [Fri, 23 Sep 2005 14:54:56 +0000 (16:54 +0200)]
[PATCH] blkparse: sequence fixes

Only set last sequence _after_ a potential break of the loop.
And remember to set last_allowed_time for stdin reads as well.

18 years ago[PATCH] blkparse: fix sorting of file events regardless of batch size
Jens Axboe [Fri, 23 Sep 2005 14:45:51 +0000 (16:45 +0200)]
[PATCH] blkparse: fix sorting of file events regardless of batch size

If the batch wraps, it would be possible to start displaying events
way into the future if parsing multiple devices.

18 years agoRevert "[PATCH] blkparse: rbtree sort primary key should be sequence, not time"
Jens Axboe [Fri, 23 Sep 2005 13:09:37 +0000 (15:09 +0200)]
Revert "[PATCH] blkparse: rbtree sort primary key should be sequence, not time"

This reverts acba447a946b022df6a30159425309f033e46ab7 commit.

18 years ago[PATCH] blktrace: Add option for setting sub buffer size
Jens Axboe [Fri, 23 Sep 2005 09:46:40 +0000 (11:46 +0200)]
[PATCH] blktrace: Add option for setting sub buffer size

Use -b to give sub buffer size in KiB, -n to give number of
sub buffers wanted. This affects the relayfs trace setup.

18 years ago[PATCH] blkparse: rbtree sort primary key should be sequence, not time
Jens Axboe [Fri, 23 Sep 2005 09:39:48 +0000 (11:39 +0200)]
[PATCH] blkparse: rbtree sort primary key should be sequence, not time

Otherwise we may lose events if the event time is a little out
of sync.

18 years ago[PATCH] blkparse: can't return on-stack variables
Jens Axboe [Fri, 23 Sep 2005 09:12:48 +0000 (11:12 +0200)]
[PATCH] blkparse: can't return on-stack variables

Just make it static, access is serialized to it.

18 years ago[PATCH] README: add link to mailing list at vger
Jens Axboe [Fri, 23 Sep 2005 08:07:54 +0000 (10:07 +0200)]
[PATCH] README: add link to mailing list at vger

18 years ago[PATCH] blkparse: don't append .log to the -o given output file
Jens Axboe [Fri, 23 Sep 2005 06:49:22 +0000 (08:49 +0200)]
[PATCH] blkparse: don't append .log to the -o given output file

18 years ago[PATCH] blkparse: make the nfiles counter live inside pdi
Nathan Scott [Fri, 23 Sep 2005 06:34:23 +0000 (08:34 +0200)]
[PATCH] blkparse: make the nfiles counter live inside pdi

Gets rid of the extra malloc and is cleaner.

18 years ago[PATCH] blkparse: fix pointer cast warnings
Nathan Scott [Fri, 23 Sep 2005 06:33:20 +0000 (08:33 +0200)]
[PATCH] blkparse: fix pointer cast warnings

18 years ago[PATCH] blkparse: correctly account number of files per device
Alan D. Brunelle [Thu, 22 Sep 2005 18:06:53 +0000 (20:06 +0200)]
[PATCH] blkparse: correctly account number of files per device

18 years agoMerge refs/heads/master from origin
Jens Axboe [Thu, 22 Sep 2005 18:02:41 +0000 (20:02 +0200)]
Merge refs/heads/master from origin

18 years ago[PATCH] blkparse: more optimizations
Jens Axboe [Thu, 22 Sep 2005 18:02:30 +0000 (20:02 +0200)]
[PATCH] blkparse: more optimizations

The sort path unification and memory cutback introduced some
overhead, mainly due to lots of memory allocations and frees
of the same structures over and over again. Add a frontend
cache for struct trace and blk_io_trace allocations.

Also concatenate a lot of the string operations in fmt_select,
it's quite high in the profiles.

This brings us to the same performance as the memory hog version.

18 years ago[PATCH] kernel: export the __blk_add_trace() symbol for modules
Jens Axboe [Thu, 22 Sep 2005 14:47:40 +0000 (16:47 +0200)]
[PATCH] kernel: export the __blk_add_trace() symbol for modules

18 years agoMerge refs/heads/master from origin
Jens Axboe [Thu, 22 Sep 2005 14:43:52 +0000 (16:43 +0200)]
Merge refs/heads/master from origin

18 years ago[PATCH] blkparse: drastically reduce memory foot print
Jens Axboe [Thu, 22 Sep 2005 14:43:48 +0000 (16:43 +0200)]
[PATCH] blkparse: drastically reduce memory foot print

File reads needed enough memory to hold all events from all CPUs.
This quickly grew to hundreds of megabytes. Bring down this number
by interleaving reads from the files and relying on the same
mechanism as the piped reads to syncronize them. This makes file reads
use the same path as pipes and actually removes more code than it adds.
For a 900,000 entry file, memory consumption here is down from ~110MiB
from about 1MiB to process it.

18 years agoRevert "[PATCH] blkparse: bad ncpus check"
Jens Axboe [Thu, 22 Sep 2005 14:40:44 +0000 (16:40 +0200)]
Revert "[PATCH] blkparse: bad ncpus check"

This reverts 08d83c704fbda722d3040d643665e2c029744aa7 commit.

18 years ago[PATCH] blkparse: bad ncpus check
Jens Axboe [Thu, 22 Sep 2005 14:15:45 +0000 (16:15 +0200)]
[PATCH] blkparse: bad ncpus check

18 years ago[PATCH] blkparse: shave 20% off the runtime
Jens Axboe [Thu, 22 Sep 2005 12:29:50 +0000 (14:29 +0200)]
[PATCH] blkparse: shave 20% off the runtime

Various optimizations and moving code around a little.

18 years ago[PATCH] Add local master repo to documentation
Jens Axboe [Thu, 22 Sep 2005 11:33:55 +0000 (13:33 +0200)]
[PATCH] Add local master repo to documentation

Note that it is using the git daemon, not rsync.

18 years ago[PATCH] Update in-program help/usage info
Jens Axboe [Thu, 22 Sep 2005 10:30:12 +0000 (12:30 +0200)]
[PATCH] Update in-program help/usage info

18 years ago[PATCH] blkparse: Fix stopwatch
Jens Axboe [Thu, 22 Sep 2005 10:11:08 +0000 (12:11 +0200)]
[PATCH] blkparse: Fix stopwatch

The option should be start:end, not start:duration (at least that is
what I intended and what is documented). Also fix the actual dumping,
we need to prune earlier entries from the rb tree or we will loop
forever.

18 years ago[PATCH] blkparse: more PC fixes
Jens Axboe [Thu, 22 Sep 2005 08:56:05 +0000 (10:56 +0200)]
[PATCH] blkparse: more PC fixes

Add size of request as well, kill space at the end of pdu dump.

18 years ago[PATCH] blkparse: fix some pc command dumping
Jens Axboe [Thu, 22 Sep 2005 08:46:06 +0000 (10:46 +0200)]
[PATCH] blkparse: fix some pc command dumping

18 years ago[PATCH] Add support for bounce and split operations
Jens Axboe [Thu, 22 Sep 2005 08:29:10 +0000 (10:29 +0200)]
[PATCH] Add support for bounce and split operations

18 years ago[PATCH] README: typo
Jens Axboe [Wed, 21 Sep 2005 16:45:29 +0000 (18:45 +0200)]
[PATCH] README: typo

18 years ago[PATCH] blkparse: just skip unknown events, there are not fatal
Jens Axboe [Wed, 21 Sep 2005 16:38:38 +0000 (18:38 +0200)]
[PATCH] blkparse: just skip unknown events, there are not fatal

18 years agoMerge branch 'master' of ssh://axboe@master.kernel.org/pub/scm/linux/kernel/git/axboe...
Jens Axboe [Wed, 21 Sep 2005 15:36:01 +0000 (17:36 +0200)]
Merge branch 'master' of ssh://axboe@/linux/kernel/git/axboe/blktrace

18 years ago[PATCH] blkparse: split queue and insert into two operations
Jens Axboe [Wed, 21 Sep 2005 15:35:30 +0000 (17:35 +0200)]
[PATCH] blkparse: split queue and insert into two operations

We need this to support md/dm actions.

18 years ago[PATCH] Fix output naming, really
Jens Axboe [Wed, 21 Sep 2005 11:59:23 +0000 (13:59 +0200)]
[PATCH] Fix output naming, really

It's still confusing since the manual name wont include _out like
the automatically generated one will. So now a trace of /dev/sda
will generate sda.blktrace.cpu and blkparse -i sda will catch
those files. Similarly, a blktrace -o sda will generate the
same filename.

18 years ago[PATCH] blkparse: fix the alphasorting
Jens Axboe [Wed, 21 Sep 2005 11:52:52 +0000 (13:52 +0200)]
[PATCH] blkparse: fix the alphasorting

Bad qsort size passed in and missing increment in ppis assignment.
Verified that it actually works now :-)

18 years ago[PATCH] blkparse: missing free of ppis
Jens Axboe [Wed, 21 Sep 2005 09:57:25 +0000 (11:57 +0200)]
[PATCH] blkparse: missing free of ppis

Doesn't really matter since we exit after this, but...

18 years ago[PATCH] blkparse: include pid in stat output
Jens Axboe [Wed, 21 Sep 2005 09:46:51 +0000 (11:46 +0200)]
[PATCH] blkparse: include pid in stat output

Avoids name conflict with threads.

18 years ago[PATCH] blkparse: sort the processes by name in stat output
Jens Axboe [Wed, 21 Sep 2005 09:45:39 +0000 (11:45 +0200)]
[PATCH] blkparse: sort the processes by name in stat output

18 years ago[PATCH] Change output/input file expectations
Jens Axboe [Tue, 20 Sep 2005 12:41:37 +0000 (14:41 +0200)]
[PATCH] Change output/input file expectations

I think it's a little confusing right now with the naming, giving
-o foo to blktrace results in foo_dev_out.cpu files. Change this
to instead generate foo.cpu files, if the option is given.

18 years ago[PATCH] blktrace: fix exit
Jens Axboe [Tue, 20 Sep 2005 07:43:49 +0000 (09:43 +0200)]
[PATCH] blktrace: fix exit

get_data() workers should return failure on is_done(), otherwise
extract() will attempt to verify a bad trace.

18 years ago[PATCH] blkparse: Add format specifiers and format field identifiers
Alan D. Brunelle [Tue, 20 Sep 2005 07:34:45 +0000 (09:34 +0200)]
[PATCH] blkparse: Add format specifiers and format field identifiers

This allows the user to completely customize the output format.

18 years ago[PATCH] blktrace: prepare for mmap usage again
Jens Axboe [Tue, 20 Sep 2005 07:13:34 +0000 (09:13 +0200)]
[PATCH] blktrace: prepare for mmap usage again

It isn't enabled right now, we need a proper consumed/produced
notification model to signal availability of data. poll() doesn't
work well enough, since it only works on a sub-buffer granularity.

18 years ago[PATCH] blkparse: account skips for file reads as well
Jens Axboe [Tue, 20 Sep 2005 07:12:34 +0000 (09:12 +0200)]
[PATCH] blkparse: account skips for file reads as well

18 years ago[PATCH] blkparse: allow a few more skips
Jens Axboe [Mon, 19 Sep 2005 11:59:30 +0000 (13:59 +0200)]
[PATCH] blkparse: allow a few more skips

18 years ago[PATCH] btrace: need space between options
Jens Axboe [Mon, 19 Sep 2005 11:46:19 +0000 (13:46 +0200)]
[PATCH] btrace: need space between options

18 years ago[PATCH] blkparse: more file breakage
Jens Axboe [Mon, 19 Sep 2005 11:34:56 +0000 (13:34 +0200)]
[PATCH] blkparse: more file breakage

If we have skipped entries in a file, don't wait around for more entries.

18 years ago[PATCH] blkparse: actually include blkparse diff
Jens Axboe [Mon, 19 Sep 2005 11:07:29 +0000 (13:07 +0200)]
[PATCH] blkparse: actually include blkparse diff

Forgot to update the cache for the file breakage patch

18 years ago[PATCH] blkparse: Fixup breakage in file reads
Jens Axboe [Mon, 19 Sep 2005 10:39:22 +0000 (12:39 +0200)]
[PATCH] blkparse: Fixup breakage in file reads

The pipe fixed broke file reads:

- Don't kill individual entries in show_entries_rb(),
  since it's a big chunk on file reads

- Need to "lookup" pdi, in case it was resized.

18 years ago[PATCH] blkparse: reinstate __find_track() failures
Jens Axboe [Mon, 19 Sep 2005 10:25:14 +0000 (12:25 +0200)]
[PATCH] blkparse: reinstate __find_track() failures

It's informational and can help find bugs.