From: Roland Schober <rolandschober1968@gmail.com>
To: linux-bluetooth@vger.kernel.org
Cc: Roland Schober <rolandschober1968@gmail.com>
Subject: [PATCH BlueZ v4 00/10] monitor: Decode Periodic Advertising with Responses
Date: Fri, 20 Nov 2026 07:38:46 +0100 [thread overview]
Message-ID: <20261120063856.16585-1-rolandschober1968@gmail.com> (raw)
This series teaches btmon the HCI commands and events of Periodic
Advertising with Responses (PAwR, Bluetooth 5.4). Until now they were
only listed by name in the opcode table and the events were printed as
unknown.
It is meant as the first step towards supporting the PAwR advertiser
role in the kernel and in BlueZ, which is what an access point of the
Electronic Shelf Label profile needs: being able to read the traces
comes before changing anything else.
Patches 1 to 7 fix existing decoders which the new ones use or which
were found on the way:
- the Connection_Handle returned by the Periodic Advertising Sync
Transfer commands and by LE Remove ISO Data Path was read as one
octet,
- LE Periodic Advertising Report trusted Data_Length and
LE Extended Create Connection trusted the number of PHY parameter
blocks, so truncated packets made them decode octets outside the
packet,
- the LE Set Periodic Advertising commands printed the advertising
handle as a connection handle,
- LE Periodic Advertising Report printed an unavailable TX power as
127 dBm and named the AoD CTE type with 1 us slots AoA,
- LE PAST Received printed the handle twice and decoded the clock
accuracy as a BR/EDR one,
- the Sync_Handle of two events and the properties of LE Set Periodic
Advertising Parameters were printed without byte order conversion.
Patch 8 splits the LE Periodic Advertising Report decoding into
helpers without changing its output, so that patch 10 can share it.
Patch 9 decodes the five commands and patch 10 the six events,
including the [v2] variants of existing events.
Testing: I have no controller with PAwR support at hand yet. The
commands were sent with hcitool to a controller which rejects them, so
that btmon decodes what the host sent, and the events and the
successful command responses were decoded from hand-made btsnoop
traces. Every command and event of these traces was also fed to the
decoders cut to every length, in a buffer of exactly that size, with
AddressSanitizer, which found no read beyond a packet. The output for
well formed packets of the existing decoders was compared with the
current tree. I have no big-endian host either: the byte order was
checked with a btmon whose little-endian conversions swap, on traces
with the fields of more than one octet reversed, which has to print
the same as the normal btmon on the normal traces.
v4:
- Add patch 7, and convert the Sync_Handle and the properties in the
new decoders as well. Copilot found two of these places, the test
described above two more.
- The fix of the LE Extended Create Connection PHY decoding no longer
adds a size to a void pointer, which CodeQL reported.
v3:
- Add the fixes of the advertising handle, of the TX power and CTE
type and of LE PAST Received, and print these fields the same way
in the new decoders.
- LE Periodic Advertising Report fix: also report octets beyond
Data_Length.
- LE Extended Create Connection fix: report extra PHY parameter
blocks, and do not report the remaining PHYs as unknown after a
truncated block.
- Commands: reuse the [v1] decoder for LE Set Periodic Advertising
Parameters [v2], print 0x00 subevent timings as none, check
Response_Data_Length, and match LE Extended Create Connection [v2]
with the events which complete it.
- Events: Data_Status 0xff keeps its reserved meaning in the [v1]
LE Periodic Advertising Report, LE Enhanced Connection Complete [v2]
is matched with the command which started the connection.
- Commands and events: report a subevent data command or a response
report with fewer entries than announced.
v2:
- Add the fixes of the LE Periodic Advertising Report Data_Length
and of the LE Extended Create Connection PHY blocks. The two
problems Copilot found in v1 were in the [v2] decoders, but they
came from the [v1] decoders they reuse, so they are fixed there.
- The patch which splits the LE Periodic Advertising Report decoding
keeps the new Data_Length check in the shared helper.
Roland Schober (10):
monitor: Fix handle decoding of status_handle_rsp
monitor: Fix data length of LE PA Report decoding
monitor: Fix PHY decoding of LE Ext Create Connection
monitor: Fix advertising handle of LE PA commands
monitor: Fix TX power and CTE Type of LE PA Report
monitor: Fix LE PAST Received decoding
monitor: Fix byte order of LE PA Sync Handle and Properties
monitor: Split LE PA Report decoding into helpers
monitor: Add PAwR HCI commands decoding
monitor: Add PAwR HCI events decoding
monitor/analyze.c | 1 +
monitor/bt.h | 174 ++++++++++++++
monitor/packet.c | 590 +++++++++++++++++++++++++++++++++++++++++-----
3 files changed, 711 insertions(+), 54 deletions(-)
--
2.47.3
next reply other threads:[~2026-11-20 6:39 UTC|newest]
Thread overview: 11+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-11-20 6:38 Roland Schober [this message]
2026-11-20 6:38 ` [PATCH BlueZ v4 01/10] monitor: Fix handle decoding of status_handle_rsp Roland Schober
2026-11-20 6:38 ` [PATCH BlueZ v4 02/10] monitor: Fix data length of LE PA Report decoding Roland Schober
2026-11-20 6:38 ` [PATCH BlueZ v4 03/10] monitor: Fix PHY decoding of LE Ext Create Connection Roland Schober
2026-11-20 6:38 ` [PATCH BlueZ v4 04/10] monitor: Fix advertising handle of LE PA commands Roland Schober
2026-11-20 6:38 ` [PATCH BlueZ v4 05/10] monitor: Fix TX power and CTE Type of LE PA Report Roland Schober
2026-11-20 6:38 ` [PATCH BlueZ v4 06/10] monitor: Fix LE PAST Received decoding Roland Schober
2026-11-20 6:38 ` [PATCH BlueZ v4 07/10] monitor: Fix byte order of LE PA Sync Handle and Properties Roland Schober
2026-11-20 6:38 ` [PATCH BlueZ v4 08/10] monitor: Split LE PA Report decoding into helpers Roland Schober
2026-11-20 6:38 ` [PATCH BlueZ v4 09/10] monitor: Add PAwR HCI commands decoding Roland Schober
2026-11-20 6:38 ` [PATCH BlueZ v4 10/10] monitor: Add PAwR HCI events decoding Roland Schober
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20261120063856.16585-1-rolandschober1968@gmail.com \
--to=rolandschober1968@gmail.com \
--cc=linux-bluetooth@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox