kernel/linux.git/sound/core/ump.c, branch v7.0.10

Convert more 'alloc_obj' cases to default GFP_KERNEL arguments

2026-02-22T04:03:00+00:00

This converts some of the visually simpler cases that have been split over multiple lines. I only did the ones that are easy to verify the resulting diff by having just that final GFP_KERNEL argument on the next line. Somebody should probably do a proper coccinelle script for this, but for me the trivial script actually resulted in an assertion failure in the middle of the script. I probably had made it a bit _too_ trivial. So after fighting that far a while I decided to just do some of the syntactically simpler cases with variations of the previous 'sed' scripts. The more syntactically complex multi-line cases would mostly really want whitespace cleanup anyway. Signed-off-by: Linus Torvalds

Convert 'alloc_obj' family to use the new default GFP_KERNEL argument

2026-02-22T01:09:51+00:00

This was done entirely with mindless brute force, using git grep -l '\

treewide: Replace kmalloc with kmalloc_obj for non-scalar types

2026-02-21T09:02:28+00:00

This is the result of running the Coccinelle script from scripts/coccinelle/api/kmalloc_objs.cocci. The script is designed to avoid scalar types (which need careful case-by-case checking), and instead replace kmalloc-family calls that allocate struct or union object instances: Single allocations: kmalloc(sizeof(TYPE), ...) are replaced with: kmalloc_obj(TYPE, ...) Array allocations: kmalloc_array(COUNT, sizeof(TYPE), ...) are replaced with: kmalloc_objs(TYPE, COUNT, ...) Flex array allocations: kmalloc(struct_size(PTR, FAM, COUNT), ...) are replaced with: kmalloc_flex(*PTR, FAM, COUNT, ...) (where TYPE may also be *VAR) The resulting allocations no longer return "void *", instead returning "TYPE *". Signed-off-by: Kees Cook

ALSA: rawmidi: Make tied_device=0 as default / unknown

2025-01-14T15:52:07+00:00

In the original change, rawmidi_info.tied_device showed -1 for the unknown or untied device. But this would require the user-space to check the protocol version and judge the value conditionally, which is rather error-prone. Instead, set the tied_device = 0 to be default as unknown, and indicate the real device with the offset 1, for achieving more backward compatibility. Suggested-by: Jaroslav Kysela Link: https://patch.msgid.link/20250114104711.19197-1-tiwai@suse.de Signed-off-by: Takashi Iwai

ALSA: seq: Notify UMP EP and FB changes

2025-01-12T12:12:21+00:00

So far we notify the sequencer client and port changes upon UMP FB changes, but those aren't really corresponding to the UMP updates. e.g. when a FB info gets updated, it's not notified but done only when some of sequencer port attribute is changed. This is no ideal behavior. This patch adds the two new sequencer event types for notifying the UMP EP and FB changes via the announce port. The new event takes snd_seq_ev_ump_notify type data, which is compatible with snd_seq_addr (where the port number is replaced with the block number). The events are sent when the EP and FB info gets updated explicitly via ioctl, or the backend UMP receives the corresponding UMP messages. The sequencer protocol version is bumped to 1.0.5 along with it. Signed-off-by: Takashi Iwai Link: https://patch.msgid.link/20250110155943.31578-9-tiwai@suse.de

ALSA: ump: Update rawmidi name per EP name update

2025-01-12T12:12:21+00:00

The rawmidi name string should be updated dynamically when the device receives the UMP EP name update, too. Both the core and legacy rawmidi names are updated. Signed-off-by: Takashi Iwai Link: https://patch.msgid.link/20250110155943.31578-7-tiwai@suse.de

ALSA: ump: Copy safe string name to rawmidi

2025-01-12T12:12:20+00:00

The UMP helper didn't set up the rawmidi name string by itself but left it to the driver. But since the only user (USB MIDI2 driver) picks up the UMP info name string to the rawmidi name as default, it's better to set up in the UMP core side. Meanwhile, UMP receives the EP name string from the device, and it might contain garbage letters. We should purify the string to be usable for the kernel as done previously for UMP Group names. This implements the copy of the UMP info name string into the rawmidi name at the creation of UMP EP object in a safe way to strip the non-ASCII or non-printable characters. Also, change the reference from the legacy rawmidi and other places to rawmidi name field instead of ump info; this assures the sane strings. Signed-off-by: Takashi Iwai Link: https://patch.msgid.link/20250110155943.31578-6-tiwai@suse.de

ALSA: ump: Copy FB name string more safely

2025-01-12T12:12:20+00:00

The UMP group names are referred as the corresponding sequencer port names, hence they should be proper ASCII strings. OTOH, the UMP group names are composed from the UMP FB strings that are received from the device; i.e. a device may give some bogus letters and we can't trust them fully. To assure that the group names consist of the proper ASCII strings, replace the normal string copy and append operations with special ones that strip the non-printable letters. Signed-off-by: Takashi Iwai Link: https://patch.msgid.link/20250110155943.31578-5-tiwai@suse.de

ALSA: rawmidi: Show substream activity in info ioctl

2025-01-12T12:12:20+00:00

The UMP legacy rawmidi may turn on/off the substream dynamically depending on the UMP Function Block information. So far, there was no direct way to know whether the substream is disabled (inactive) or not; at most one can take a look at the substream name string or try to open and get -ENODEV. This patch extends the rawmidi info ioctl to show the current inactive state of the given substream. When the selected substream is inactive, info flags field contains the new bit flag SNDRV_RAWMIDI_INFO_STREAM_INACTIVE. Signed-off-by: Takashi Iwai Link: https://patch.msgid.link/20250110155943.31578-3-tiwai@suse.de

ALSA: rawmidi: Expose the tied device number in info ioctl

2025-01-12T12:12:20+00:00

The UMP legacy rawmidi is derived from the UMP rawmidi, but currently there is no way to know which device is involved in other side. This patch extends the rawmidi info ioctl to show the tied device number. As default it stores -1, indicating that no tied device. Signed-off-by: Takashi Iwai Link: https://patch.msgid.link/20250110155943.31578-2-tiwai@suse.de