July 2025 - gerrit-log - lists.osmocom.org

[M] Change in libosmocore[master]: Automatically increase io_uring, if too small.

by laforge

Attention is currently required from: jolly, pespin. laforge has posted comments on this change by jolly. ( https://gerrit.osmocom.org/c/libosmocore/+/40725?usp=email ) Change subject: Automatically increase io_uring, if too small. ...................................................................... Patch Set 2: (6 comments) File src/core/osmo_io_internal.h: https://gerrit.osmocom.org/c/libosmocore/+/40725/comment/a53bcc51_10cebfac?… : PS2, Line 110: void *read_ring[IOFD_MSGHDR_READ_SQES]; some comments here would be useful to explan what those fields are and how hey are used (not just the read_ring now, but also the previously-introduced read_msghdr, read_len, ... Also: why are those pointers void? if it points to a ring, shouldn't it be 'struct io_uring *' for read_ring and write_ring? File src/core/osmo_io_uring.c: https://gerrit.osmocom.org/c/libosmocore/+/40725/comment/a0eb6d91_79685d7d?… : PS2, Line 67: g_io_uring_queue pleaes use a more descriptive name. The variable is not a qeuue, but it is the size of the ring. Something like g_io_uring_size or g_io_uring_num_entries would be more descriptive. https://gerrit.osmocom.org/c/libosmocore/+/40725/comment/06962c19_78cb567d?… : PS2, Line 113: OSMO_IO_URING_QUEUE like the variable name in the code, the environment variable name shold be self-descriptive. Also, since it's the *initial* size, something like OSMO_IO_URING_INITIAL_{SIZE,NUM_ENTRIES} would make most sense. Naming does matter... https://gerrit.osmocom.org/c/libosmocore/+/40725/comment/76d8c11b_e4366e52?… : PS2, Line 163: osmo_iofd_uring_get_sqe the 'osmo_' prefix is reserverd for exported functions. You are adding a static function, it should not have an osmo_ prefix. https://gerrit.osmocom.org/c/libosmocore/+/40725/comment/4b3570e8_75a14c96?… : PS2, Line 172: LOGP(DLIO, LOGL_NOTICE, "io_uring too small to handle all SQEs, increasing size to %d.\n", g_io_uring_queue); minor: while the current implementation alsways doubles the size, it might still be useful to print the old [too small] size? https://gerrit.osmocom.org/c/libosmocore/+/40725/comment/40117db5_b6cf2823?… : PS2, Line 175: is there code to clean up the old ring after completion of the last cqe of the old ring? -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/40725?usp=email To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings?usp=email Gerrit-MessageType: comment Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: Id9230146acc8d54bfd44834e783c31b37bd64bca Gerrit-Change-Number: 40725 Gerrit-PatchSet: 2 Gerrit-Owner: jolly <andreas(a)eversberg.eu> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: laforge <laforge(a)osmocom.org> Gerrit-CC: pespin <pespin(a)sysmocom.de> Gerrit-Attention: jolly <andreas(a)eversberg.eu> Gerrit-Attention: pespin <pespin(a)sysmocom.de> Gerrit-Comment-Date: Wed, 23 Jul 2025 11:16:38 +0000 Gerrit-HasComments: Yes Gerrit-Has-Labels: No

3 days, 22 hours

1
0
0 0

[M] Change in libosmocore[master]: Add multiple messages buffers to io_uring write operations

by jolly

Attention is currently required from: pespin. jolly has posted comments on this change by jolly. ( https://gerrit.osmocom.org/c/libosmocore/+/40493?usp=email ) Change subject: Add multiple messages buffers to io_uring write operations ...................................................................... Patch Set 7: (5 comments) File src/core/osmo_io.c: https://gerrit.osmocom.org/c/libosmocore/+/40493/comment/38629ece_28e3e269?… : PS4, Line 248: lh = iofd->tx_queue.msg_queue.prev; > You added it to libosmocore but apparently not using it here? I fogot to remove this function. I was already using llist_last_entry_or_null(). https://gerrit.osmocom.org/c/libosmocore/+/40493/comment/1d4d96d5_8c0d3497?… : PS4, Line 444: /* Re-enqueue remaining buffers. */ > Here we should improve it by adding new msg from the tx_queue into the iov if we aren't yet doing it […] Then we must change the way the tx_queue stores. Currently it stores msghdrs, but we would need to store msgbs. It makes this much more complex. https://gerrit.osmocom.org/c/libosmocore/+/40493/comment/bfff97ec_18737813?… : PS4, Line 465: iofd->io_ops.write_cb(iofd, (rc < 0) ? rc : chunk, msg); > Again, by incorporation the negative value case into chunk variable, we can simplify all these lines […] Done https://gerrit.osmocom.org/c/libosmocore/+/40493/comment/d0e4fd57_ed6a5e70?… : PS4, Line 479: msgb_free(msghdr->msg[idx]); > msghdr->msg[idx] = NULL; Done https://gerrit.osmocom.org/c/libosmocore/+/40493/comment/6e737520_f1bf4096?… : PS4, Line 517: msghdr = iofd_txqueue_tail(iofd); > Not done, not answered, no comment explaining in code. I wrote it abovethe function call llist_last_entry_or_null(). -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/40493?usp=email To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings?usp=email Gerrit-MessageType: comment Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: I8c4e0a785cf66becd7fb5b2caf718c9724b56686 Gerrit-Change-Number: 40493 Gerrit-PatchSet: 7 Gerrit-Owner: jolly <andreas(a)eversberg.eu> Gerrit-Reviewer: Jenkins Builder Gerrit-CC: pespin <pespin(a)sysmocom.de> Gerrit-Attention: pespin <pespin(a)sysmocom.de> Gerrit-Comment-Date: Wed, 23 Jul 2025 10:46:31 +0000 Gerrit-HasComments: Yes Gerrit-Has-Labels: No Comment-In-Reply-To: jolly <andreas(a)eversberg.eu> Comment-In-Reply-To: pespin <pespin(a)sysmocom.de>

3 days, 23 hours

1
0
0 0

[M] Change in libosmocore[master]: Add multiple messages buffers to struct iofd_msghdr

by jolly

Attention is currently required from: laforge, pespin. jolly has posted comments on this change by jolly. ( https://gerrit.osmocom.org/c/libosmocore/+/40491?usp=email ) Change subject: Add multiple messages buffers to struct iofd_msghdr ...................................................................... Patch Set 6: (2 comments) File src/core/osmo_io_internal.h: https://gerrit.osmocom.org/c/libosmocore/+/40491/comment/7e36cd82_3a36aa3c?… : PS5, Line 143: int io_len; > I wonder whether this is needed or can be inferred when needed by looking up NULL ptrs in "msg" fiel […] Without it, we would have extra loops. I want to avoid that. Alternatively we could use a list of msgb+iovec in an extra structure. This would make it much more complex. File src/core/osmo_io_internal.h: https://gerrit.osmocom.org/c/libosmocore/+/40491/comment/5b827776_3017afd8?… : PS3, Line 140: int msg_max; > This should definetly be configurable over API per osmo_iofd object. […] Done -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/40491?usp=email To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings?usp=email Gerrit-MessageType: comment Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: I4fb1067de4615cc22cc6caf99b481491e7f2ef92 Gerrit-Change-Number: 40491 Gerrit-PatchSet: 6 Gerrit-Owner: jolly <andreas(a)eversberg.eu> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: pespin <pespin(a)sysmocom.de> Gerrit-CC: laforge <laforge(a)osmocom.org> Gerrit-Attention: laforge <laforge(a)osmocom.org> Gerrit-Attention: pespin <pespin(a)sysmocom.de> Gerrit-Comment-Date: Wed, 23 Jul 2025 10:46:23 +0000 Gerrit-HasComments: Yes Gerrit-Has-Labels: No Comment-In-Reply-To: jolly <andreas(a)eversberg.eu> Comment-In-Reply-To: laforge <laforge(a)osmocom.org> Comment-In-Reply-To: pespin <pespin(a)sysmocom.de>

3 days, 23 hours

1
0
0 0

[M] Change in libosmocore[master]: Avoid reusing pending buffer; append incoming data instead

by jolly

Attention is currently required from: laforge, pespin. jolly has posted comments on this change by jolly. ( https://gerrit.osmocom.org/c/libosmocore/+/40584?usp=email ) Change subject: Avoid reusing pending buffer; append incoming data instead ...................................................................... Patch Set 4: (5 comments) Commit Message: https://gerrit.osmocom.org/c/libosmocore/+/40584/comment/ff2050e4_17cbe464?… : PS3, Line 17: This change ensures that newly received data is appended to the existing > First you say we don't want to reuse the pending buffer, but here you say "This change ensures that […] Maybe I should state that I don't want to reuse the pending buffer for the read operation. Instead I copy from the buffer that was used for the read operation to the pending buffer. File src/core/osmo_io.c: https://gerrit.osmocom.org/c/libosmocore/+/40584/comment/2b7b3610_96fc97ba?… : PS3, Line 177: /*! convenience wrapper to call msgb_alloc with parameters from osmo_io_fd (with extra size) */ > "with extra size" here doesn't apply anymore? Done https://gerrit.osmocom.org/c/libosmocore/+/40584/comment/d54ad17c_59b4ea84?… : PS3, Line 178: struct msgb *iofd_msgb_alloc2(struct osmo_io_fd *iofd, size_t size) > This is probably breaking the promise that msgb passed to the app are at least of a certain size con […] In case the pending buffer is used, the size may be larger, but never smaller. I don't see any problem with that. It happens when reading/receiing. The application should only care about the length of data in the buffer not the size. https://gerrit.osmocom.org/c/libosmocore/+/40584/comment/3ca7bf60_cd0f8648?… : PS3, Line 335: * If the pending message is not large enough, create a larger message. */ > I may be wrong, but I'd expect user to configure iofd with a max_size big enough to allow whatever s […] Imagine the pending buffer is almost full. Now we read extra data that is larger than the remaining space of the pending buffer. In order to merge them, a larger buffer is needed. https://gerrit.osmocom.org/c/libosmocore/+/40584/comment/f6d10647_ee937061?… : PS3, Line 344: memcpy(msgb_put(iofd->pending, msgb_length(msg)), msgb_data(msg), msgb_length(msg)); > Can we avoid 2 memcpys (here and above) in the specific code path? How? The fist memcopy is used, if the pending buffer must be enlarged. The second memcopy will append the newly received data to the pending buffer. -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/40584?usp=email To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings?usp=email Gerrit-MessageType: comment Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: I08df9736ccc5e9a7df61ca6dcf94629ee010752f Gerrit-Change-Number: 40584 Gerrit-PatchSet: 4 Gerrit-Owner: jolly <andreas(a)eversberg.eu> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: laforge <laforge(a)osmocom.org> Gerrit-Reviewer: pespin <pespin(a)sysmocom.de> Gerrit-Attention: laforge <laforge(a)osmocom.org> Gerrit-Attention: pespin <pespin(a)sysmocom.de> Gerrit-Comment-Date: Wed, 23 Jul 2025 10:46:18 +0000 Gerrit-HasComments: Yes Gerrit-Has-Labels: No Comment-In-Reply-To: pespin <pespin(a)sysmocom.de>

3 days, 23 hours

1
0
0 0

[S] Change in libosmocore[master]: Allow io_uring_submit batching just ahead of poll/select

by jolly

Attention is currently required from: laforge, pespin. jolly has posted comments on this change by jolly. ( https://gerrit.osmocom.org/c/libosmocore/+/40489?usp=email ) Change subject: Allow io_uring_submit batching just ahead of poll/select ...................................................................... Patch Set 4: (4 comments) Patchset: PS3: > I'm still not convinced we want to add more envvars. […] Right now I use the envvars for running TTCN3 tests with different io_uring features. It makes it easier to compare. File src/core/osmo_io_uring.c: https://gerrit.osmocom.org/c/libosmocore/+/40489/comment/889cdea7_e7f90d3f?… : PS2, Line 186: if (!g_io_uring_batch) > maybe use the OSMO_LIKELY macro for branch prediction optimizaton, as we assume the default is the c […] Done https://gerrit.osmocom.org/c/libosmocore/+/40489/comment/afbdc520_829fc90e?… : PS2, Line 329: if (!g_io_uring_batch) > same here and further below Done File src/core/select.c: https://gerrit.osmocom.org/c/libosmocore/+/40489/comment/a9058573_f6aff482?… : PS3, Line 440: osmo_io_uring_submit(); > I'd say it makes more sense to have the OSMO_UNLIKELY if here directly, to avoid an extra call+ret o […] Done -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/40489?usp=email To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings?usp=email Gerrit-MessageType: comment Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: Id34fe2ced32c63d15b14810e145744f7509064cc Gerrit-Change-Number: 40489 Gerrit-PatchSet: 4 Gerrit-Owner: jolly <andreas(a)eversberg.eu> Gerrit-Reviewer: Jenkins Builder Gerrit-CC: laforge <laforge(a)osmocom.org> Gerrit-CC: pespin <pespin(a)sysmocom.de> Gerrit-Attention: laforge <laforge(a)osmocom.org> Gerrit-Attention: pespin <pespin(a)sysmocom.de> Gerrit-Comment-Date: Wed, 23 Jul 2025 10:46:11 +0000 Gerrit-HasComments: Yes Gerrit-Has-Labels: No Comment-In-Reply-To: laforge <laforge(a)osmocom.org> Comment-In-Reply-To: pespin <pespin(a)sysmocom.de>

3 days, 23 hours

1
0
0 0

[M] Change in libosmocore[master]: Send multiple read/recvfrom SQEs in advance

by jolly

Attention is currently required from: jolly. Hello Jenkins Builder, I'd like you to reexamine a change. Please visit https://gerrit.osmocom.org/c/libosmocore/+/40494?usp=email to look at the new patch set (#7). The following approvals got outdated and were removed: Verified+1 by Jenkins Builder Change subject: Send multiple read/recvfrom SQEs in advance ...................................................................... Send multiple read/recvfrom SQEs in advance Multiple read or recvfrom operations can be submitted via SQEs when using io_uring. This allows reading multiple packet / more data between calls of osmo_select_main() the main loop. Having more than one SQE submitted is optional and the number can be controlled via environment variable. Related: OS#6705 Change-Id: Id50a8900fa2fe6de553e5025feae7e1e8d501e30 --- M src/core/osmo_io_internal.h M src/core/osmo_io_uring.c 2 files changed, 46 insertions(+), 13 deletions(-) git pull ssh://gerrit.osmocom.org:29418/libosmocore refs/changes/94/40494/7 -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/40494?usp=email To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings?usp=email Gerrit-MessageType: newpatchset Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: Id50a8900fa2fe6de553e5025feae7e1e8d501e30 Gerrit-Change-Number: 40494 Gerrit-PatchSet: 7 Gerrit-Owner: jolly <andreas(a)eversberg.eu> Gerrit-Reviewer: Jenkins Builder Gerrit-CC: pespin <pespin(a)sysmocom.de> Gerrit-Attention: jolly <andreas(a)eversberg.eu>

3 days, 23 hours

1
0
0 0

[M] Change in libosmocore[master]: Automatically increase io_uring, if too small.

by jolly

Attention is currently required from: jolly, pespin. Hello Jenkins Builder, laforge, I'd like you to reexamine a change. Please visit https://gerrit.osmocom.org/c/libosmocore/+/40725?usp=email to look at the new patch set (#3). The following approvals got outdated and were removed: Verified+1 by Jenkins Builder Change subject: Automatically increase io_uring, if too small. ...................................................................... Automatically increase io_uring, if too small. The ring may be too small to store all SQEs before the kernel can handle them. If this happens, a new ring is allocated with twice of the size of the old ring. The old ring will not be destroyed, as it still contains uncompleted elements. Some of them may never be completed. A pointer to the current ring will be stored within the msghdr structure. It is used when cancelling an SQE. The cancellation must be performed in the same ring where it was created. It is quite unlikely that the old ring cannot store the cancellation SQE. If this happens, the kernel is given some time to handle the SQEs. The initial size of the ring can be changed via environment variable. Related: OS#6705 Change-Id: Id9230146acc8d54bfd44834e783c31b37bd64bca --- M src/core/osmo_io_internal.h M src/core/osmo_io_uring.c 2 files changed, 65 insertions(+), 25 deletions(-) git pull ssh://gerrit.osmocom.org:29418/libosmocore refs/changes/25/40725/3 -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/40725?usp=email To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings?usp=email Gerrit-MessageType: newpatchset Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: Id9230146acc8d54bfd44834e783c31b37bd64bca Gerrit-Change-Number: 40725 Gerrit-PatchSet: 3 Gerrit-Owner: jolly <andreas(a)eversberg.eu> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: laforge <laforge(a)osmocom.org> Gerrit-CC: pespin <pespin(a)sysmocom.de> Gerrit-Attention: jolly <andreas(a)eversberg.eu> Gerrit-Attention: pespin <pespin(a)sysmocom.de>

3 days, 23 hours

1
0
0 0

[M] Change in libosmocore[master]: Add multiple messages buffers to io_uring write operations

by jolly

Attention is currently required from: jolly. Hello Jenkins Builder, I'd like you to reexamine a change. Please visit https://gerrit.osmocom.org/c/libosmocore/+/40493?usp=email to look at the new patch set (#7). The following approvals got outdated and were removed: Verified+1 by Jenkins Builder Change subject: Add multiple messages buffers to io_uring write operations ...................................................................... Add multiple messages buffers to io_uring write operations Multiple message buffers can be writen by sending a single SQE when using io_uring. If there is less data written, the completely written buffers are removed and the partly written buffers are truncated. Afterwards they are re-queued for next write operation. Having more than one buffer is optional and the number can be controlled via environment variable. Related: OS#6705 Change-Id: I8c4e0a785cf66becd7fb5b2caf718c9724b56686 --- M src/core/osmo_io.c M src/core/osmo_io_uring.c 2 files changed, 104 insertions(+), 51 deletions(-) git pull ssh://gerrit.osmocom.org:29418/libosmocore refs/changes/93/40493/7 -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/40493?usp=email To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings?usp=email Gerrit-MessageType: newpatchset Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: I8c4e0a785cf66becd7fb5b2caf718c9724b56686 Gerrit-Change-Number: 40493 Gerrit-PatchSet: 7 Gerrit-Owner: jolly <andreas(a)eversberg.eu> Gerrit-Reviewer: Jenkins Builder Gerrit-CC: pespin <pespin(a)sysmocom.de> Gerrit-Attention: jolly <andreas(a)eversberg.eu>

3 days, 23 hours

1
0
0 0

[M] Change in libosmocore[master]: Add multiple messages buffers to struct iofd_msghdr

by jolly

Attention is currently required from: jolly, laforge. Hello Jenkins Builder, pespin, I'd like you to reexamine a change. Please visit https://gerrit.osmocom.org/c/libosmocore/+/40491?usp=email to look at the new patch set (#6). The following approvals got outdated and were removed: Verified+1 by Jenkins Builder Change subject: Add multiple messages buffers to struct iofd_msghdr ...................................................................... Add multiple messages buffers to struct iofd_msghdr This is a prerequisite for transfering multiple buffers with a single SQE or CQE when using osmo_io with io_uring. The message buffers at struct iofd_msghr, as well as the io vectors are arrays. A maximum number is stored, as well as the current number. Having more than one buffer is optional and the number can be controlled via environment variable. Related: OS#6705 Change-Id: I4fb1067de4615cc22cc6caf99b481491e7f2ef92 --- M include/osmocom/core/osmo_io.h M src/core/osmo_io.c M src/core/osmo_io_internal.h M src/core/osmo_io_poll.c M src/core/osmo_io_uring.c 5 files changed, 48 insertions(+), 16 deletions(-) git pull ssh://gerrit.osmocom.org:29418/libosmocore refs/changes/91/40491/6 -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/40491?usp=email To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings?usp=email Gerrit-MessageType: newpatchset Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: I4fb1067de4615cc22cc6caf99b481491e7f2ef92 Gerrit-Change-Number: 40491 Gerrit-PatchSet: 6 Gerrit-Owner: jolly <andreas(a)eversberg.eu> Gerrit-Reviewer: Jenkins Builder Gerrit-Reviewer: pespin <pespin(a)sysmocom.de> Gerrit-CC: laforge <laforge(a)osmocom.org> Gerrit-Attention: jolly <andreas(a)eversberg.eu> Gerrit-Attention: laforge <laforge(a)osmocom.org>

3 days, 23 hours

1
0
0 0

[S] Change in libosmocore[master]: Allow io_uring_submit batching just ahead of poll/select

by jolly

Attention is currently required from: jolly, laforge. Hello Jenkins Builder, I'd like you to reexamine a change. Please visit https://gerrit.osmocom.org/c/libosmocore/+/40489?usp=email to look at the new patch set (#4). The following approvals got outdated and were removed: Verified+1 by Jenkins Builder Change subject: Allow io_uring_submit batching just ahead of poll/select ...................................................................... Allow io_uring_submit batching just ahead of poll/select Let's add a mode (enabled via the LIBOSMO_IO_URING_BATCH environment variable), where we don't call io_uring_submit() after every operation we add to the submission queue. Rather, do that once before we go into poll. This should massively reduce the amount of io_uring_enter() syscalls we're seeing. Related: OS#6705 Change-Id: Id34fe2ced32c63d15b14810e145744f7509064cc --- M src/core/osmo_io_uring.c M src/core/select.c 2 files changed, 36 insertions(+), 4 deletions(-) git pull ssh://gerrit.osmocom.org:29418/libosmocore refs/changes/89/40489/4 -- To view, visit https://gerrit.osmocom.org/c/libosmocore/+/40489?usp=email To unsubscribe, or for help writing mail filters, visit https://gerrit.osmocom.org/settings?usp=email Gerrit-MessageType: newpatchset Gerrit-Project: libosmocore Gerrit-Branch: master Gerrit-Change-Id: Id34fe2ced32c63d15b14810e145744f7509064cc Gerrit-Change-Number: 40489 Gerrit-PatchSet: 4 Gerrit-Owner: jolly <andreas(a)eversberg.eu> Gerrit-Reviewer: Jenkins Builder Gerrit-CC: laforge <laforge(a)osmocom.org> Gerrit-CC: pespin <pespin(a)sysmocom.de> Gerrit-Attention: jolly <andreas(a)eversberg.eu> Gerrit-Attention: laforge <laforge(a)osmocom.org>

3 days, 23 hours

1
0
0 0

2025

2024

2023

2022

gerrit-log July 2025