- Jun 23, 2022
-
-
Jens Axboe authored
In prior kernels, we did file assignment always at prep time. This meant that req->task == current. But after deferring that assignment and then pushing the inflight tracking back in, we've got the inflight tracking using current when it should in fact now be using req->task. Fixup that error introduced by adding the inflight tracking back after file assignments got modifed. Fixes: 9cae36a0 ("io_uring: reinstate the inflight tracking") Signed-off-by:
Jens Axboe <axboe@kernel.dk>
-
Dylan Yudaken authored
The TP_printk macro's are not supposed to use custom code ([1]) or else tools such as perf cannot use these events. Convert the opcode string representation to use the __string wiring that the event framework provides ([2]). [1]: https://lwn.net/Articles/379903/ [2]: https://lwn.net/Articles/381064/ Fixes: 033b87d2 ("io_uring: use the text representation of ops in trace") Signed-off-by:
Dylan Yudaken <dylany@fb.com> Link: https://lore.kernel.org/r/20220623083743.2648321-1-dylany@fb.com [axboe: fixup spurious removal of sq_thread assignment] Signed-off-by:
-
- Jun 21, 2022
-
-
Pavel Begunkov authored
We have re-polling for partial IO, so a request can be polled twice. If it used two poll entries the first time then on the second io_arm_poll_handler() it will find the old apoll entry and NULL kmalloc()'ed second entry, i.e. apoll->double_poll, so leaking it. Fixes: 10c87333 ("io_uring: allow re-poll if we made progress") Signed-off-by:
Pavel Begunkov <asml.silence@gmail.com> Link: https://lore.kernel.org/r/fee2452494222ecc7f1f88c8fb659baef971414a.1655852245.git.asml.silence@gmail.com Signed-off-by:
-
Pavel Begunkov authored
Leaving ip.error set when a request was punted to task_work execution is problematic, don't forget to clear it. Fixes: aa43477b ("io_uring: poll rework") Signed-off-by:
-
Pavel Begunkov authored
Don't forget to cancel all linked requests of poll request when __io_arm_poll_handler() failed. Fixes: aa43477b ("io_uring: poll rework") Signed-off-by:
-
Pavel Begunkov authored
apoll_events should be set once in the beginning of poll arming just as poll->events and not change after. However, currently io_uring resets it on each __io_poll_execute() for no clear reason. There is also a place in __io_arm_poll_handler() where we add EPOLLONESHOT to downgrade a multishot, but forget to do the same thing with ->apoll_events, which is buggy. Fixes: 81459350 ("io_uring: cache req->apoll->events in req->cflags") Signed-off-by:
Hao Xu <howeyxu@tencent.com> Link: https://lore.kernel.org/r/0aef40399ba75b1a4d2c2e85e6e8fd93c02fc6e4.1655814213.git.asml.silence@gmail.com Signed-off-by:
-
Jens Axboe authored
With the dropping of the IOPOLL checking in the per-opcode handlers, we inadvertently left two checks in the recv/recvmsg and send/sendmsg prep handlers for the same thing, and one of them includes addr2 which holds the flags for these opcodes. Fix it up and kill the redundant checks. Signed-off-by:
-
- Jun 20, 2022
-
-
Jens Axboe authored
If we mark for reissue, we assume that the buffer will remain stable. Hence if are using a provided buffer, we need to ensure that we stick with it for the duration of that request. This only affects block devices that use provided buffers, as those are the only ones that get marked with REQ_F_REISSUE. Signed-off-by:
-
- Jun 17, 2022
-
-
Jens Axboe authored
io_arm_poll_handler() will recycle the buffer appropriately if we end up arming poll (or if we're ready to retry), but not for the io-wq case if we have attempted poll first. Explicitly recycle the buffer to avoid both hanging on to it too long, but also to avoid multiple reads grabbing the same one. This can happen for ring mapped buffers, since it hasn't necessarily been committed. Fixes: c7fb1942 ("io_uring: add support for ring mapped supplied buffers") Link: https://github.com/axboe/liburing/issues/605 Signed-off-by:
-
- Jun 16, 2022
-
-
Dylan Yudaken authored
io_req_task_prio_work_add has a strict assumption that it will only be used with io_req_task_complete. There is a codepath that assumes this is the case and will not even call the completion function if it is hit. For uring_cmd with an arbitrary completion function change the call to the correct non-priority version. Fixes: ee692a21 ("fs,io_uring: add infrastructure for uring-cmd") Signed-off-by:
-
Jens Axboe authored
For recv/recvmsg, IO either completes immediately or gets queued for a retry. This isn't the case for read/readv, if eg a normal file or a block device is used. Here, an operation can get queued with the block layer. If this happens, ring mapped buffers must get committed immediately to avoid that the next read can consume the same buffer. Check if we're dealing with pollable file, when getting a new ring mapped provided buffer. If it's not, commit it immediately rather than wait post issue. If we don't wait, we can race with completions coming in, or just plain buffer reuse by committing after a retry where others could have grabbed the same buffer. Fixes: c7fb1942 ("io_uring: add support for ring mapped supplied buffers") Reviewed-by:
-
- Jun 15, 2022
-
-
Pavel Begunkov authored
Don't let io_fill_cqe_aux() post 16B cqes for CQE32 rings, neither the kernel nor the userspace expect this to happen. Fixes: 76c68fbf ("io_uring: enable CQE32") Signed-off-by:
-
Pavel Begunkov authored
In preparation for the following patch, inline __io_fill_cqe(), there is only one user. Signed-off-by:
-
Pavel Begunkov authored
We don't really know the state of req->extra{1,2] fields in __io_fill_cqe_req(), if an opcode handler is not aware of CQE32 option, it never sets them up properly. Track the state of those fields with a request flag. Fixes: 76c68fbf ("io_uring: enable CQE32") Signed-off-by: -
Pavel Begunkov authored
The only user of io_req_complete32()-like functions is cmd requests. Instead of keeping the whole complete32 family, remove them and provide the extras in already added for inline completions req->extra{1,2}. When fill_cqe_res() finds CQE32 option enabled it'll use those fields to fill a 32B cqe. Signed-off-by: -
Pavel Begunkov authored
We want just one function that will handle both normal cqes and 32B cqes. Combine __io_fill_cqe_req() and __io_fill_cqe_req32(). It's still not entirely correct yet, but saves us from cases when we fill an CQE of a wrong size. Fixes: 76c68fbf ("io_uring: enable CQE32") Signed-off-by:
-
Pavel Begunkov authored
There are too many cqe filling helpers, kill __io_fill_cqe{32}_req(), use __io_fill_cqe{32}_req_filled() instead, and then rename it. It'll simplify fixing in following patches. Signed-off-by:
-
- Jun 14, 2022
-
-
Pavel Begunkov authored
This partially reverts a7c41b46 Even though IORING_CLOSE_FD_AND_FILE_SLOT might save cycles for some users, but it tries to do two things at a time and it's not clear how to handle errors and what to return in a single result field when one part fails and another completes well. Kill it for now. Signed-off-by:
-
Pavel Begunkov authored
This reverts commit 3d200242. Buffer selection with nops was used for debugging and benchmarking but is useless in real life. Let's revert it before it's released. Signed-off-by:
-
Pavel Begunkov authored
This reverts commit 2bb04df7. CQE32 nops were used for debugging and benchmarking but it doesn't target any real use case. Revert it, we can return it back if someone finds a good way to use it. Signed-off-by:
-
- Jun 13, 2022
-
-
https://github.com/isilence/linuxJens Axboe authored
Pull io_uring fixes from Pavel. * 'io_uring/io_uring-5.19' of https://github.com/isilence/linux: io_uring: fix double unlock for pbuf select io_uring: kbuf: fix bug of not consuming ring buffer in partial io case io_uring: openclose: fix bug of closing wrong fixed file io_uring: fix not locked access to fixed buf table io_uring: fix races with buffer table unregister io_uring: fix races with file table unregister
-
Dylan Yudaken authored
The type of head and tail do not allow more than 2^15 entries in a provided buffer ring, so do not allow this. At 2^16 while each entry can be indexed, there is no way to disambiguate full vs empty. Signed-off-by:
-
Dylan Yudaken authored
The type of head needs to match that of tail in order for rollover and comparisons to work correctly. Without this change the comparison of tail to head might incorrectly allow io_uring to use a buffer that userspace had not given it. Fixes: c7fb1942 ("io_uring: add support for ring mapped supplied buffers") Signed-off-by:
-
Dylan Yudaken authored
When indexing into a provided buffer ring, do not subtract 1 from the index. Fixes: c7fb1942 ("io_uring: add support for ring mapped supplied buffers") Signed-off-by:
-
Pavel Begunkov authored
io_buffer_select(), which is the only caller of io_ring_buffer_select(), fully handles locking, mutex unlock in io_ring_buffer_select() will lead to double unlock. Fixes: c7fb1942 ("io_uring: add support for ring mapped supplied buffers") Signed-off-by:
-
Hao Xu authored
When we use ring-mapped provided buffer, we should consume it before arm poll if partial io has been done. Otherwise the buffer may be used by other requests and thus we lost the data. Fixes: c7fb1942 ("io_uring: add support for ring mapped supplied buffers") Signed-off-by:
-
Hao Xu authored
Don't update ret until fixed file is closed, otherwise the file slot becomes the error code. Fixes: a7c41b46 ("io_uring: let IORING_OP_FILES_UPDATE support choosing fixed file slots") Signed-off-by:
-
Pavel Begunkov authored
We can look inside the fixed buffer table only while holding ->uring_lock, however in some cases we don't do the right async prep for IORING_OP_{WRITE,READ}_FIXED ending up with NULL req->imu forcing making an io-wq worker to try to resolve the fixed buffer without proper locking. Move req->imu setup into early req init paths, i.e. io_prep_rw(), which is called unconditionally for rw requests and under uring_lock. Fixes: 634d00df ("io_uring: add full-fledged dynamic buffers support") Signed-off-by: -
Pavel Begunkov authored
Fixed buffer table quiesce might unlock ->uring_lock, potentially letting new requests to be submitted, don't allow those requests to use the table as they will race with unregistration. Reported-and-tested-by:
van fantasy <g1042620637@gmail.com> Fixes: bd54b6fe ("io_uring: implement fixed buffers registration similar to fixed files") Signed-off-by:
-
Pavel Begunkov authored
Fixed file table quiesce might unlock ->uring_lock, potentially letting new requests to be submitted, don't allow those requests to use the table as they will race with unregistration. Reported-and-tested-by:
-
- Jun 12, 2022
-
-
Linus Torvalds authored
-
Linus Torvalds authored
Merge tag 'platform-drivers-x86-v5.19-2' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86 Pull x86 platform driver fixes from Hans de Goede: "Highlights: - Fix hp-wmi regression on HP Omen laptops introduced in 5.18 - Several hardware-id additions - A couple of other tiny fixes" * tag 'platform-drivers-x86-v5.19-2' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86: platform/x86/intel: hid: Add Surface Go to VGBS allow list platform/x86: hp-wmi: Use zero insize parameter only when supported platform/x86: hp-wmi: Resolve WMI query failures on some devices platform/x86: gigabyte-wmi: Add support for B450M DS3H-CF platform/x86: gigabyte-wmi: Add Z690M AORUS ELITE AX DDR4 support platform/x86: barco-p50-gpio: Add check for platform_driver_register platform/x86/intel: pmc: Support Intel Raptorlake P platform/x86/intel: Fix pmt_crashlog array reference platform/mellanox: Add static in struct declaration. platform/mellanox: Spelling s/platfom/platform/
-
git://git.kernel.org/pub/scm/linux/kernel/git/tj/wqLinus Torvalds authored
Pull workqueue fixes from Tejun Heo: "Tetsuo's patch to trigger build warnings if system-wide wq's are flushed along with a TP type update and trivial comment update" * tag 'wq-for-5.19-rc1-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq: workqueue: Switch to new kerneldoc syntax for named variable macro argument workqueue: Fix type of cpu in trace event workqueue: Wrap flush_workqueue() using a macro
-
Linus Torvalds authored
Merge tag 'kbuild-fixes-v5.19' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild Pull Kbuild fixes from Masahiro Yamada: - Make the *.mod build rule portable for POSIX awk - Fix regression of 'make nsdeps' - Make scripts/check-local-export working for older bash versions - Fix scripts/gdb to extract the .config data from vmlinux * tag 'kbuild-fixes-v5.19' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild: scripts/gdb: change kernel config dumping method scripts/check-local-export: avoid 'wait $!' for process substitution scripts/nsdeps: adjust to the format change of *.mod files kbuild: avoid regex RS for POSIX awk
-
git://git.samba.org/sfrench/cifs-2.6Linus Torvalds authored
Pull cifs client fixes from Steve French: "Three reconnect fixes, all for stable as well. One of these three reconnect fixes does address a problem with multichannel reconnect, but this does not include the additional fix (still being tested) for dynamically detecting multichannel adapter changes which will improve those reconnect scenarios even more" * tag '5.19-rc1-smb3-client-fixes' of git://git.samba.org/sfrench/cifs-2.6: cifs: populate empty hostnames for extra channels cifs: return errors during session setup during reconnects cifs: fix reconnect on smb3 mount types
-
git://git.kernel.org/pub/scm/linux/kernel/git/crng/randomLinus Torvalds authored
Pull random number generator fixes from Jason Donenfeld: - A fix for a 5.19 regression for a case in which early device tree initializes the RNG, which flips a static branch. On most plaforms, jump labels aren't initialized until much later, so this caused splats. On a few mailing list threads, we cooked up easy fixes for arm64, arm32, and risc-v. But then things looked slightly more involved for xtensa, powerpc, arc, and mips. And at that point, when we're patching 7 architectures in a place before the console is even available, it seems like the cost/risk just wasn't worth it. So random.c works around it now by checking the already exported `static_key_initialized` boolean, as though somebody already ran into this issue in the past. I'm not super jazzed about that; it'd be prettier to not have to complicate downstream code. But I suppose it's practical. - A few small code nits and adding a missing __init annotation. - A change to the default config values to use the cpu and bootloader's seeds for initializing the RNG earlier. This brings them into line with what all the distros do (Fedora/RHEL, Debian, Ubuntu, Gentoo, Arch, NixOS, Alpine, SUSE, and Void... at least), and moreover will now give us test coverage in various test beds that might have caught the above device tree bug earlier. - A change to WireGuard CI's configuration to increase test coverage around the RNG. - A documentation comment fix to unrelated maintainerless CRC code that I was asked to take, I guess because it has to do with polynomials (which the RNG thankfully no longer uses). * tag 'random-5.19-rc2-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/crng/random: wireguard: selftests: use maximum cpu features and allow rng seeding random: remove rng_has_arch_random() random: credit cpu and bootloader seeds by default random: do not use jump labels before they are initialized random: account for arch randomness in bits random: mark bootloader randomness code as __init random: avoid checking crng_ready() twice in random_init() crc-itu-t: fix typo in CRC ITU-T polynomial comment
-
Duke Lee authored
The Surface Go reports Chassis Type 9 (Laptop,) so the device needs to be added to dmi_vgbs_allow_list to enable tablet mode when an attached Type Cover is folded back. BugLink: https://github.com/linux-surface/linux-surface/issues/837 Signed-off-by:
Duke Lee <krnhotwings@gmail.com> Reviewed-by:
Andy Shevchenko <andy.shevchenko@gmail.com> Link: https://lore.kernel.org/r/20220607213654.5567-1-krnhotwings@gmail.com Reviewed-by:
Hans de Goede <hdegoede@redhat.com> Signed-off-by:
-
Bedant Patnaik authored
commit be9d73e6 ("platform/x86: hp-wmi: Fix 0x05 error code reported by several WMI calls") and commit 12b19f14 ("platform/x86: hp-wmi: Fix hp_wmi_read_int() reporting error (0x05)") cause ACPI BIOS Error (bug): Attempt to CreateField of length zero (20211217/dsopcode-133) because of the ACPI method HWMC, which unconditionally creates a Field of size (insize*8) bits: CreateField (Arg1, 0x80, (Local5 * 0x08), DAIN) In cases where args->insize = 0, the Field size is 0, resulting in an error. Fix this by using zero insize only if 0x5 error code is returned Tested on Omen 15 AMD (2020) board ID: 8786. Fixes: be9d73e6 ("platform/x86: hp-wmi: Fix 0x05 error code reported by several WMI calls") Signed-off-by:
Bedant Patnaik <bedant.patnaik@gmail.com> Tested-by:
Jorge Lopez <jorge.lopez2@hp.com> Link: https://lore.kernel.org/r/41be46743d21c78741232a47bbb5f1cdbcc3d21e.camel@gmail.com Reviewed-by:
-
Jorge Lopez authored
WMI queries fail on some devices where the ACPI method HWMC unconditionally attempts to create Fields beyond the buffer if the buffer is too small, this breaks essential features such as power profiles: CreateByteField (Arg1, 0x10, D008) CreateByteField (Arg1, 0x11, D009) CreateByteField (Arg1, 0x12, D010) CreateDWordField (Arg1, 0x10, D032) CreateField (Arg1, 0x80, 0x0400, D128) In cases where args->data had zero length, ACPI BIOS Error (bug): AE_AML_BUFFER_LIMIT, Field [D008] at bit offset/length 128/8 exceeds size of target Buffer (128 bits) (20211217/dsopcode-198) was obtained. ACPI BIOS Error (bug): AE_AML_BUFFER_LIMIT, Field [D009] at bit offset/length 136/8 exceeds size of target Buffer (136bits) (20211217/dsopcode-198) The original code created a buffer size of 128 bytes regardless if the WMI call required a smaller buffer or not. This particular behavior occurs in older BIOS and reproduced in OMEN laptops. Newer BIOS handles buffer sizes properly and meets the latest specification requirements. This is the reason why testing with a dynamically allocated buffer did not uncover any failures with the test systems at hand. This patch was tested on several OMEN, Elite, and Zbooks. It was confirmed the patch resolves HPWMI_FAN GET/SET calls in an OMEN Laptop 15-ek0xxx. No problems were reported when testing on several Elite and Zbooks notebooks. Fixes: 4b4967cb ("platform/x86: hp-wmi: Changing bios_args.data to be dynamically allocated") Signed-off-by: -
Jonathan Neuschäfer authored
The syntax without dots is available since commit 43756e34 ("scripts/kernel-doc: Add support for named variable macro arguments"). The same HTML output is produced with and without this patch. Signed-off-by:
Jonathan Neuschäfer <j.neuschaefer@gmx.net> Acked-by:
Tejun Heo <tj@kernel.org> Signed-off-by:
-
