aboutsummaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
* libtrivfs: Fix memleak on errorSamuel Thibault2023-01-011-0/+2
|
* fstab: Fix memleak on errorSamuel Thibault2023-01-011-0/+1
|
* libps: Fix erroneous allocation testSamuel Thibault2023-01-011-1/+1
|
* getty: Fix fd leak on errorSamuel Thibault2023-01-011-1/+3
|
* fstab: Fix memleak on errorSamuel Thibault2023-01-011-1/+1
|
* libtrivfs: Avoid passing uninitialized seconds fieldSamuel Thibault2023-01-011-0/+2
| | | | | Even if microseconds being -1 is enough, better not leak uninitialized values.
* term: Drop spurious derefSamuel Thibault2023-01-011-9/+2
| | | | That was a remnant of manual port lookup.
* libnetfs: fix memory leak on errorSamuel Thibault2023-01-011-0/+3
|
* libcons: Do not send event to server when out of rangeSamuel Thibault2023-01-011-8/+6
|
* libftpconn: Fix memory leak on errorSamuel Thibault2023-01-011-1/+4
|
* boot: Fix memory leak on S_io_read errorSamuel Thibault2023-01-011-0/+2
|
* make_peropen: Fix memory leaks on errorSamuel Thibault2023-01-012-4/+15
|
* fsys_get_children: Fix double-free on errorSamuel Thibault2023-01-012-3/+3
|
* mach-defpager: document lock orderingSamuel Thibault2022-12-311-0/+22
|
* proc: Do not reference a task port when allocation failedSamuel Thibault2022-12-241-1/+2
|
* S_mach_notify_new_task: Ignore notification of dead tasksSamuel Thibault2022-12-211-0/+3
| | | | | | | In case the new task notification gets late, the task may have terminated already, and thus a dead name gets provided in the notification. We should just ignore this, otherwise the task hashing would get collisions on all such dead tasks ports.
* proc: Do not deallocate input ports on errorSamuel Thibault2022-12-211-11/+4
| | | | | The error handling code would already deallocate them, thus leading to duplicate deallocation, spurious destroy, and later on port mixup.
* proc: Add task to hashes before requesting death notifySamuel Thibault2022-12-212-6/+5
| | | | | In case the task dies very early and thus the notification arrives very early.
* Fix *printf specifier for user space mach ports.Flavio Cruz2022-12-0613-23/+23
| | | | | mach_port_t are mach_port_name_t and thus require %u instead of %lu. Message-Id: <Y42RELMbulK4xaKM@reue>
* mach-defpager: assert_backtrace() on panicSamuel Thibault2022-12-041-1/+2
| | | | | instead of just exiting: better print a backtrace so we get to know what happened.
* startup: Make essential tasks essential to machSamuel Thibault2022-12-031-0/+2
| | | | | | | | Whenever a Hurd essential task crashes, startup just reboots the system since there's not much that can be done at that point. When we have a kernel debugger, however, we could at least get crashing information, so let's make Hurd's startup tell the kernel which tasks are essential, so it can trigger the debugger whenever they crash.
* mach-defpager: Mark as essentialSamuel Thibault2022-12-031-0/+13
| | | | | So we get to know why the system gets stuck if mach-defpager happens to die.
* hurd_types.defs: redefine io_statbuf_t, timespec_t and fsys_statfsbuf_t as ↵Flavio Cruz2022-11-211-5/+96
| | | | | | true structs Message-Id: <Y3BpLsrddifLOR1z@viriathus>
* Remove include of hurd/hurd_types.defs in iioctl.defs and rioctl.defsFlavio Cruz2022-11-152-2/+0
| | | | | hurd/hurd_types.defs is already included through ioctl_types.defs. Message-Id: <Y3HUZ3KzKnCchYvI@viriathus>
* Add ioctl_types.defs to avoid duplication of type definitions.Flavio Cruz2022-11-133-18/+42
| | | | | Also change type srtentry_t to use the new struct format. Message-Id: <Y3CKiDS3wszn0G4a@viriathus>
* hurd: ensure subsystem directive appears before anything else in *.defsFlavio Cruz2022-11-134-13/+12
| | | | | | | | | | mig is incorrectly initializing basic types twice (init_type) for kernel user and kernel server definitions. To fix this bug, we should initialize the types during a subsystem directive only. This commit ensures all hurd definitions have subsystem at the very start so that these definitions work correctly with a patched mig. Message-Id: <Y3B/9ZCdqYCcd/zf@viriathus>
* Define mig type flock_t as a struct instead of array of intsFlavio Cruz2022-11-111-1/+7
| | | | Message-Id: <Y23Sua0dRTrqcxZl@viriathus>
* Replace vsprintf with vsnprintf in ext2fs/msg.cFlavio Cruz2022-11-091-3/+3
| | | | Message-Id: <Y2s/UPSCuGap3cv6@viriathus>
* Remove duplicate pfinet linux headers that have glue versionsFlavio Cruz2022-11-0633-5044/+0
| | | | Message-Id: <Y2NNAegVOKg/X7OC@viriathus>
* libps: Fix memory leakRichard Braun2022-11-051-0/+1
| | | | Message-Id: <20221105141505.4684-1-rbraun@sceen.net>
* Fix parallel buildsSamuel Thibault2022-11-011-5/+17
| | | | | | | | | | | | | | | | | | | | As explained on https://www.gnu.org/software/automake/manual/html_node/Multiple-Outputs.html a b: c foo -o a -o b is actually equivalent to a: c foo -o a -o b b: c foo -o a -o b and can thus break in parallel builds, since the foo command may then be called twice. Also, we are here using pattern rules. The automake manual contains various levels of fixing this, using a mere stamp file should be enough for us.
* Add --without-rump configure optionSamuel Thibault2022-10-311-21/+26
| | | | To allow cross-builds which cannot use AC_LINK_IFELSE at all.
* Add --without-acpica configure optionSamuel Thibault2022-10-311-5/+9
| | | | To allow cross-builds which cannot use AC_CHECK_LIB at all.
* Also do not realpath "procfs" pseudo-deviceSamuel Thibault2022-10-241-1/+2
| | | | * sutils/fstab.c (fstab_find_device): Do not realpath "procfs" pseudo-device.
* rumpdisk: Link in PIIX IDE driver and ATA common codeSamuel Thibault2022-09-261-1/+1
|
* pfinet: Fix warningSamuel Thibault2022-09-221-4/+3
|
* lwip: Fix warningSamuel Thibault2022-09-221-2/+2
|
* iioctl: restore documenting 10 and 11 as unusedSamuel Thibault2022-09-221-5/+1
|
* rioctl: Use r not i group ioctl for SIOCADDRT/SIOCDELRTDamien Zammit2022-09-2210-16/+74
| | | | Message-Id: <20220922004105.961869-1-damien@zamaudio.com>
* lwip: Add missing stubDamien Zammit2022-09-221-0/+10
| | | | Message-Id: <20220922004018.961808-1-damien@zamaudio.com>
* Use route.h from glibcSamuel Thibault2022-09-213-43/+9
| | | | | | Unfortunately the Linux source also uses <net/route.h>, so we have to both make that include the glibc-provided one, and avoid making it emit Linuxish definitions when they would conflict with glibc definitions.
* procfs: Populate /proc/route with network routesDamien Zammit2022-09-202-1/+78
|
* acpi: Convert translator to an emulated mach deviceDamien Zammit2022-09-206-25/+207
| | | | | This makes acpi usable as a bootstrap translator. Message-Id: <20220920030035.931113-1-damien@zamaudio.com>
* acpi: Remove references to /dev/memDamien Zammit2022-09-121-75/+34
| | | | Message-Id: <20220912103837.556815-5-damien@zamaudio.com>
* shutdown: Use new acpi RPC to halt machine, clean upDamien Zammit2022-09-124-88/+14
| | | | | | | This allows clean shutdown of all modern x86 machines (not just qemu) by using the acpi translator to call into libacpica code. Message-Id: <20220912103837.556815-3-damien@zamaudio.com>
* acpi: Link translator to libacpica and provide RPCsDamien Zammit2022-09-1215-13/+233
| | | | | | | | | Provides two new acpi RPCs to sleep the machine and to get the irq of any pci device. ACPI mode is enabled by default when the translator is started. NB: Merging this commit means libacpica is a build dep. Message-Id: <20220912103837.556815-2-damien@zamaudio.com>
* machdev, pci-arbiter, rumpdisk: Fix race condition in bootstrapDamien Zammit2022-09-114-27/+29
| | | | | | | | This fixes a known race condition in bootstrapping by separating the fsys_startup call from the server demuxer loop into two separate functions that the caller can decide when to call. Message-Id: <20220908093229.499494-1-damien@zamaudio.com>
* pci-arbiter: Fix type of printf paramDamien Zammit2022-09-111-1/+1
| | | | Message-Id: <20220908093034.499415-1-damien@zamaudio.com>
* Add new RPC for pfinet network routesDamien Zammit2022-09-117-4/+223
| | | | Message-Id: <20220909094234.517165-1-damien@zamaudio.com>
* defs: Fix typing of iioctlDamien Zammit2022-08-311-1/+1
| | | | | | After much digging this fixes the new routing ioctls Message-Id: <20220831143032.401732-1-damien@zamaudio.com>
generated by cgit v1.2.3 (git 2.39.1) at 2025-04-14 18:51:08 +0000