Include <sys/types.h> explicitly instead of depending on that
Include <stdarg.h> instead of <machine/stdarg.h> when compiled as
Detect clang on macOS. The version string is slightly different.
Include "util.h", not <util.h>. The header is in the same directory
When compiling on macOS or Linux, __dead can be defined already.
I've been thinking about encoding this with the partition specification (i.e. as part of the -p option), but the specification is cryptic enough as it is. Maybe when there's a different way of specifying partitions (say, using a JSON or YAML file given instead of a number of -p options), will it be easier to add a "bootme" attribute that translates into setting the active flag. For now, this is simple and it works...
o Provide a private definition for UUIDs (mkimg_uuid_t) because
Add LORDER, TSORT and TSORTFLAGS variables and replace the
Switch to using the portable partition scheme headers.
Re-apply change 306811 or alternatively, revert change 307385.
Revert change 306811 so that the change can be re-done using
In order to allow mkimg(1) (and other tools) to become a build tool
Fixes to the previous diff:
o Have the generic GPT uuid definition match the UUID constants so that the header is self-consistent and usable without "external" definition for UUIDs.
o Typedef gpt_uuid_t from GPT_UUID_TYPE to make it more palatable for applications to use the GPT type for UUIDs themselves.
o Fix a comment in sys/sys/disklabel.h
Replace OFF_MAX with INT64_MAX. The former is defined on Linux.
Prefer <stdint.h> over <sys/types.h>. While here remove redundant
Replace STAILQ with TAILQ. TAILQs are portable enough that they can
marcel retitled
D8115: Move disk scheme definitions under <sys/disk/$scheme.h> from
to
Move disk scheme definitions under <sys/disk/$scheme.h>.
Avoid depending on the <sys/endian.h> header for le*enc and be*enc.
Eliminate the use of EDOOFUS. The error code was used to signal
Replace the use of linker sets with constructors for both the
Document the ".pico" extension for object files.
Relocatable object files are renamed from *.So to *.pico
When MAKEOBJDIRPREFIX points to a case-insensitive file system, the
Update local variable 'block' after calling capacity_resize(),
When MAKEOBJDIRPREFIX points to a case-insensitive file system, the
Author mentioned SVN revision with fix.
I can document .po, .pico, etc but prefer to do so in a separate commit. The need for such documentation isn't created by this change, but has always been there. I'll do a followup commit (plus email to , etc) after this change is approved.
marcel retitled
D7906: Switch from .So to .pico for object files compiled with -fpic from
Which from .So to .pico for object files compiled with -fpic to
Switch from .So to .pico for object files compiled with -fpic.
marcel retitled
D7906: Switch from .So to .pico for object files compiled with -fpic from
to
Which from .So to .pico for object files compiled with -fpic.
When MAKEOBJDIRPREFIX points to a case-insensitive file system, the
Adding my response to bdrewery@ for posterity:
marcel retitled
D7893: Fix lib/libc for object tree on case-insensitive file system. from
to
Fix lib/libc for object tree on case-insensitive file system..
Probably good to get feedback from people on arm@ or re@.
I'll approve in case there has been silence so far...
If mkimg still returns 0 when it doesn't like the command line options, then we should definitely fix that. You want scripts to be able to check for (or just fail) in that case and not create a broken image.
Support version 4 of the userboot structure by implementing the
Bump VM_MAX_MEMSEGS from 2 to 3 to match the number of VM segment
Add option -C to have the guest memory included in core files.
Add vm_set_register() and vm_set_desc() callbacks. These callbacks
Check that the userboot version is at least 3, rather than
I double-checked the code and there's actually no binary search logic. It's all linear. Technically speaking, the reordering was not necessary, but I really appreciate the effort. It's a lot easier now to implement a binary search for cases where the linear search gets too time consuming.
IIRC, aliases must be sorted so that the binary search works. ms-recovery before ms-reserved, please.
I see we disordered ntfs before netbsd-* as well, so we have a pre-existing bug if the binary search still holds.
If we can't open the file, skip devclose() for the exclusive_file_system
Add option -l for specifying which OS loader to dlopen(3). By default
I would not rework the patch. While mav@'s comments are good, they go beyond the simple problem that the hold solved. The hold only solved the holding off of mounting root while interrupt-driven discovery was still ongoing. You patch, trasz@, avoids that we wait mounting root when there was no point waiting, as the device was already there. That is a good change.
As expected, things aren't as simple as hoped. Consequently, we have
Auto-detect the UGA frame buffer and stride on a MacBook. We're
My MacBook has UGA only, but we fail to detect any changes
For UGA, the frame buffer address obtained by scanning the
Add support for the UGA draw protocol. This includes adding a
Add a gop command to help diagnose VT efifb problems. The gop
Overall: I like the change. You may want to get a second thumbs up though.
Looking at the ifconfig in 4.4 lite, it looks like the netmask has always been printed as a hex number and I've always found that to be less than ideal. Having options for how the subnet is displayed is a good suggestion. I like the CIDR notion for example.
An error of -1 from parse_mount() indicates that the specification
MFC r286808, r286809, r286867, r286868
MFC r286660, r286419, r286417, r286395, r286215, r284883
Fix build for architectures that define wchar_t as an unsigned int.
o Add vdsk_trim() to support BOP_DELETE.
Add a convenience wrapper around vmrun.sh
Add 24 more page table pages we allocate on boot-up. 16MB slop
Support frame buffers that are larger than the default screen
Improve support for Macs that have a stride not equal to the
Improve the VT initialization message: have it say what the
s/as/at/ in previous commit.
Change md(4) to use weak symbols as start, end and size for the embedded
Instead of having separate do_sync functions for ARM_ARCH 6 vs.
The Broadcom BCM56060 chip has a Cortex-A9R4 core.
Add support for the Broadcom TruManage integrated serial port.
Better support memory mapped console devices, such as VGA and EFI
Use bus_alloc_resource_any(), rather than bus_alloc_resource()
Document the application interface.
Fix typo introduced in previous commit.
Rebase after r286395: rounding fix for dynamic VHD