ian (Ian Lepore)
User

Projects

User Details

User Since
May 11 2014, 1:02 PM (192 w, 4 d)

Recent Activity

Tue, Jan 16

ian committed rS328069: Remove redundant critical_enter/exit() calls. The block of code delimited.
Remove redundant critical_enter/exit() calls. The block of code delimited
Tue, Jan 16, 11:19 PM
ian committed rS328068: Move some code around and rename a couple variables; no functional changes..
Move some code around and rename a couple variables; no functional changes.
Tue, Jan 16, 11:14 PM
ian abandoned D13731: Convert the x86 RTC driver to use new validated BCD<->timespec conversions.

Committed as r328005

Tue, Jan 16, 4:57 AM
ian abandoned D13730: Add RTC clock conversions for BCD values, with non-panic validation..

Committed as r327971

Tue, Jan 16, 4:56 AM
ian committed rS328039: Add static inline rtcin_locked() and rtcout_locked() functions for doing a.
Add static inline rtcin_locked() and rtcout_locked() functions for doing a
Tue, Jan 16, 3:02 AM

Mon, Jan 15

ian committed rS328005: Convert the x86 RTC driver to use new validated BCD<->timespec conversions..
Convert the x86 RTC driver to use new validated BCD<->timespec conversions.
Mon, Jan 15, 4:41 PM
ian updated the test plan for D13731: Convert the x86 RTC driver to use new validated BCD<->timespec conversions.
Mon, Jan 15, 4:39 PM

Sun, Jan 14

ian committed rS327971: Add RTC clock conversions for BCD values, with non-panic validation..
Add RTC clock conversions for BCD values, with non-panic validation.
Sun, Jan 14, 5:01 PM

Sat, Jan 13

ian committed rS327909: Fix fallout from applying a patch twice..
Fix fallout from applying a patch twice.
Sat, Jan 13, 3:12 AM

Thu, Jan 11

ian committed rS327843: Add description files for WITH/WITHOUT_LINT. These should have been part.
Add description files for WITH/WITHOUT_LINT. These should have been part
Thu, Jan 11, 7:41 PM
ian committed rS327837: Do not build lint(1) by default on stable-11, add WITH_LINT to enable.
Do not build lint(1) by default on stable-11, add WITH_LINT to enable
Thu, Jan 11, 5:51 PM
ian closed D13799: Do not build lint(1) by default on stable-11, add WITH_LINT to enable building it..
Thu, Jan 11, 5:51 PM

Wed, Jan 10

ian updated the diff for D13731: Convert the x86 RTC driver to use new validated BCD<->timespec conversions.

Follow changes in D13730... the conversion routines now take a bool flag to indicate whether AM/PM mode conversions need to be done. atrtc always operates in 24-hour mode.

Wed, Jan 10, 4:21 PM
ian updated the diff for D13730: Add RTC clock conversions for BCD values, with non-panic validation..

Add support for handling AM/PM mode (12 vs 24 hour clock) to the conversion routines. Many RTC chips/drivers support an AM/PM mode, and it turns out to be difficult to do the math on BCD values in each individual driver. It's much easier to commonize the support and do the math while the numbers are in binary form during conversion.

Wed, Jan 10, 4:18 PM
ian committed rS327758: Convert a collection of unrelated bitwise flags to a collection of boolean.
Convert a collection of unrelated bitwise flags to a collection of boolean
Wed, Jan 10, 2:57 AM
ian committed rS327757: Bugfix: don't lose the am/pm mode flag when setting the time. Unlike most.
Bugfix: don't lose the am/pm mode flag when setting the time. Unlike most
Wed, Jan 10, 2:32 AM
ian committed rS327756: Bugfix: on RTC chips with a 32-bit binary counter, after reading the time,.
Bugfix: on RTC chips with a 32-bit binary counter, after reading the time,
Wed, Jan 10, 2:28 AM

Tue, Jan 9

ian added a comment to D13799: Do not build lint(1) by default on stable-11, add WITH_LINT to enable building it..

lint could be treated as a bootstrap tool to fix the build problem. What that provide, that this change doesn't, is a broken version of lint (it has been broken for 3 years) which can't really be used for anything other than making the build complete without errors.

Tue, Jan 9, 3:14 PM

Mon, Jan 8

ian created D13799: Do not build lint(1) by default on stable-11, add WITH_LINT to enable building it..
Mon, Jan 8, 7:32 PM
ian added inline comments to D13793: Introduce aw_syscon(4) for earlier attachment.
Mon, Jan 8, 3:49 PM

Sun, Jan 7

ian committed rS327679: Use EVENTHANDLER_DIRECT_INVOKE for [un]mount events, for better performance..
Use EVENTHANDLER_DIRECT_INVOKE for [un]mount events, for better performance.
Sun, Jan 7, 6:07 PM
ian committed rS327678: Use EVENTHANDLER_DIRECT_INVOKE() for device events, for better performance..
Use EVENTHANDLER_DIRECT_INVOKE() for device events, for better performance.
Sun, Jan 7, 6:06 PM
ian added a comment to D13131: Replace the incorrect C arm fabs(3) code with an asm version..

As usual, the clang/llvm documentation ranges from murky to non-existant on this stuff. I did track down in the gcc docs that you can always explicitly call __builtin_foo() even when -fno-builtin and -ffreestanding are in effect (such as when we build libstand and loader(8)). Since clang tries to be gcc compatible, I guess we have to assume it will behave the same.

Sun, Jan 7, 3:20 PM
ian committed rS327668: MFC r327220-r327221.
MFC r327220-r327221
Sun, Jan 7, 2:59 AM
ian committed rS327667: MFC r327222:.
MFC r327222:
Sun, Jan 7, 2:57 AM
ian committed rS327664: MFC r327226, r327356.
MFC r327226, r327356
Sun, Jan 7, 1:08 AM
ian committed rS327663: MFC r327439:.
MFC r327439:
Sun, Jan 7, 1:06 AM
ian committed rS327658: MFC r327367:.
MFC r327367:
Sun, Jan 7, 12:04 AM

Sat, Jan 6

ian committed rS327656: MFC r327048-r327050.
MFC r327048-r327050
Sat, Jan 6, 11:25 PM
ian committed rS327655: MFC r327032:.
MFC r327032:
Sat, Jan 6, 11:21 PM
ian committed rS327654: MFC r326924-r326925.
MFC r326924-r326925
Sat, Jan 6, 11:12 PM
ian committed rS327653: MFC r326750:.
MFC r326750:
Sat, Jan 6, 11:04 PM
ian committed rS327651: MFC r325108:.
MFC r325108:
Sat, Jan 6, 10:57 PM
ian added a comment to D13131: Replace the incorrect C arm fabs(3) code with an asm version..

I don't think we can rely on VFP hardware being available. It is at least theoretically possible to build a kernel using "nooptions VFP" and build all of userland with --float-abi=soft and any use of a vfp instruction will be a SIGILL, right?

Sat, Jan 6, 9:31 PM
ian committed rS327638: MFC r325045, r325054-r325056, r325061, r325063, r325065.
MFC r325045, r325054-r325056, r325061, r325063, r325065
Sat, Jan 6, 9:20 PM
ian committed rS327637: MFC r325060:.
MFC r325060:
Sat, Jan 6, 8:52 PM
ian committed rS327633: Fix imx6 hdmi init after r323553..
Fix imx6 hdmi init after r323553.
Sat, Jan 6, 7:25 PM
ian committed rS327626: MFC r324413, r324415.
MFC r324413, r324415
Sat, Jan 6, 5:21 PM

Fri, Jan 5

ian committed rS327595: MFC r324185:.
MFC r324185:
Fri, Jan 5, 7:28 PM
ian committed rS327594: MFC r324169:.
MFC r324169:
Fri, Jan 5, 7:21 PM
ian committed rS327592: MFC r319987, r324107-r324108.
MFC r319987, r324107-r324108
Fri, Jan 5, 7:13 PM

Mon, Jan 1

ian added a dependency for D13731: Convert the x86 RTC driver to use new validated BCD<->timespec conversions: D13730: Add RTC clock conversions for BCD values, with non-panic validation..
Mon, Jan 1, 11:53 PM
ian added a dependent revision for D13730: Add RTC clock conversions for BCD values, with non-panic validation.: D13731: Convert the x86 RTC driver to use new validated BCD<->timespec conversions.
Mon, Jan 1, 11:53 PM
ian created D13731: Convert the x86 RTC driver to use new validated BCD<->timespec conversions.
Mon, Jan 1, 11:52 PM
ian created D13730: Add RTC clock conversions for BCD values, with non-panic validation..
Mon, Jan 1, 11:45 PM

Sun, Dec 31

ian committed rS327453: Add a validbcd() routine that uses the bcd2bin_data[] array and returns a.
Add a validbcd() routine that uses the bcd2bin_data[] array and returns a
Sun, Dec 31, 10:43 PM
ian committed rS327446: Chase r327432... sparc64 clock.c now needs to include sys/tslog.h.
Chase r327432... sparc64 clock.c now needs to include sys/tslog.h
Sun, Dec 31, 8:31 PM
ian committed rS327439: Allow use of .WAIT in the LOCAL_DIRS and LOCAL_LIB_DIRS lists..
Allow use of .WAIT in the LOCAL_DIRS and LOCAL_LIB_DIRS lists.
Sun, Dec 31, 6:53 PM
ian closed D13622: Allow use of .WAIT in the LOCAL_DIRS and LOCAL_LIB_DIRS lists..
Sun, Dec 31, 6:53 PM

Sat, Dec 30

ian committed rS327367: Make kernel option KERNVIRTADDR optional, remove it from std.<platform>.
Make kernel option KERNVIRTADDR optional, remove it from std.<platform>
Sat, Dec 30, 12:21 AM

Fri, Dec 29

ian committed rS327356: Correct a mistake and reword a couple sentences to clarify that "the value".
Correct a mistake and reword a couple sentences to clarify that "the value"
Fri, Dec 29, 8:00 PM

Tue, Dec 26

ian committed rS327226: Add a section describing how to tune ARM kernel options to use an MD_ROOT.
Add a section describing how to tune ARM kernel options to use an MD_ROOT
Tue, Dec 26, 10:06 PM
ian committed rS327222: Add a new ARM kernel option, LOCORE_MAP_MB, to control the size of the.
Add a new ARM kernel option, LOCORE_MAP_MB, to control the size of the
Tue, Dec 26, 7:03 PM
ian committed rS327221: Complete the changing of the old "i2c-address" property to the modern "reg".
Complete the changing of the old "i2c-address" property to the modern "reg"
Tue, Dec 26, 6:42 PM
ian committed rS327220: Update the FDT example for the lm75 sensor to match current devicetree.
Update the FDT example for the lm75 sensor to match current devicetree
Tue, Dec 26, 6:34 PM
ian accepted D13625: sys/arm/lpc: Remove support for it.
Tue, Dec 26, 5:16 PM
ian requested changes to D13625: sys/arm/lpc: Remove support for it.

I'm fine with the removal, I've never heard from a single person that used this platform. The only mention of it on the arm@ list was a request in 2010, before we actually imported the current lpc code.

Tue, Dec 26, 5:08 PM

Mon, Dec 25

ian created D13622: Allow use of .WAIT in the LOCAL_DIRS and LOCAL_LIB_DIRS lists..
Mon, Dec 25, 7:29 PM

Wed, Dec 20

ian committed rS327050: If a temporary mapping is made to support EARLY_PRINTF, undo that mapping.
If a temporary mapping is made to support EARLY_PRINTF, undo that mapping
Wed, Dec 20, 10:19 PM
ian committed rS327049: Allow pmap_kremove() to remove 1MB section mappings as well as 4K pages..
Allow pmap_kremove() to remove 1MB section mappings as well as 4K pages.
Wed, Dec 20, 10:17 PM
ian committed rS327048: Restore the ability to use EARLY_PRINTF support during most of initarm()..
Restore the ability to use EARLY_PRINTF support during most of initarm().
Wed, Dec 20, 8:46 PM
ian committed rS327047: Remove arm-specific implementations of atomic_load/store_xxx() now that.
Remove arm-specific implementations of atomic_load/store_xxx() now that
Wed, Dec 20, 8:42 PM
ian committed rS327032: Add a new kernel config option, MD_ROOT_READONLY, which forces on the.
Add a new kernel config option, MD_ROOT_READONLY, which forces on the
Wed, Dec 20, 6:23 PM
ian closed D13411: Add MD_ROOT_READONLY option to disallow writeable remount of root fs.
Wed, Dec 20, 6:23 PM

Dec 18 2017

ian committed rS326934: MFC (conceptually) r326752, r326754:.
MFC (conceptually) r326752, r326754:
Dec 18 2017, 5:17 PM
ian committed rS326925: Do not attempt to refill the TX fifo if there is no data left to transfer..
Do not attempt to refill the TX fifo if there is no data left to transfer.
Dec 18 2017, 2:34 AM
ian committed rS326924: Fix debugging output, fallout from something like s/read/readctl/g.
Fix debugging output, fallout from something like s/read/readctl/g
Dec 18 2017, 12:16 AM

Dec 10 2017

ian committed rS326754: When building for arm arches, set PKGALIGN to the max cache line size.
When building for arm arches, set PKGALIGN to the max cache line size
Dec 10 2017, 11:07 PM
ian committed rS326752: Save and restore r9 register in arm ubldr. In old gcc 4.2, r9 was a callee-.
Save and restore r9 register in arm ubldr. In old gcc 4.2, r9 was a callee-
Dec 10 2017, 9:51 PM
ian committed rS326750: Do not give up if writing to the chip's control and status registers fails.
Do not give up if writing to the chip's control and status registers fails
Dec 10 2017, 6:56 PM

Dec 7 2017

ian created D13411: Add MD_ROOT_READONLY option to disallow writeable remount of root fs.
Dec 7 2017, 6:21 PM

Nov 26 2017

ian requested changes to D13244: Remove fortune(6).

People have existing references/calls to fortune in their shell startup scripts. Removing the program could potentially leave people unable to login remotely in a worst case. I don't believe removing the binary is a viable option.

Nov 26 2017, 5:09 PM

Nov 7 2017

ian accepted D12961: DS1307: Add the mcp7941x enable bit.

Looks good.

Nov 7 2017, 4:47 PM

Nov 5 2017

ian requested changes to D12961: DS1307: Add the mcp7941x enable bit.
Nov 5 2017, 7:53 PM

Oct 30 2017

ian abandoned D12821: Allow faster eventhandler dispatching by keeping pointers to handler lists..

Abandoned in favor of the nearly-identical D12814.

Oct 30 2017, 4:00 PM

Oct 29 2017

ian committed rS325108: Improve the performance of the hpet timer in bhyve guests by making the.
Improve the performance of the hpet timer in bhyve guests by making the
Oct 29 2017, 8:50 PM
ian updated the diff for D12821: Allow faster eventhandler dispatching by keeping pointers to handler lists..

Update to remove unrelated changes (moving lines from eventhandler.h to proc_events.h, and rearranging the order of some macros in eventhandler.h).

Oct 29 2017, 7:42 PM
ian added a comment to D12821: Allow faster eventhandler dispatching by keeping pointers to handler lists..
In D12821#266158, @kib wrote:

Why do we need to keep the compatibility and to create the _LIST KPI in addition to the existing KPI ?

As I understand, the goal of the change is to remove the lookup of the eventhandler list by name. What if you make the 'slow' handlers require the EVENTHANDLER_DEFINE() presence in the source code and then use the eventhandler_tag as an argument for _INVOKE and pehaps even to _REGISTER/_UNREGISTER. This probably would also require some module magic. In other words, the eventhandler would become very similar to the MALLOC_DECLARE/MALLOC_DEFINE.

I understand that this is larger rototiling of the KPI, perhaps in the backward-incompatible way, but I do not see a reason to keep the string-identified handlers machinery around, if it is not useful. Is there a handler kind which cannot be easily converted into the stringless model ?

Oct 29 2017, 3:17 PM
ian created D12821: Allow faster eventhandler dispatching by keeping pointers to handler lists..
Oct 29 2017, 2:15 AM

Oct 28 2017

ian committed rS325065: Split the hardware type enum and the hw feature flags bits into separate.
Split the hardware type enum and the hw feature flags bits into separate
Oct 28 2017, 8:03 PM
ian committed rS325063: Use the 16-bit receive shift feature in ffec hardware that supports it..
Use the 16-bit receive shift feature in ffec hardware that supports it.
Oct 28 2017, 7:34 PM
ian committed rS325061: Support up to 3 IRQs in the ffec driver..
Support up to 3 IRQs in the ffec driver.
Oct 28 2017, 7:08 PM
ian committed rS325060: Add a #define RESOURCE_SPEC_END. Similar to DEVMETHOD_END and.
Add a #define RESOURCE_SPEC_END. Similar to DEVMETHOD_END and
Oct 28 2017, 6:56 PM
ian committed rS325056: Avoid AXI bus issues due to a MAC reset on imx6sx and imx7..
Avoid AXI bus issues due to a MAC reset on imx6sx and imx7.
Oct 28 2017, 5:31 PM
ian committed rS325055: Add FECFLAG_AVB variant flag to support new features on imx7..
Add FECFLAG_AVB variant flag to support new features on imx7.
Oct 28 2017, 5:06 PM
ian committed rS325054: Increase the alignment of the rx/tx descriptor ring buffers to 64 bytes..
Increase the alignment of the rx/tx descriptor ring buffers to 64 bytes.
Oct 28 2017, 4:51 PM

Oct 27 2017

ian committed rS325045: Actually release resources in detach() rather than just returning EBUSY..
Actually release resources in detach() rather than just returning EBUSY.
Oct 27 2017, 5:22 PM

Oct 26 2017

ian committed rS325023: MFC r314914:.
MFC r314914:
Oct 26 2017, 10:19 PM

Oct 23 2017

ian added a comment to D8554: Verified execution (veriexec) as a MAC module..

Note the comments I've added here are a couple simple things I noticed when skimming the code to get familiar with it. I.e., this was not an in-depth review with any understanding of the code.

Oct 23 2017, 11:57 PM

Oct 21 2017

ian added a reviewer for D12736: i2c support for mediatek soc: ian.
Oct 21 2017, 5:31 PM
ian added a comment to D12736: i2c support for mediatek soc.

IMO, this entire driver is wrong. Because of the way the hardware works, it cannot support the start, repeat_start, stop, read, and write methods for iicbus_if, and it must not try to "fake" that support. It must implement only callback, reset, and transfer. The transfer method should be basically just

Oct 21 2017, 5:31 PM

Oct 19 2017

ian committed rS324760: MFC r323997-r323998.
MFC r323997-r323998
Oct 19 2017, 5:26 PM
ian committed rS324756: MFC r323985:.
MFC r323985:
Oct 19 2017, 4:16 PM
ian committed rS324755: MFC r323392:.
MFC r323392:
Oct 19 2017, 4:08 PM

Oct 8 2017

ian committed rS324416: Fix imx6 hdmi init after r323553, which used a config_intrhook to defer the.
Fix imx6 hdmi init after r323553, which used a config_intrhook to defer the
Oct 8 2017, 6:38 PM
ian committed rS324415: Add eventhandler notifications for newbus device attach/detach..
Add eventhandler notifications for newbus device attach/detach.
Oct 8 2017, 5:34 PM
ian closed D12557: Add eventhandler notifications for newbus device attach/detach..
Oct 8 2017, 5:34 PM
ian committed rS324413: Restore the ability to deregister an eventhandler from within the callback..
Restore the ability to deregister an eventhandler from within the callback.
Oct 8 2017, 5:21 PM
ian closed D12561: Restore the ability to deregister an eventhandler callback from within the callback..
Oct 8 2017, 5:21 PM

Oct 2 2017

ian committed rS324186: Define a single instance of ahci_devclass and reference it from all the.
Define a single instance of ahci_devclass and reference it from all the
Oct 2 2017, 2:58 AM