Page MenuHomeFreeBSD

royger (Roger Pau Monné)
User

Projects

User Details

User Since
May 18 2014, 2:32 AM (336 w, 1 d)

Recent Activity

Aug 27 2020

royger committed rP546327: emulators/xen-kernel & sysutils/xen-tools: update to 4.14.0.
emulators/xen-kernel & sysutils/xen-tools: update to 4.14.0
Aug 27 2020, 7:59 AM

Aug 26 2020

royger committed rP546292: misc/seabios: update to 1.14.0.
misc/seabios: update to 1.14.0
Aug 26 2020, 4:10 PM

Aug 21 2020

royger added a comment to D6611: net: translate inbound checksum offloading flags to outbound when forwarding.
In D6611#580267, @jhb wrote:

I think one question is how can you really identify these packets? You can send a packet into a VM (e.g. via an if_tap interface from a localhost connection) but that VM might itself forward the packet internally to another interface that then heads out onto the wire. You can't really know that before you send the packet into the if_tap, and the VM itself doesn't know which interfaces are internal and which are not.

I considered & rejected some magic solution where we constrain virtual interface mac addresses to a range that is unique and which can be easily checked. Then we'd know if a packet came in on a virtual interface & needs checksumming. But that doesn't work if the VM has pass-thru access to a physical NIC and can send a packet w/o going through the host.

Aug 21 2020, 1:25 PM

Aug 18 2020

royger added a comment to D6656: net/tso: use the most restrictive segment size to perform the calculations.

I'm confused. t_maxsegsize is basically the connection's MSS, eg, what the NIC is supposed to break the TSO into. However if_hw_tsomaxsegsize is the max size of the TSO to be broken up. Eg, on a NIC without oddball limitations, if_hw_tsomaxsegcount might be 64, if_hw_tsomaxsegsize might be 65535. That means that TSO is supposed to send a max of 64K down to the NIC, in a max of 64 packets. If t_maxseg is less than 1000, then we'd have to limit the TSO to 64 packets.

It seems like by restricting if_hw_tsomaxsegsize to t_maxseg, you're essentially turning off TSO, since you're only allowing one MSS to be sent at a time.

Ugh, I'm wrong. I was confusing t_tsomax with t_tsomaxsegsize. I forgot that some NICs are apparently so broken that they limit TSO seg sizes.

Aug 18 2020, 12:49 PM
royger added a comment to D6611: net: translate inbound checksum offloading flags to outbound when forwarding.

This will break some NICs (see comments on D6656 -- basically, if you're setting CSUM_TCP then you need to replace the TCP checksum with the pseudohdr checksum, and store the offset of the TCP header's checksum in csum_data).

The important question is: If the packet has already had its checksum checked, then the checksum is correct. Why not update the checksum normally and send the packet without checksum offload enabled?

Aug 18 2020, 11:22 AM

Aug 13 2020

royger added a comment to D6611: net: translate inbound checksum offloading flags to outbound when forwarding.

Is there any concern or additional work needed for ipV6?

Aug 13 2020, 2:40 PM

Jun 23 2020

royger closed D24482: sysutils/xen-tools: fix build with clang and lld 10.
Jun 23 2020, 3:16 AM

Jun 18 2020

royger committed rS362331: MFC r352925: xen/ctrl: acknowledge all control requests.
MFC r352925: xen/ctrl: acknowledge all control requests
Jun 18 2020, 3:45 PM
royger committed rS362329: MFC r352925: xen/ctrl: acknowledge all control requests.
MFC r352925: xen/ctrl: acknowledge all control requests
Jun 18 2020, 3:15 PM

May 28 2020

royger committed rS361580: xen/control: short circuit xctrl_on_watch_event on spurious event.
xen/control: short circuit xctrl_on_watch_event on spurious event
May 28 2020, 8:20 AM
royger committed rS361579: xen/blkfront: use the correct type for disk sectors.
xen/blkfront: use the correct type for disk sectors
May 28 2020, 8:19 AM
royger committed rS361578: xenpv: do not use low 1MB for Xen mappings on i386.
xenpv: do not use low 1MB for Xen mappings on i386
May 28 2020, 8:18 AM

May 26 2020

royger committed rS361500: xen-locore: fix size in GDT descriptor.
xen-locore: fix size in GDT descriptor
May 26 2020, 10:24 AM

May 25 2020

royger committed rP536484: misc/seabios: update to 1.13.0.
misc/seabios: update to 1.13.0
May 25 2020, 5:21 PM
royger committed rP536472: sysutils/xen-tools + emulators/xen-kernel: update to 4.13.1.
sysutils/xen-tools + emulators/xen-kernel: update to 4.13.1
May 25 2020, 2:38 PM

May 20 2020

royger committed rS361274: dev/xenstore: fix return with locks held.
dev/xenstore: fix return with locks held
May 20 2020, 11:01 AM

May 13 2020

royger committed rP535106: sysutils/xen-tools: properly fix Wno-misleading-indentation errors.
sysutils/xen-tools: properly fix Wno-misleading-indentation errors
May 13 2020, 12:25 PM

May 4 2020

royger committed rP533907: emulators/xen-kernel: improve package message.
emulators/xen-kernel: improve package message
May 4 2020, 8:43 AM

Apr 30 2020

royger committed rP533418: sysutils/xen-tools: fix build with clang and lld 10.
sysutils/xen-tools: fix build with clang and lld 10
Apr 30 2020, 8:37 AM

Apr 27 2020

royger added a comment to D24482: sysutils/xen-tools: fix build with clang and lld 10.

Ping? I would like to get this committed so xen-tools package builds again on HEAD. Thanks!

Apr 27 2020, 3:39 PM

Apr 18 2020

royger retitled D24482: sysutils/xen-tools: fix build with clang and lld 10 from sysutils/xen-tools: fix build with lld 10 to sysutils/xen-tools: fix build with clang and lld 10.
Apr 18 2020, 10:22 AM
royger updated the diff for D24482: sysutils/xen-tools: fix build with clang and lld 10.

Also fix clang 10 build, by adding -Wno-misleading-indentation to
tools build

Apr 18 2020, 10:21 AM
royger created D24482: sysutils/xen-tools: fix build with clang and lld 10.
Apr 18 2020, 9:52 AM

Apr 17 2020

royger accepted D24384: xen-locore: Silence DWARF2 section warning.

Thanks!

Apr 17 2020, 6:10 PM

Mar 23 2020

royger closed D24160: sysutils/xen-tools: fix makefile typo.
Mar 23 2020, 5:40 PM
royger committed rP528995: sysutils/xen-tools: fix makefile typo.
sysutils/xen-tools: fix makefile typo
Mar 23 2020, 5:40 PM
royger planned changes to D24160: sysutils/xen-tools: fix makefile typo.
Mar 23 2020, 5:37 PM
royger updated the diff for D24160: sysutils/xen-tools: fix makefile typo.

bump port revision

Mar 23 2020, 12:03 PM
royger created D24160: sysutils/xen-tools: fix makefile typo.
Mar 23 2020, 11:02 AM

Mar 20 2020

royger committed rP528771: sysutils/xen-tools: fix build with clang 10.0.
sysutils/xen-tools: fix build with clang 10.0
Mar 20 2020, 9:55 AM
royger closed D24097: sysutils/xen-tools: fix build with clang 10.0.
Mar 20 2020, 9:55 AM

Mar 17 2020

royger retitled D24097: sysutils/xen-tools: fix build with clang 10.0 from xen-tools: fix build with clang 10.0 Add upstream fix for clang 10.0 build. to xen-tools: fix build with clang 10.0.
Mar 17 2020, 5:47 PM
royger created D24097: sysutils/xen-tools: fix build with clang 10.0.
Mar 17 2020, 2:16 PM

Feb 24 2020

royger accepted D23638: Mark more nodes as CTLFLAG_MPSAFE or CTLFLAG_NEEDGIANT (M of N).
Feb 24 2020, 6:34 PM

Feb 17 2020

royger committed rP526355: xen-{kernel/tools}: update to 4.13.
xen-{kernel/tools}: update to 4.13
Feb 17 2020, 9:18 AM
royger closed D23170: xen: update to 4.13.0.
Feb 17 2020, 9:18 AM

Feb 14 2020

royger added a comment to D23638: Mark more nodes as CTLFLAG_MPSAFE or CTLFLAG_NEEDGIANT (M of N).

Thanks, I have a couple of comments about the need for GIANT on some of the sysctls. I have to admit I'm not very familiar with sysctls, so I might be missing something.

Feb 14 2020, 11:56 AM

Feb 6 2020

royger committed rS357616: xen/console: fix priority of Xen console.
xen/console: fix priority of Xen console
Feb 6 2020, 2:03 PM

Jan 29 2020

royger committed rS357251: MFC r354637:.
MFC r354637:
Jan 29 2020, 9:34 AM

Jan 14 2020

royger updated the diff for D23170: xen: update to 4.13.0.

Moved iconv dependency from LIB_DEPENDS to USES.

Jan 14 2020, 2:08 PM
royger created D23170: xen: update to 4.13.0.
Jan 14 2020, 12:01 PM

Dec 17 2019

royger created D22851: release: enable serial console by default for VM images.
Dec 17 2019, 11:11 AM
royger created D22850: release: allow using DESTDIR with ftp target.
Dec 17 2019, 11:10 AM

Nov 12 2019

royger committed rS354637: xen: fix dispatching of NMIs.
xen: fix dispatching of NMIs
Nov 12 2019, 10:31 AM

Oct 1 2019

royger accepted D21834: Fix nondeterministic load of fdt_slicer.

LGTM, it's a shame we don't have a more declarative way to express such dependencies.

Oct 1 2019, 9:13 AM
royger committed rS352925: xen/ctrl: acknowledge all control requests.
xen/ctrl: acknowledge all control requests
Oct 1 2019, 8:21 AM

Aug 13 2019

royger committed rP508824: xen-{tools/kernel}: update to 4.12.1.
xen-{tools/kernel}: update to 4.12.1
Aug 13 2019, 1:40 PM
royger closed D21227: xen-{tools/kernel}: update to 4.12.1.
Aug 13 2019, 1:40 PM

Aug 12 2019

royger created D21227: xen-{tools/kernel}: update to 4.12.1.
Aug 12 2019, 2:14 PM

Jul 29 2019

royger closed D21028: emulators/xen-kernel: fix foreign mappings.
Jul 29 2019, 2:59 PM
royger committed rP507571: emulators/xen-kernel: fix foreign mappings.
emulators/xen-kernel: fix foreign mappings
Jul 29 2019, 2:59 PM
royger added a comment to D21028: emulators/xen-kernel: fix foreign mappings.

tyvm!

Jul 29 2019, 2:58 PM
royger added a comment to D21028: emulators/xen-kernel: fix foreign mappings.

Hello, can I please get this approved?

Jul 29 2019, 10:44 AM

Jul 26 2019

royger added a comment to D21028: emulators/xen-kernel: fix foreign mappings.

Ping? This is an almost trivial change, since it's a straight backport from an upstream commit.

Jul 26 2019, 8:13 AM

Jul 22 2019

royger created D21028: emulators/xen-kernel: fix foreign mappings.
Jul 22 2019, 4:25 PM

Jul 3 2019

royger committed rP505754: xen-{kernel/tools}: switch back to LLVM toolchain.
xen-{kernel/tools}: switch back to LLVM toolchain
Jul 3 2019, 2:25 PM
royger closed D20816: xen-{kernel/tools}: switch back to LLVM toolchain.
Jul 3 2019, 2:25 PM

Jul 1 2019

royger retitled D20816: xen-{kernel/tools}: switch back to LLVM toolchain from xen-{kernel/tools}: switch back to LLVM toolchain Backport upstream fixes to correctly build Xen using LLVM 8 toolchain. Note the fixes are not strictly needed for xen-tools since the Xen shim build is only ever supposed to be booted using the... to xen-{kernel/tools}: switch back to LLVM toolchain.
Jul 1 2019, 9:58 AM
royger created D20816: xen-{kernel/tools}: switch back to LLVM toolchain.
Jul 1 2019, 9:51 AM

Jun 28 2019

royger added a comment to D20770: arm64: Move "segments" array from bus_dma_tag to bus_dmamap..

Interestingly enough on arm and powerpc that array is stored as a part of dma map structure, so we might run into that anyway.
Anyway I agree with you that we need to have the same behavior across architectures.

Jun 28 2019, 2:08 PM
royger added a comment to D20770: arm64: Move "segments" array from bus_dma_tag to bus_dmamap..

Can you be more specific about the flow that can lead to overwrite the segs array?

Basically if you have a lot of threads doing DMA loads on the same tag concurrently there is a small chance of the segs array being overwritten before it is read in the callback.
It is a rare race, but reproducible.

Hm, I'm not sure I see how that's supposed to happen. busdma_swi already takes a lock (dmat->common.lockfunc) to prevent issuing multiple loads on the same tag, thus there should be no overwriting while the caller callback is using tag->segs?

Well, I'm not sure how dusdma_swi can help us in this case.
If you go through the flow starting at bus_dmamap_load, then going into bounce_bus_dmamap_load_buffer no lock is acquired there.

Right, the lock is only used for the deferred callback.

I think this is fine, but should be done for x86 also at least (that's where arm64 busdma code is coming from). This expected behaviour should also be mentioned on the busdma man page.

I think the general expectation of busdma code is that you serialize loads when using the same tag, but I cannot find an explicit reference to it on the man page, and IMO the locking section is too vague and doesn't explicitly say that you are not supposed to issue concurrent load calls using the same tag.

Ok, so as you said we have two ways to proceed from here:

  1. Modify the manual to explicitly require the user to serialize all loads on the same tag.

Do we know if all existing drivers follow that rule?

  1. Move the segmets array on all architectures.

I could prepare the patch, but frankly I don't feel competent enough to make such design decision.

Jun 28 2019, 1:31 PM

Jun 27 2019

royger added a comment to D20770: arm64: Move "segments" array from bus_dma_tag to bus_dmamap..

Can you be more specific about the flow that can lead to overwrite the segs array?

Basically if you have a lot of threads doing DMA loads on the same tag concurrently there is a small chance of the segs array being overwritten before it is read in the callback.
It is a rare race, but reproducible.

Hm, I'm not sure I see how that's supposed to happen. busdma_swi already takes a lock (dmat->common.lockfunc) to prevent issuing multiple loads on the same tag, thus there should be no overwriting while the caller callback is using tag->segs?

Well, I'm not sure how dusdma_swi can help us in this case.
If you go through the flow starting at bus_dmamap_load, then going into bounce_bus_dmamap_load_buffer no lock is acquired there.

Jun 27 2019, 2:55 PM

Jun 26 2019

royger added a comment to D20770: arm64: Move "segments" array from bus_dma_tag to bus_dmamap..

Can you be more specific about the flow that can lead to overwrite the segs array?

Basically if you have a lot of threads doing DMA loads on the same tag concurrently there is a small chance of the segs array being overwritten before it is read in the callback.
It is a rare race, but reproducible.

Jun 26 2019, 3:37 PM
royger added a comment to D20770: arm64: Move "segments" array from bus_dma_tag to bus_dmamap..

I'm not sure I understand the motivation behind that change, is this fixing an existing issue with a specific driver?

Jun 26 2019, 2:43 PM

Jun 20 2019

royger committed rP504632: emulators/xen-kernel: use binutils to workaround LLD 8 bug.
emulators/xen-kernel: use binutils to workaround LLD 8 bug
Jun 20 2019, 3:25 PM
royger closed D20706: emulators/xen-kernel: use binutils to workaround LLD 8 bug.
Jun 20 2019, 3:25 PM
royger created D20706: emulators/xen-kernel: use binutils to workaround LLD 8 bug.
Jun 20 2019, 9:26 AM

May 15 2019

royger committed rP501706: xen: add fixes for XSA-297 (MDS).
xen: add fixes for XSA-297 (MDS)
May 15 2019, 10:25 AM
royger closed D20265: xen: add fixes for XSA-297 (MDS).
May 15 2019, 10:25 AM
royger added a comment to D20265: xen: add fixes for XSA-297 (MDS).

Fixed in the new version, thanks!

May 15 2019, 9:53 AM
royger updated the diff for D20265: xen: add fixes for XSA-297 (MDS).

s/FILESDIR/PATCHDIR/

May 15 2019, 9:52 AM
royger created D20265: xen: add fixes for XSA-297 (MDS).
May 15 2019, 8:14 AM

May 6 2019

royger committed rS347183: geom: fix initialization order.
geom: fix initialization order
May 6 2019, 9:48 AM
royger closed D20148: geom: fix initialization order.
May 6 2019, 9:48 AM

May 3 2019

royger added a comment to D20148: geom: fix initialization order.

I suspect you'll need to bump some g_raid parts as well- it has modules declared at SI_SUB_DRIVERS/SI_ORDER_{FIRST,SECOND,THIRD} that I suspect could/should be bumped to SECOND/THIRD/FOURTH -- but I haven't had time to dig into it and examine the dependencies there.

May 3 2019, 2:31 PM
royger created D20148: geom: fix initialization order.
May 3 2019, 2:13 PM

Apr 24 2019

royger committed rP499848: xen: update to 4.12.
xen: update to 4.12
Apr 24 2019, 2:21 PM
royger closed D20038: xen: update to 4.12.
Apr 24 2019, 2:21 PM
royger created D20038: xen: update to 4.12.
Apr 24 2019, 9:18 AM

Mar 15 2019

royger committed rP495786: sysutils/xen-tools: add QEMU migration fix.
sysutils/xen-tools: add QEMU migration fix
Mar 15 2019, 12:38 PM
royger closed D19591: sysutils/xen-tools: add QEMU migration fix.
Mar 15 2019, 12:38 PM
royger added a comment to D19591: sysutils/xen-tools: add QEMU migration fix.

Shouldn't this patch be added to EXTRA_PATCHES?

Mar 15 2019, 11:51 AM
royger updated the diff for D19592: sysutils/xen-tools: remove docs option.
  • Use a explicit target to only build man pages, instead of using the generic docs target and expecting only the man pages to be built.
  • Remove the markdown dependency, the html files can be easily fetched from https://xenbits.xen.org/docs/.
Mar 15 2019, 11:45 AM
royger added a comment to D19592: sysutils/xen-tools: remove docs option.
In D19592#419471, @mat wrote:

What happened to all the html files in DOCSDIR?

Mar 15 2019, 11:26 AM
royger updated the summary of D19592: sysutils/xen-tools: remove docs option.
Mar 15 2019, 10:22 AM
royger created D19592: sysutils/xen-tools: remove docs option.
Mar 15 2019, 10:20 AM
royger created D19591: sysutils/xen-tools: add QEMU migration fix.
Mar 15 2019, 10:01 AM

Mar 13 2019

royger added a comment to D19521: sysutils/xen-tools: fix plist when DOCS option is disabled.
In D19521#418756, @mat wrote:

The manpages should not be behind the DOCS option, they should either always be installed, or be behind a MANPAGES option (that defaults to on).

Mar 13 2019, 3:01 PM

Mar 12 2019

royger committed rP495458: emulators/xen-kernel: backport fixes and apply XSAs.
emulators/xen-kernel: backport fixes and apply XSAs
Mar 12 2019, 3:02 PM
royger closed D19413: emulators/xen-kernel: backport fixes.
Mar 12 2019, 3:02 PM
royger added a comment to D19521: sysutils/xen-tools: fix plist when DOCS option is disabled.

Forgot to ask but can you commit this?

Mar 12 2019, 2:56 PM
royger updated the diff for D19413: emulators/xen-kernel: backport fixes.

Update to add patches for XSA-{284,287,290,292-294}

Mar 12 2019, 11:38 AM

Mar 11 2019

royger accepted D19521: sysutils/xen-tools: fix plist when DOCS option is disabled.

LGTM, thanks!

Mar 11 2019, 4:31 PM
royger added a comment to D19413: emulators/xen-kernel: backport fixes.

Ping?

Mar 11 2019, 4:31 PM

Mar 1 2019

royger updated the diff for D19413: emulators/xen-kernel: backport fixes.

Add a couple more of backported fixes that should improve the
user-experience, ie: prevent Xen from crashing or deadlocking.

Mar 1 2019, 5:27 PM
royger retitled D19413: emulators/xen-kernel: backport fixes from emulators/xen-kernel: fix for dom0 lapic EOIs Backport a patch to correctly forward dom0 lapic EOIs to the underlying lapic when required. to emulators/xen-kernel: backport fixes.
Mar 1 2019, 5:25 PM
royger created D19413: emulators/xen-kernel: backport fixes.
Mar 1 2019, 8:24 AM

Feb 27 2019

royger committed rP494055: xen-kernel & xen-tools: fix ignore message.
xen-kernel & xen-tools: fix ignore message
Feb 27 2019, 2:01 PM
royger closed D19387: xen-kernel & xen-tools: fix ignore message.
Feb 27 2019, 2:01 PM
royger created D19387: xen-kernel & xen-tools: fix ignore message.
Feb 27 2019, 1:32 PM