Re: USB stack

On Mon, 8 Jan 2018 13:17:22 +0800 "blubee blubeeme" <gurenchan_at_gmail.com> said

> On Mon, Jan 8, 2018 at 8:03 AM, Jon Brawn <jon_at_brawn.org> wrote:
> 
> >
> >
> > > On Jan 7, 2018, at 5:44 PM, Jon Brawn <jon_at_brawn.org> wrote:
> > >
> > >
> > >> On Jan 6, 2018, at 10:18 PM, blubee blubeeme <gurenchan_at_gmail.com>
> > wrote:
> > >>
> > >> On Sun, Jan 7, 2018 at 12:11 PM, Warner Losh <imp_at_bsdimp.com> wrote:
> > >>
> > >>>
> > >>>
> > >>> On Sat, Jan 6, 2018 at 8:56 PM, blubee blubeeme <gurenchan_at_gmail.com>
> > >>> wrote:
> > >>>
> > >>>> I ask does FreeBSD usb stack actually implements USB spec 2.0 or
> > greater
> > >>>> and the topic gets derailed...?
> > >>>>
> > >>>
> > >>> Yes, it does.
> > >>>
> > >>>
> > >>>> Are you guys saying that 7-8MB/s is USB speeds?
> > >>>>
> > >>>
> > >>> I've gotten up to 24MB/s for maybe a decade. That's not possible with
> > USB
> > >>> 1.x. More recently, I've maxed out the writes on a USB stick at about
> > >>> 75MB/s (the fastest it will do), which isn't possible with USB 2.0...
> > I've
> > >>> not tried USB3 with an SSD that can do more....
> > >>>
> > >>> Warner
> > >>>
> > >>>
> > >>>> On Thu, Jan 4, 2018 at 6:44 PM, O'Connor, Daniel <darius_at_dons.net.au>
> > >>>> wrote:
> > >>>>
> > >>>>>
> > >>>>>
> > >>>>>> On 4 Jan 2018, at 09:23, Gary Jennejohn <gljennjohn_at_gmail.com>
> > wrote:
> > >>>>>>> What is an "LG v30"?
> > >>>>>>>
> > >>>>>> It's a smartphone from LG and only supports USB2 speed.  The
> > reported
> > >>>>>> transfer rate is no big surprise.
> > >>>>>
> > >>>>> OK thanks.
> > >>>>>
> > >>>>> --
> > >>>>> Daniel O'Connor
> > >>>>> "The nice thing about standards is that there
> > >>>>> are so many of them to choose from."
> > >>>>> -- Andrew Tanenbaum
> > >>>>> GPG Fingerprint - 5596 B766 97C0 0E94 4347 295E E593 DC20 7B3F CE8C
> > >>>>>
> > >>>>>
> > >>>> _______________________________________________
> > >>>> freebsd-current_at_freebsd.org mailing list
> > >>>> https://lists.freebsd.org/mailman/listinfo/freebsd-current
> > >>>> To unsubscribe, send any mail to "freebsd-current-unsubscribe_at_
> > freebsd.org
> > >>>> "
> > >>>>
> > >>>
> > >>> I just connected a Transcend StorageJet 1TB hdd not a mobile phone
> > >> -------------------------------------------------------------------
> > >> Jan  7 11:56:56 blubee kernel: umass0 on uhub0
> > >> Jan  7 11:56:56 blubee kernel: umass0: <StoreJet Transcend StoreJet
> > >> Transcend, class 0/0, rev 3.00/80.00, addr 4> on usbus0
> > >> Jan  7 11:56:56 blubee kernel: umass0:  SCSI over Bulk-Only; quirks =
> > 0x0100
> > >> Jan  7 11:56:56 blubee kernel: umass0:3:0: Attached to scbus3
> > >> Jan  7 11:56:56 blubee kernel: da0 at umass-sim0 bus 0 scbus3 target 0
> > lun 0
> > >> Jan  7 11:56:56 blubee kernel: da0: <StoreJet Transcend 0> Fixed Direct
> > >> Access SPC-4 SCSI device
> > >> Jan  7 11:56:56 blubee kernel: da0: Serial Number W9328YZN
> > >> Jan  7 11:56:56 blubee kernel: da0: 400.000MB/s transfers
> > >> Jan  7 11:56:56 blubee kernel: da0: 953869MB (1953525168 512 byte
> > sectors)
> > >> Jan  7 11:56:56 blubee kernel: da0: quirks=0x2<NO_6_BYTE>
> > >> Jan  7 12:06:08 blubee kernel: lock order reversal:
> > >> Jan  7 12:06:08 blubee kernel:  1st 0xfffffe07c26336c0 bufwait
> > (bufwait) _at_
> > >> /usr/src/sys/vm/vm_pager.c:374
> > >> Jan  7 12:06:08 blubee kernel:  2nd 0xfffff80148c425f0 zfs (zfs) _at_
> > >> /usr/src/sys/dev/md/md.c:952
> > >> Jan  7 12:06:08 blubee kernel: stack backtrace:
> > >> Jan  7 12:06:08 blubee kernel: #0 0xffffffff80acfa03 at
> > >> witness_debugger+0x73
> > >> Jan  7 12:06:08 blubee kernel: #1 0xffffffff80acf882 at
> > >> witness_checkorder+0xe02
> > >> Jan  7 12:06:08 blubee kernel: #2 0xffffffff80a41b8e at
> > >> lockmgr_lock_fast_path+0x1ae
> > >> Jan  7 12:06:08 blubee kernel: #3 0xffffffff81094309 at
> > VOP_LOCK1_APV+0xd9
> > >> Jan  7 12:06:08 blubee kernel: #4 0xffffffff80b4ac36 at _vn_lock+0x66
> > >> Jan  7 12:06:08 blubee kernel: #5 0xffffffff80611d32 at
> > mdstart_vnode+0x442
> > >> Jan  7 12:06:08 blubee kernel: #6 0xffffffff806102ce at md_kthread+0x1fe
> > >> Jan  7 12:06:08 blubee kernel: #7 0xffffffff80a2d654 at fork_exit+0x84
> > >> Jan  7 12:06:08 blubee kernel: #8 0xffffffff80ef5e0e at
> > fork_trampoline+0xe
> > >> Jan  7 12:06:15 blubee kernel: lock order reversal:
> > >> Jan  7 12:06:15 blubee kernel:  1st 0xfffffe07c41d5dc0 bufwait
> > (bufwait) _at_
> > >> /usr/src/sys/kern/vfs_bio.c:3562
> > >> Jan  7 12:06:15 blubee kernel:  2nd 0xfffff8002bb31a00 dirhash
> > (dirhash) _at_
> > >> /usr/src/sys/ufs/ufs/ufs_dirhash.c:281
> > >> Jan  7 12:06:15 blubee kernel: stack backtrace:
> > >> Jan  7 12:06:15 blubee kernel: #0 0xffffffff80acfa03 at
> > >> witness_debugger+0x73
> > >> Jan  7 12:06:15 blubee kernel: #1 0xffffffff80acf882 at
> > >> witness_checkorder+0xe02
> > >> Jan  7 12:06:15 blubee kernel: #2 0xffffffff80a748a8 at _sx_xlock+0x68
> > >> Jan  7 12:06:15 blubee kernel: #3 0xffffffff80d6a28d at
> > ufsdirhash_add+0x3d
> > >> Jan  7 12:06:15 blubee kernel: #4 0xffffffff80d6d119 at
> > ufs_direnter+0x459
> > >> Jan  7 12:06:15 blubee kernel: #5 0xffffffff80d76313 at
> > ufs_makeinode+0x613
> > >> Jan  7 12:06:15 blubee kernel: #6 0xffffffff80d71ff4 at ufs_create+0x34
> > >> Jan  7 12:06:15 blubee kernel: #7 0xffffffff810919e3 at
> > VOP_CREATE_APV+0xd3
> > >> Jan  7 12:06:15 blubee kernel: #8 0xffffffff80b4a53d at
> > vn_open_cred+0x2ad
> > >> Jan  7 12:06:15 blubee kernel: #9 0xffffffff80b42e92 at
> > kern_openat+0x212
> > >> Jan  7 12:06:15 blubee kernel: #10 0xffffffff80f16d2b at
> > amd64_syscall+0x79b
> > >> Jan  7 12:06:15 blubee kernel: #11 0xffffffff80ef5b7b at
> > Xfast_syscall+0xfb
> > >>
> > >>
> > >> Is the slow transfers user error?
> > >
> > > Wotcha!
> > >
> > > I don’t see any read or write performance figures anywhere? Also, is
> > this CURRENT? If so, aren’t all the debug / warning features that are
> > turned on by default in CURRENT at the moment going to have an effect on
> > throughput? Especially if you’re writing through a filesystem where
> > directory and file accesses will each require a lock to be taken, if only
> > for a short while? If you want to get closer to the true USB speed of the
> > device, stop mounting it and copying files to the filesystem, but instead
> > just dd data onto and off of the device directly, and measure how fast that
> > goes. Remember to backup your data from the card first…
> > >
> > > Jon.
> > >
> > >
> >
> > Also, is the SD card physically inside the phone, and you are using a USB
> > cord to connect the phone to the FreeBSD computer by any chance?
> >
> > Jon
> >
> > _at_Mark Millard
> I use sysutils/simple-mtpfs to mount the android device.
> when I mount the phone through USB this is the relevant section:
> /dev/fuse                                   356311 78912 277398    22%
> /mnt
> 
> That's the most complicated mount process that I use,
> for the ssd it's just a simple mount /dev/device /mnt
> relevant output:
> /dev/da0                                    923913 121450 728550    14%
> /mnt
> 
> Can you tell me what information you're looking for so that I can gather it
> all up and send it.
> 
> _at_Jon Brawn
> I am running current because I handle admin a few other boxes that are on
> RELEASE so I have
> to run in current to make sure they don't have it.
It's not CURRENT that's the problem, but GENERIC (WITNESS et al; that causes
contention) -- see; performance loss. :-)

> I do wonder about those locks as well but, they should only affect the
> multiple small files,
> not so much the larger files.
> 
> The microsd card is physically inside the phone.
--Chris