qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Max Krasnyansky <maxk@kernel.org>
To: Anthony Liguori <anthony@codemonkey.ws>
Cc: qemu-devel@nongnu.org, kvm@vger.kernel.org
Subject: [Qemu-devel] Re: [PATCH 2/5] husb: support for USB host device auto connect.
Date: Thu, 14 Aug 2008 12:38:59 -0700	[thread overview]
Message-ID: <48A489D3.5070900@kernel.org> (raw)
In-Reply-To: <48A46033.3070200@codemonkey.ws>

Anthony Liguori wrote:
> Max Krasnyansky wrote:
>> Anyway, it's implemented using a periodic timer that scans host devices
>> and grabs those that match the filter. Timer is started when the first
>> filter is added.
>>   
> 
> Again, there has to be a way to get notified of usb device add/remove in
> Linux.

Yes the best way is probably registering with HAL via dbus.
Do you mind if we do something like that incrementally ?
ie Merge the current patch then change polling to HAL registration.
That would simplify testing/patch stacking for me. Existing users are
not affected anyway since the timer is activated only when new filter is added.

>> +static USBHostDevice *hostdev_find(int bus_num, int addr)
>> +{
>> +    USBHostDevice *s = hostdev_list;
>> +    while (s) {
>> +        if (s->bus_num == bus_num && s->addr == addr)
>> +            return s;
>> +        s = s->next;
>> +    }
>> +    return NULL;
>> +}
>>   
> 
> There are generic list macros in sys-queue.h.  They aren't universally
> used though so it's up to you whether you think it's appropriate.
Oh, I looked around (mostly USB code) and did not see any generic lists.
I'd say ideally we should clean things up to use generic stuff as much as
possible. There seems to be a lot of duplication currently.

>> +struct USBAutoFilter {
>> +    struct USBAutoFilter *next;
>> +    int bus_num;
>> +    int addr;
>> +    int vendor_id;
>> +    int product_id;
>> +};
>> +
>> +static QEMUTimer *usb_auto_timer;
>> +static struct USBAutoFilter *usb_auto_filter;
>> +
>> +static int usb_host_auto_scan(void *opaque, int bus_num, int addr,
>> +                     int class_id, int vendor_id, int product_id,
>> +                     const char *product_name, int speed)
>> +{
>> +    struct USBAutoFilter *f;
>> +    struct USBDevice *dev;
>> +
>> +    /* Ignore hubs */
>> +    if (class_id == 9)
>> +        return 0;
>> +
>> +    for (f = usb_auto_filter; f; f = f->next) {
>> +        // printf("Auto match: bus_num %d addr %d vid %d pid %d\n",
>> +    //    bus_num, addr, vendor_id, product_id);
>>   
> 
> An #ifdef guard would be nicer.
Yep. Missed that one.

>> +    if (f->bus_num >= 0 && f->bus_num != bus_num)
>> +            continue;
>> +
>> +    if (f->addr >= 0 && f->addr != addr)
>> +            continue;
>> +
>> +    if (f->vendor_id >= 0 && f->vendor_id != vendor_id)
>> +            continue;
>> +
>> +    if (f->product_id >= 0 && f->product_id != product_id)
>> +            continue;
>> +
>> +        /* We got a match */
>> +
>> +        /* Allredy attached ? */
>> +        if (hostdev_find(bus_num, addr))
>> +            return 0;
>> +
>> +        printf("Auto open: bus_num %d addr %d\n", bus_num, addr);
>>   
> 
> Please make this debug too.
Will do.


> 
>> +    dev = usb_host_device_open_addr(bus_num, addr, product_name);
>> +    if (dev)
>> +        usb_device_add_dev(dev);
>> +    }
>> +
>> +    return 0;
>> +}
>> +
>> +static void usb_host_auto_timer(void *unused)
>> +{
>> +    usb_host_scan(NULL, usb_host_auto_scan);
>> +    qemu_mod_timer(usb_auto_timer, qemu_get_clock(rt_clock) + 2000);
>> +}
>> +
>> +/*
>> + * Add autoconnect filter
>> + * -1 means 'any' (device, vendor, etc)
>> + */
>> +static void usb_host_auto_add(int bus_num, int addr, int vendor_id,
>> int product_id)
>> +{
>> +    struct USBAutoFilter *f = qemu_mallocz(sizeof(*f));
>> +    if (!f) {
>> +        printf("Failed to allocate auto filter\n");
>> +    return;
>> +    }
>> +
>> +    f->bus_num = bus_num;
>> +    f->addr    = addr;
>> +    f->vendor_id  = vendor_id;
>> +    f->product_id = product_id;
>> +
>> +    if (!usb_auto_filter) {
>> +        /*
>> +         * First entry. Init and start the monitor.
>> +         * Right now we're using timer to check for new devices.
>> +         * If this turns out to be too expensive we can move that
>> into a +         * separate thread.
>> +         */
>> +    usb_auto_timer = qemu_new_timer(rt_clock, usb_host_auto_timer,
>> NULL);
>> +    if (!usb_auto_timer) {
>> +            printf("Failed to allocate timer\n");
>>   
> 
> Please print errors to stderr.

Will do.

Max

  reply	other threads:[~2008-08-14 19:39 UTC|newest]

Thread overview: 37+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2008-08-14  4:22 [Qemu-devel] [PATCH 0/5] Various USB fixes and improvements Max Krasnyansky
2008-08-14  4:22 ` [Qemu-devel] [PATCH 1/5] husb: support for USB host device auto disconnect Max Krasnyansky
2008-08-14 16:28   ` [Qemu-devel] " Anthony Liguori
2008-08-14 19:26     ` Max Krasnyansky
2008-08-14 21:41     ` Max Krasnyansky
2008-08-14  4:22 ` [Qemu-devel] [PATCH 2/5] husb: support for USB host device auto connect Max Krasnyansky
2008-08-14 16:41   ` [Qemu-devel] " Anthony Liguori
2008-08-14 19:38     ` Max Krasnyansky [this message]
2008-08-14 20:21       ` Anthony Liguori
2008-08-14 20:34         ` Max Krasnyansky
2008-08-14 20:41           ` Anthony Liguori
2008-08-14 21:14             ` François Revol
2008-08-15  7:46             ` Guido Günther
2008-08-15 18:24               ` Max Krasnyansky
2008-08-15 18:31                 ` Javier Guerra
2008-08-18 18:21                   ` Max Krasnyansky
2008-08-18 18:52                     ` Javier Guerra
2008-08-18 18:56                       ` Jamie Lokier
2008-08-17  7:52                 ` Avi Kivity
2008-08-18 18:46                   ` Max Krasnyansky
2008-08-18 14:11                 ` Anthony Liguori
2008-08-18 18:16                   ` Max Krasnyansky
2008-08-14  4:22 ` [Qemu-devel] [PATCH 3/5] usb: generic packet handler cleanup and documentation Max Krasnyansky
2008-08-14  4:22 ` [Qemu-devel] [PATCH 4/5] uhci: rewrite UHCI emulator, fully async operation with multiple outstanding transactions Max Krasnyansky
2008-08-14 17:51   ` [Qemu-devel] " Anthony Liguori
2008-08-14 19:49     ` Max Krasnyansky
2008-10-11 23:54   ` [Qemu-devel] " Juergen Lock
2008-10-15 19:54     ` Max Krasnyansky
2008-10-15 22:05       ` andrzej zaborowski
2008-10-16 21:25         ` Juergen Lock
2008-08-14  4:22 ` [Qemu-devel] [PATCH 5/5] husb: rewrite Linux host USB layer, fully async operation Max Krasnyansky
2008-08-15 14:24   ` Paul Brook
2008-08-15 19:04     ` Max Krasnyansky
2008-08-15 19:53       ` Paul Brook
2008-08-18 18:40         ` Max Krasnyansky
2008-08-14 17:55 ` [Qemu-devel] Re: [PATCH 0/5] Various USB fixes and improvements Anthony Liguori
2008-08-14 19:55   ` Max Krasnyansky

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=48A489D3.5070900@kernel.org \
    --to=maxk@kernel.org \
    --cc=anthony@codemonkey.ws \
    --cc=kvm@vger.kernel.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).