[stgt] RFC [patch] Search /sys/class/bsg/<ctl>/dev for bsg major number.
Mark Harvey
markh794 at gmail.com
Sun Aug 22 01:53:12 CEST 2010
Cheers
I'll look tomorrow when time permits :)
Sent from my iPhone
On 22/08/2010, at 9:02, "Nicholas A. Bellinger" <nab at linux-iscsi.org> wrote:
> On Sat, 2010-08-21 at 16:07 +1000, Mark Harvey wrote:
>> Apologies for including this patch as an attachment..
>> I haven't figured out how to successfully do it from the office.
>>
>> Included inline for easier reading / review.
>>
>> On Fri, Aug 20, 2010 at 7:29 PM, Nicholas A. Bellinger
>> <nab at linux-iscsi.org> wrote:
>>> On Fri, 2010-08-20 at 11:22 +0900, FUJITA Tomonori wrote:
>>>> On Thu, 19 Aug 2010 12:51:54 -0700
>>>> "Nicholas A. Bellinger" <nab at linux-iscsi.org> wrote:
>>>>> So it looks like the failure is attributed to my hardcoding of major 254
>>>>> in the original STGT/BSG patch (note the comment above the conditional
>>>>> check). Here is a quick fix:
>>>>>
>>>>> diff --git a/usr/bs_sg.c b/usr/bs_sg.c
>>>>> index cda5cd2..a369570 100644
>>>>> --- a/usr/bs_sg.c
>>>>> +++ b/usr/bs_sg.c
>>>>> @@ -289,7 +289,7 @@ static int chk_sg_device(char *path)
>>>>> return 0;
>>>>>
>>>>> /* This is not yet defined in include/linux/major.h.. */
>>>>> - if (major(st.st_rdev) == 254)
>>>>> + if ((major(st.st_rdev) == 254) || (major(st.st_rdev) == 253))
>>>>> return 1;
>>>>>
>>>>> return -1;
>>>>
>>>> Well, seems that nobody uses the bsg support except for you. :)
>>>>
>>>> The above patch is still wrong.
>>>>
>>>> bsg major number can be anything. So you need to get it from
>>>> /proc/devices.
>>>>
>>>> Or you can also get it from /sys/class/bsg/, which is created by
>>>> kernel so available on any distributions.
>>>>
>>>> You can steal the bsg device handling code from sg3-utils. Maybe
>>>> someone could create something like libbsg.so
>>>
>>> Hmmm indeed, thanks for point this one out. I will have look at
>>> adressing for this the BSG efforts with STGT and QEMU-KVM using the
>>> recommended major:minor output from /sys/class/bsg/$HCTL/dev.
>>>
>>> Best,
>>>
>>> --nab
>>>
>>>
>>
>>> From a90611630879205ef497821552474e894d463c94 Mon Sep 17 00:00:00 2001
>> From: Mark Harvey <markh794 at gmail.com>
>> Date: Sat, 21 Aug 2010 15:22:48 +1000
>> Subject: Check bsg major number from /sys/class/bsg/<c:t:l>/dev
>>
>> Signed-off-by: Mark Harvey <markh794 at gmail.com>
>> ---
>> usr/bs_sg.c | 35 ++++++++++++++++++++++++++++++++---
>> 1 files changed, 32 insertions(+), 3 deletions(-)
>>
>> diff --git a/usr/bs_sg.c b/usr/bs_sg.c
>> index dfd49c3..7fc9787 100644
>> --- a/usr/bs_sg.c
>> +++ b/usr/bs_sg.c
>> @@ -269,6 +269,34 @@ static void bs_sg_cmd_complete(int fd, int
>> events, void *data)
>> cmd->scsi_cmd_done(cmd, io_hdr.status);
>> }
>>
>> +static int get_bsg_major(char *path)
>> +{
>> + FILE *devfd;
>> + char majorno[8];
>> + char dev[128];
>> + int ch;
>> + int i;
>> +
>> + dev[0] = '\0';
>> + strncat(dev, "/sys/class/bsg/", 16);
>> + strncat(dev, &path[9], 64);
>> + strncat(dev, "/dev", 5);
>> + devfd = fopen(dev, "r");
>> + if (!devfd)
>> + return -1;
>> + ch = fgetc(devfd);
>> + for (i = 0; i < 7; i++) {
>> + if (ch == ':') {
>> + majorno[i] = '\0';
>> + break;
>> + }
>> + majorno[i] = ch;
>> + ch = fgetc(devfd);
>> + }
>> + fclose(devfd);
>> + return atoi(majorno);
>> +}
>> +
>> static int chk_sg_device(char *path)
>> {
>> struct stat st;
>> @@ -287,9 +315,10 @@ static int chk_sg_device(char *path)
>> if (major(st.st_rdev) == SCSI_GENERIC_MAJOR)
>> return 0;
>>
>> - /* This is not yet defined in include/linux/major.h.. */
>> - if (major(st.st_rdev) == 254)
>> - return 1;
>> + if (!strncmp("/dev/bsg", path, 8)) {
>> + if (major(st.st_rdev) == get_bsg_major(path))
>> + return 0;
>> + }
>>
>> return -1;
>> }
>
> Hi Mark,
>
> Here is my updated version of your initial code for the megasas
> 8708EM2 / SGL passthrough friendly qemu-kvm.git/scsi-bsg:
>
> http://git.kernel.org/?p=virt/kvm/nab/qemu-kvm.git;a=commitdiff;h=075901d9bc1fea670bef84f55806f1bc8b2a5e9e
>
> Please feel free to improve your patch using this code and respin
> another patch for tomo-san to review.
>
> Thanks!
>
> --nab
>
>
--
To unsubscribe from this list: send the line "unsubscribe stgt" in
the body of a message to majordomo at vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
More information about the stgt
mailing list