> Ok. Can you move the is_visible check to scsi_sysfs_add_devices, should be fine to place it just after the sdev_state check. Sure, will move it in next patch. -----Original Message----- From: 'Christoph Hellwig' [mailto:hch@xxxxxxxxxxxxx] Sent: Wednesday, September 24, 2014 9:38 AM To: Subhash Jadavani Cc: 'Christoph Hellwig'; 'Dolev Raviv'; James.Bottomley@xxxxxxxxxxxxxxxxxxxxx; linux-scsi@xxxxxxxxxxxxxxx; linux-scsi-owner@xxxxxxxxxxxxxxx; linux-arm-msm@xxxxxxxxxxxxxxx; santoshsy@xxxxxxxxx Subject: Re: [PATCH V5 02/17] scsi: sysfs: don't add scsi_device if its already added On Wed, Sep 24, 2014 at 09:27:47AM -0700, Subhash Jadavani wrote: > No, It happens in this sequence of events: > 1. LLD calls the __scsi_add_device() for well known logical units > before > scsi_scan_host() (This is done as part of [PATCH V5 10/17] scsi: ufs: > manually add well known logical units). __scsi_add_device() will also > add the scsi device to sysfs by calling scsi_sysfs_add_sdev() 2. Now > LDD calls the scsi_scan_host() (Note that CONFIG_SCSI_SCAN_ASYNC is > enabled) which will schedule the async scan. Device reports only > normal LUs (no w-lus reported here) when REPORT LUNs in sent with > SELECT REPORT cleared. At the end of async scan, > scsi_finish_async_scan() calls scsi_sysfs_add_devices(). > scsi_sysfs_add_devices() iterates through all the scsi_device > instances attached to Scsi_Host hence end up calling > scsi_sysfs_add_sdev() 2nd time for the scsi_device which were already > added explicitly by LDD. Ok. Can you move the is_visible check to scsi_sysfs_add_devices, should be fine to place it just after the sdev_state check. -- To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
