only finds 26 scsi devices

[keithbusch]

The pattern match only looks for sda - sdz, but we have more than that. Need this:

diff --git a/linux/DtaDevOS.cpp b/linux/DtaDevOS.cpp
index 5261e73..fc0f33e 100644
--- a/linux/DtaDevOS.cpp
+++ b/linux/DtaDevOS.cpp
@@ -135,7 +135,8 @@ int  DtaDevOS::diskScan()
     if(dir!=NULL)
     {
         while((dirent=readdir(dir))!=NULL) {
-            if((!fnmatch("sd[a-z]",dirent->d_name,0)) ||
+            if((!fnmatch("sd[a-z]",dirent->d_name,0)) ||
+                    (!fnmatch("sd[a-z][a-z]",dirent->d_name,0)) ||
                     (!fnmatch("nvme[0-9]",dirent->d_name,0)) ||
                     (!fnmatch("nvme[0-9][0-9]",dirent->d_name,0))
                     ) {

[keithbusch]

found my keys, able to send pull requests again. #480

[dtasupport]

How about this: std::vector<std::string> DtaLinux::generateDtaDriveDevRefs() { std::vector<std::string> devrefs; DIR *dir = opendir("/dev"); if (dir==NULL) { LOG(E) << "Can't read /dev ?!"; return devrefs; } struct dirent *dirent; while (NULL != (dirent=readdir(dir))) { std::string devref=std::string("/dev/")+dirent->d_name; struct stat s; stat(devref.c_str(), &s); // LOG(E) << devref // << " st_ino:" << HEXON(4) << s.st_ino // << " st_dev:" << HEXON(4) << s.st_dev // << " st_rdev:" << HEXON(4) << s.st_rdev // << " st_rdev>>8:" << HEXON(4) << (s.st_rdev>>8) // << " st_rdev&0xF:" << HEXON(4) << (s.st_rdev & 0x000F) // ; const unsigned long device_type=s.st_rdev >> 8; const unsigned char device_part=s.st_rdev & 15; typedef enum _rdev_type { SCSI_DRIVE=8, NVME_DRIVE=259, } rdev_type; if (device_part==0 && (device_type==rdev_type::SCSI_DRIVE || device_type==rdev_type::NVME_DRIVE)) { // LOG(E) << devref << " accepted." // ; devrefs.push_back(devref); } } closedir(dir); std::sort(devrefs.begin(),devrefs.end()); return devrefs; } As you can tell from the commented-out-debug-logging-to-stderr, this is tips code. The idea is that this is what each OS needs to be able to do, and then enumerateDtaDriveDevRefs is just a OS-independent filtering loop that validates our access to those devices (OK, using an OS-specific routine openDeviceHandle there, too).

…

On Thu, Sep 5, 2024 at 12:40 PM 'Keith Busch' via Drive Trust Alliance - Support ***@***.***> wrote: found my keys, able to send pull requests again. #480 <#480> — Reply to this email directly, view it on GitHub <#479 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AD52TFBYJBU6MJCL7XTOKO3ZVCCQXAVCNFSM6AAAAABNM2XLZSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMZSGE4TCNJQG4> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[keithbusch]

{ SCSI_DRIVE=8, NVME_DRIVE=259, }

The 259 value is not reserved for the NVMe major. Any block device can use that.