2.4.0-test6 boot problems
We have a Lion w/ 3gb of RAM, 2gb of which begins at 0x100000000 physical (i.e. 4gb), that will not boot 2.4.0-test6 unless we limit it to using the memory below 4gb. The system panics because it is unable to mount the root file system. Root cannot be mounted because we do not find a valid partiton table on the disk (note that this same disk boots test4 fine and will boot test6 with memory restricted to below 4gb).
We don't find a partition table because the bread() of it fails. We don't get an I/O error, but rather it seems the read isn't being placing the data into the chosen buffer. I verifed this by placing debug code into bread(). I pre-fill the buffer with all ones (0xfff.....) and the buffer is unchanged upon completion of the read. I suspect the problem may lie within the CONFIG_SWIOTLB, it seems we're truncating the upper 32 bits somewhere.
Has anyone else come across this, or know if a fix exists? I'd be happy to try out a patch if one exists.
Thanks,
Pat
Signature · 2 lines
Patrick O'Rourke [email protected]
Hello,
If you are using QLogic, please make sure you have enabled Greater than 4gb addressing in the Firmware (use the Alt-Q) option. If that still failes, let me know.
Thanks,
Goutham
-----Original Message-----
From: Pat O'Rourke [mailto:[email protected]]
Sent: Monday, August 21, 2000 8:45 AM
To: [email protected]
Subject: [Linux-ia64] 2.4.0-test6 boot problems
We have a Lion w/ 3gb of RAM, 2gb of which begins at 0x100000000 physical (i.e. 4gb), that will not boot 2.4.0-test6 unless we limit it to using the memory below 4gb. The system panics because it is unable to mount the root file system. Root cannot be mounted because we do not find a valid partiton table on the disk (note that this same disk boots test4 fine and will boot test6 with memory restricted to below 4gb).
We don't find a partition table because the bread() of it fails. We don't get an I/O error, but rather it seems the read isn't being placing the data into the chosen buffer. I verifed this by placing debug code into bread(). I pre-fill the buffer with all ones (0xfff.....) and the buffer is unchanged upon completion of the read. I suspect the problem may lie within the CONFIG_SWIOTLB, it seems we're truncating the upper 32 bits somewhere.
Has anyone else come across this, or know if a fix exists? I'd be happy to try out a patch if one exists.
Thanks,
Pat
Signature · 9 lines
Patrick O'Rourke [email protected] _______________________________________________ Linux-IA64 mailing list [email protected] http://lists.linuxia64.org/lists/listinfo/linux-ia64
Quoted 1 line · Pat O'Rourke
We don't find a partition table because the
bread()of it fails.
You need to enable the 4GB support in the Qlogic SCSI BIOS.
Bill
Thanks to all those who responded. Yes the problem was that I had 4gb support turned off in the Qlogic. Once I enabled the system booted fine.
Pat
Signature · 2 lines
Patrick O'Rourke [email protected]