Loading DSDT from initrd, can it be done from grub instead?
A message that is not in the archive (sent off-list or not yet received): [email protected]
Quoted 6 lines · David Härdeman
In commit
9a9e0d685553af76cb6ae2af93cca4913e7fcd47theACPI_CUSTOM_DSDT_INITRDoption was removed.I was wondering if it'd be feasible to hack grub to load a custom DSDT instead (since grub can read filesystems) and then pass a pointer to a memory location containing the DSDT to the kernel?
Um, why? What stops you from reverting the offending commit?
Without further info, I'd assume the feature was removed on policy grounds. I.e. it's not nice to be effectively patching the BIOS yourself, and vendors have fixed most of their BIOS's / linux is slightly more tolerant to broken ones. I don't know whats wrong with reading files from the initrd.
Alan
Quoted 8 lines · Alan Jenkins
Quoted 6 lines · David Härdeman
In commit
9a9e0d685553af76cb6ae2af93cca4913e7fcd47theACPI_CUSTOM_DSDT_INITRDoption was removed.I was wondering if it'd be feasible to hack grub to load a custom DSDT instead (since grub can read filesystems) and then pass a pointer to a memory location containing the DSDT to the kernel?
Quoted 7 lines
Um, why? What stops you from reverting the offending commit?
Without further info, I'd assume the feature was removed on policy grounds. I.e. it's not nice to be effectively patching the BIOS yourself, and vendors have fixed most of their BIOS's / linux is slightly more tolerant to broken ones. I don't know whats wrong with reading files from the initrd.
It was reverted because the code was not really ready and caused some problems.
Quoted 2 lines
Alan
Gabriel
Quoted 9 lines · Alan Jenkins
Quoted 6 lines · David Härdeman
In commit
9a9e0d685553af76cb6ae2af93cca4913e7fcd47theACPI_CUSTOM_DSDT_INITRDoption was removed.I was wondering if it'd be feasible to hack grub to load a custom DSDT instead (since grub can read filesystems) and then pass a pointer to a memory location containing the DSDT to the kernel?
Um, why? What stops you from reverting the offending commit?
Because doing it from grub (if possible) would allow custom DSDT's to be loaded without patching the kernel (e.g. allows distro kernels to be used) and without introducing code that opens files from kernel space and mucks with both the VFS and ACPI in ugly ways?
Quoted 5 lines
Without further info, I'd assume the feature was removed on policy grounds. I.e. it's not nice to be effectively patching the BIOS yourself, and vendors have fixed most of their BIOS's / linux is slightly more tolerant to broken ones. I don't know whats wrong with reading files from the initrd.
If you do not understand the issue at hand, why comment?
Signature · 1 line
David Härdeman
Quoted 20 lines · David Härdeman
Quoted 11 lines · Alan Jenkins
Quoted 7 lines · David Härdeman
In commit
9a9e0d685553af76cb6ae2af93cca4913e7fcd47theACPI_CUSTOM_DSDT_INITRDoption was removed.I was wondering if it'd be feasible to hack grub to load a custom DSDT instead (since grub can read filesystems) and then pass a pointer to a memory location containing the DSDT to the kernel?
Um, why? What stops you from reverting the offending commit?
Because doing it from grub (if possible) would allow custom DSDT's to be loaded without patching the kernel (e.g. allows distro kernels to be used) and without introducing code that opens files from kernel space and mucks with both the VFS and ACPI in ugly ways?
Quoted 9 lines
Quoted 6 lines
Without further info, I'd assume the feature was removed on policy grounds. I.e. it's not nice to be effectively patching the BIOS yourself, and vendors have fixed most of their BIOS's / linux is slightly more tolerant to broken ones. I don't know whats wrong with reading files from the initrd.
If you do not understand the issue at hand, why comment?
Mea cupla. Apologies if my tone was also inappropriate.
I saw a short question on an interesting subject, without a context to suggest an ongoing
conversation. I would have saved myself some embarassment if I remembered that linux really doesn't like in-kernel filesystem access; I just didn't make the connection.
I guessed wrong, didn't see where you were coming from, and didn't take the effort to resolve the SHA commit id. I literally don't have the space for a checkout on my main PC and I feel slightly guilty when I hit on the gitweb host.
Ironically, I just looked it up and Google knew what the SHA ID was. You might be interested to know that another DH (Dave Hansen) made the same suggestion, but using kexec instead of grub to load the DSDT and boot kernel.
Alan